diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 098033fca3a8553f85600759e98ebe97db5d4897..96e8cd1c1478825ec5705a81d9848043f77d58e9 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": "20260701T000029Z", - "generated_at": "2026-07-01T00:11:01Z", + "snapshot_id": "20260701T060031Z", + "generated_at": "2026-07-01T06:12:27Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1152,7 +1152,7 @@ "cluster_id": "cluster-46894-6", "summary": "Cluster of 3 issues and 3 PRs centered on issue #46894.", "status": "open", - "confidence": 0.6, + "confidence": 0.8, "canonical_issue_number": 46894, "canonical_pr_number": 46903, "issue_numbers": [ @@ -1167,6 +1167,7 @@ ], "evidence_types": [ "closing_reference", + "shared_issue_target", "soft_similarity" ], "pr_comparisons": [ @@ -1363,235 +1364,6 @@ } ] }, - { - "cluster_id": "cluster-46927-6", - "summary": "Cluster of 6 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 46927, - "issue_numbers": [], - "pr_numbers": [ - 46927, - 46949, - 46950, - 46951, - 46970, - 46971 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 46927, - "right_pr_number": 46949, - "code_similarity": 0.157, - "size_similarity": 0.786, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46927, - "right_pr_number": 46950, - "code_similarity": 0.2, - "size_similarity": 0.966, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.049, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46927, - "right_pr_number": 46951, - "code_similarity": 0.079, - "size_similarity": 0.393, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46927, - "right_pr_number": 46970, - "code_similarity": 0.195, - "size_similarity": 0.929, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.059, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46927, - "right_pr_number": 46971, - "code_similarity": 0.162, - "size_similarity": 0.786, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.029, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46949, - "right_pr_number": 46950, - "code_similarity": 0.152, - "size_similarity": 0.759, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46949, - "right_pr_number": 46951, - "code_similarity": 0.1, - "size_similarity": 0.5, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46949, - "right_pr_number": 46970, - "code_similarity": 0.169, - "size_similarity": 0.846, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46949, - "right_pr_number": 46971, - "code_similarity": 0.206, - "size_similarity": 1.0, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.038, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46950, - "right_pr_number": 46951, - "code_similarity": 0.076, - "size_similarity": 0.379, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46950, - "right_pr_number": 46970, - "code_similarity": 0.182, - "size_similarity": 0.897, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.017, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46950, - "right_pr_number": 46971, - "code_similarity": 0.393, - "size_similarity": 0.759, - "file_overlap": 0.5, - "area_overlap": 0.182, - "patch_similarity": 0.184, - "shared_filenames": [ - "tests/models/mamba2/test_modeling_mamba2.py" - ], - "shared_file_areas": [ - { - "filename": "tests/models/mamba2/test_modeling_mamba2.py", - "left_ranges": [ - [ - 418, - 424 - ], - [ - 446, - 451 - ], - [ - 475, - 480 - ], - [ - 498, - 504 - ] - ], - "right_ranges": [ - [ - 418, - 424 - ], - [ - 450, - 456 - ], - [ - 480, - 497 - ], - [ - 508, - 514 - ] - ] - } - ] - }, - { - "left_pr_number": 46951, - "right_pr_number": 46970, - "code_similarity": 0.085, - "size_similarity": 0.423, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46951, - "right_pr_number": 46971, - "code_similarity": 0.106, - "size_similarity": 0.5, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.038, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46970, - "right_pr_number": 46971, - "code_similarity": 0.169, - "size_similarity": 0.846, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -2276,31 +2048,113 @@ ] }, { - "cluster_id": "cluster-46957-2", + "cluster_id": "cluster-46831-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, - "canonical_pr_number": 46957, + "canonical_pr_number": 46831, "issue_numbers": [], "pr_numbers": [ - 46957, - 46968 + 46831, + 46978 ], "evidence_types": [ "soft_similarity" ], "pr_comparisons": [ { - "left_pr_number": 46957, - "right_pr_number": 46968, - "code_similarity": 0.131, - "size_similarity": 0.608, - "file_overlap": 0.0, + "left_pr_number": 46831, + "right_pr_number": 46978, + "code_similarity": 0.113, + "size_similarity": 0.229, + "file_overlap": 0.182, "area_overlap": 0.0, - "patch_similarity": 0.063, - "shared_filenames": [], - "shared_file_areas": [] + "patch_similarity": 0.083, + "shared_filenames": [ + "src/transformers/models/qwen3_5/modeling_qwen3_5.py", + "src/transformers/models/qwen3_5_moe/modeling_qwen3_5_moe.py", + "src/transformers/models/qwen3_next/modeling_qwen3_next.py", + "src/transformers/models/qwen3_next/modular_qwen3_next.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/qwen3_5/modeling_qwen3_5.py", + "left_ranges": [ + [ + 466, + 474 + ] + ], + "right_ranges": [ + [ + 33, + 39 + ], + [ + 437, + 443 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_5_moe/modeling_qwen3_5_moe.py", + "left_ranges": [ + [ + 463, + 471 + ] + ], + "right_ranges": [ + [ + 33, + 39 + ], + [ + 434, + 440 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_next/modeling_qwen3_next.py", + "left_ranges": [ + [ + 620, + 628 + ] + ], + "right_ranges": [ + [ + 30, + 36 + ], + [ + 592, + 598 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_next/modular_qwen3_next.py", + "left_ranges": [ + [ + 461, + 469 + ] + ], + "right_ranges": [ + [ + 23, + 29 + ], + [ + 433, + 439 + ] + ] + } + ] } ] }, @@ -2588,7 +2442,7 @@ 46897, 46901 ], - "reason": "Issues in cluster-46894-6 are treated as duplicates because they share closing_reference, soft_similarity evidence." + "reason": "Issues in cluster-46894-6 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." } ], "duplicate_prs": [ @@ -2681,6 +2535,15 @@ "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-46831-2", + "canonical_pr_number": 46831, + "duplicate_pr_numbers": [ + 46978 + ], + "target_issue_number": null, + "reason": "PRs in cluster-46831-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-46869-3", "canonical_pr_number": 46886, @@ -2698,29 +2561,7 @@ 46958 ], "target_issue_number": 46894, - "reason": "PRs in cluster-46894-6 are treated as duplicates because they converge on issue #46894 with closing_reference, soft_similarity evidence." - }, - { - "cluster_id": "cluster-46927-6", - "canonical_pr_number": 46927, - "duplicate_pr_numbers": [ - 46949, - 46950, - 46951, - 46970, - 46971 - ], - "target_issue_number": null, - "reason": "PRs in cluster-46927-6 are treated as duplicates because they share soft_similarity evidence." - }, - { - "cluster_id": "cluster-46957-2", - "canonical_pr_number": 46957, - "duplicate_pr_numbers": [ - 46968 - ], - "target_issue_number": null, - "reason": "PRs in cluster-46957-2 are treated as duplicates because they share soft_similarity evidence." + "reason": "PRs in cluster-46894-6 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." } ], "best_issue": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 186c9bdcbe5403dbc92e3a5fc54f834957230785..ebcf79bc1af6f3938eee2ef95556ff4b1107e696 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": "20260701T000029Z", - "generated_at": "2026-07-01T00:11:01Z", + "snapshot_id": "20260701T060031Z", + "generated_at": "2026-07-01T06:12:27Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -132,418 +132,418 @@ "issue:45237|issue:46015", "issue:43039|issue:44805" ], - "prepared_review_unit_hash": "d73b86881442560dcf92f363c754708dfaace06f9acf3993cdde9ea1d6d9d6cb", + "prepared_review_unit_hash": "2f811f2e920d66a277e7a03e2286cf312b291b2d0cacb07e93d3335e6e4ca450", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 63, - "serialized_chars": 46569, - "estimated_input_tokens": 11643, - "estimated_eval_tokens": 23542 + "serialized_chars": 46580, + "estimated_input_tokens": 11645, + "estimated_eval_tokens": 23546 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue reports with a few genuine duplicate subclusters: DeepSpeed ZeRO-3/BERT index errors, Qwen3.5 MoE rotary-embedding/offload bugs, and a torch 2.4.1 compatibility report. The rest are broad-similarity false positives across different models, backends, or docs topics.", - "confidence": 0.76, - "canonical_issue_reason": "issue:43596 is the clearest representative of the DeepSpeed ZeRO-3/BERT index-out-of-bounds duplicate cluster and the earliest concise report of that concrete failure mode.", + "summary": "This set is mostly heterogeneous. I would only merge two soft edges as true duplicates: the Bert/DeepSpeed ZeRO-3 IndexError pair and the Qwen3.5 MoE rotary-embedding/offload pair. The remaining pairs are related by broad subsystem or wording, but not the same concrete bug/change.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43596 is the best canonical anchor for the ZeRO-3/Bert IndexError subcluster: it states the concrete failure mode clearly, and issue:43638 reads like the same bug with a slightly narrower phrasing.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 best captures a real duplicate cluster with a specific error, setup, and likely shared root cause; it is the most useful single issue to keep as the representative report.", + "best_issue_reason": "issue:43596 is the strongest representative of the only clearly duplicated issue pair and is the most actionable target for de-duplication.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Same family (Qwen3.5) but different failures: config attribute error vs streaming chat/completions behavior." + "reason": "Different failures in the Qwen3.5 family: missing config attribute vs streaming chat/completions behavior." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different hardware/backend problems: AMD GPU forward failure vs MPS histogram not implemented." + "reason": "Both mention hardware/backend problems, but one is GPT-OSS on AMD GPUs and the other is an MPS histogram op error; not the same bug." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different model paths and symptoms; image-preprocessor dtype mismatch is not the same as GPT-OSS AMD GPU failure." + "reason": "Different model-specific failures: Mistral image/text dtype mismatch vs GPT-OSS GPU backend issue." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are dtype-related, but they affect different models and code paths; too broad to merge." + "reason": "Both are dtype/device related, but they affect different models and code paths." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Different models and failures; an example breakage is not the same bug as an AMD GPU runtime issue." + "reason": "Unrelated model/example failures; no shared concrete failure mode." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different concrete problems in different MoE models; not the same code-path bug." + "reason": "Different MoE/model problems: AMD GPU execution vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collectives mismatch and ONNX export error are unrelated failures." + "reason": "DeepSpeed collective mismatch is unrelated to ModernBERT ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Both mention model/runtime issues, but ONNX export and Qwen3Moe FP8 failures are different bugs." + "reason": "ONNX export error and FineGrainedFP8 failure are separate model/tooling issues." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different models and failure modes; export error vs device-map cuda assert." + "reason": "Export-time ONNX failure and auto device-map CUDA assert are different code paths." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "A task crash and outdated RAG examples are unrelated." + "reason": "Table QA crash vs RAG documentation freshness; not the same bug." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export failure and GenerationMixin import error are distinct issues." + "reason": "ONNX export issue and GenerationMixin import error are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanups for v5 removals, but they reference different removed surfaces and are not the same change." + "reason": "Both are v5 doc cleanups, but they remove different references and would not be one concrete fix." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "Different subsystems: TRL embedding unpacking vs resuming locally downloaded remote-code models." + "reason": "TRL SFT embedding unpacking failure is unrelated to resuming training with trust-remote-code." }, { "left": "issue:46165", "right": "issue:46869", "accept": false, - "reason": "Resume/trust-remote-code problem is unrelated to YOLO multi-GPU device-map failure." + "reason": "Resume/trust-remote-code bug is unrelated to YOLO multi-GPU device-map failure." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Both are device/shape-ish symptoms, but one is a mask shape mismatch and the other is a device mismatch in Gemma4 PLE." + "reason": "Mask-shape IndexError and Gemma4 PLE device mismatch are different forward-pass errors." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both involve tensor shape errors, but the reported setups and concrete failures differ." + "reason": "Both are IndexErrors, but one is ZeRO-3/Bert initialization and the other is a mask/tensor shape mismatch." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "DeepSpeed ZeRO-3 BERT loading and remote-code resume are unrelated." + "reason": "ZeRO-3 Bert initialization bug is unrelated to resume/trust-remote-code behavior." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different error signatures and code paths; index-out-of-bounds in ZeRO-3 vs mask/index mismatch." + "reason": "Different IndexError families: ZeRO-3 empty tensor initialization vs mask/tensor shape mismatch." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask shape mismatch and video token/feature mismatch are different multimodal bugs." + "reason": "Different multimodal failures: mask indexing vs video token/feature mismatch in chat template." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Unrelated failures: tensor mask mismatch vs GPT-OSS TP/MoE issue." + "reason": "Mask shape mismatch and GPT-OSS tensor-parallel failure are different problems." }, { "left": "issue:46165", "right": "issue:46468", "accept": false, - "reason": "Remote-code resume bug and broken official Nemotron example are not the same underlying issue." + "reason": "Training resume issue and Nemotron example breakage do not share a concrete failure mode." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration and resume/trust-remote-code are different failures." + "reason": "Video StopIteration in Qwen3-vl-embedding is unrelated to training resume with remote code." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset and ModernBERT ONNX export are unrelated." + "reason": "BarkProcessor voice preset bug and ModernBERT ONNX export are unrelated." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both mention DeepSpeed, but collective mismatch and BERT ZeRO-3 index errors are different bugs." + "reason": "DeepSpeed collective mismatch is not the same as ZeRO-3 Bert initialization IndexError." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same DeepSpeed ZeRO-3/BERT index-out-of-bounds failure; the titles differ only in whether it happens at init or training and pretrained vs non-pretrained BERT." + "reason": "Same underlying ZeRO-3/Bert out-of-bounds IndexError during model initialization/loading; the titles describe the same failure with minor wording differences." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy call and GPT-OSS tensor-parallel failure are not the same bug." + "reason": "Liger Kernel cross-entropy call issue is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Incorrect detokenization and LoRA merge/save mismatch are unrelated." + "reason": "Different defects: DeepSeek detokenization in v5 vs LoRA merge/save/reload mismatch." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen-3 NVFP4 runtime error and LoRA reload mismatch are different problems." + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA merge/save inconsistency are unrelated." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Processor failure for Voxtral and remote-code model_class config_class error are distinct." + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class/config_class attribute errors." }, { "left": "issue:43716", "right": "issue:46869", "accept": false, - "reason": "Different models and runtimes; image-preprocessor dtype mismatch is not a YOLO device-map bug." + "reason": "Mistral dtype mismatch and YOLO multi-GPU device-map failure are different model/back-end bugs." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "FSDP checkpointing recomputation mismatch is unrelated to a mask shape/indexing error." + "reason": "FSDP activation-checkpointing tensor recompute error is unrelated to the mask indexing bug." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Both involve DeepSpeed, but the specific runtime errors and reproduction paths differ." + "reason": "DeepSpeed training weight-shape error and DeepSpeed collective mismatch are different failures." }, { "left": "issue:45685", "right": "issue:46869", "accept": false, - "reason": "MPS histogram op issue and YOLO multi-GPU device-map failure are unrelated." + "reason": "MPS histogram op error is unrelated to YOLO multi-GPU device-map behavior." }, { "left": "issue:45859", "right": "issue:45902", "accept": true, - "reason": "Same Qwen3_5MoeTextRotaryEmbedding code path and same underlying offload/materialization bug around rotary frequencies." + "reason": "Same Qwen3.5 MoE rotary-embedding/device-materialization bug: one report says CPU offload incompatibility, the other identifies uninitialized inv_freq after meta\u2192to_empty(cuda)." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "BERT ZeRO-3 index-out-of-bounds and GPT-OSS TP failure are unrelated." + "reason": "ZeRO-3 Bert IndexError and GPT-OSS TP failure are unrelated." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation checkpointing/FSDP recompute mismatch and trainer signature change are different issues." + "reason": "Activation-checkpointing/FSDP recomputation mismatch is unrelated to CustomTrainer.compute_loss signature breakage." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "DeepSpeed BERT loading bug and GPT-OSS tensor-parallel bug do not share the same code path." + "reason": "ZeRO-3 Bert initialization bug is unrelated to GPT-OSS TP failure." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different models and symptoms; dtype mismatch in preprocessing is not an invalid method call in GraniteMoEHybrid." + "reason": "Different model-specific issues: Mistral dtype mismatch vs GraniteMoEHybrid invalid method." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Different multimodal bugs: reverse embedding lookup vs video feature/token mismatch." + "reason": "Both are Gemma4/Gemma-related multimodal issues, but one is an expensive reverse-embedding lookup and the other is a video token/feature mismatch." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model families and backend failures; NVFP4 runtime error is not the GPT-OSS AMD GPU issue." + "reason": "Qwen-3-8B-NVFP4 runtime failure and GPT-OSS AMD GPU failure are different." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "vLLM Gemma3TextConfig attribute error and Mistral tokenizer regex patch crash are unrelated." + "reason": "Gemma3 config attribute missing for vLLM and Mistral tokenizer regex patch crash are unrelated." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Gemma4 device mismatch and expensive reverse embedding lookup are different bugs." + "reason": "Different Gemma4 bugs: device mismatch in forward vs costly reverse embedding lookup." }, { "left": "issue:45800", "right": "issue:46515", - "accept": true, - "reason": "Both report the same torch 2.4.1 incompatibility; one is just a more specific version of the same compatibility break." + "accept": false, + "reason": "Both are torch 2.4.1 incompatibility reports, but they are not clearly the same concrete failure and may target different stacks/versions." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP activation checkpointing mismatch and Qwen3 auto device-map cuda assert are unrelated." + "reason": "FSDP activation-checkpointing issue and Qwen3 auto device-map CUDA assert are different problems." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Generic torch incompatibility and Qwen-3 NVFP4 runtime error are different issues." + "reason": "General torch incompatibility and Qwen-3-8B-NVFP4 runtime failure are not the same bug." }, { "left": "issue:43828", - "right": "issue:45685", + "right": "issue:45161", "accept": false, - "reason": "Different backends and failures: autocast dtype mismatch vs MPS integer histogram support." + "reason": "Phi-tiny-MoE autocast dtype mismatch and GPT-OSS TP failure are different issues." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Mixtral torch dynamo tracing error and Mistral-3 image dtype mismatch are unrelated." + "reason": "Mixtral torch-dynamo grouped_mm error and Mistral dtype mismatch are different code paths." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Both are runtime failures, but the models and concrete errors differ." + "reason": "Different model/runtime failures; only a broad 'inference issue' similarity." }, { "left": "issue:43716", "right": "issue:46468", "accept": false, - "reason": "Mistral-3 image dtype mismatch and Nemotron example breakage are unrelated." + "reason": "Mistral dtype mismatch is unrelated to Nemotron example breakage." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Different models and failure modes; NVFP4 runtime error is not the TRL embedding unpacking issue." + "reason": "Qwen-3-8B-NVFP4 error and TRL SFT embedding unpacking are unrelated." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "Invalid method call in GraniteMoEHybrid and MPS histogram Int support are unrelated." + "reason": "GraniteMoEHybrid invalid method call and MPS histogram op error are unrelated." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different model families and error signatures." + "reason": "Mistral dtype mismatch and Qwen-3-8B-NVFP4 runtime error are distinct." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "TRL SFT embedding unpacking and GPT-OSS TP failure are different bugs." + "reason": "TRL SFT embedding unpacking failure is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:43716", "right": "issue:45685", "accept": false, - "reason": "Different models, devices, and errors; not the same underlying bug." + "reason": "Mistral dtype mismatch and MPS histogram op error are unrelated." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger Kernel cross-entropy being called and TRL embedding unpacking are unrelated." + "reason": "Liger Kernel cross-entropy issue is unrelated to TRL SFT embedding unpacking." }, { "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Gemma3 ZeRO-3 loading and BERT ZeRO-3 training index errors are similar only at a high level; the concrete models and code paths differ." + "reason": "Gemma3 ZeRO-3 loading IndexError and Bert ZeRO-3 loading IndexError are different model families." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions in v5, but they target different removed documentation surfaces." + "reason": "Both are docs regressions in v5, but one is about pipeline-task docs and the other about the run command; separate fixes." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Repeated pair: autocast dtype mismatch and MPS histogram error are not the same bug." + "reason": "Phi-tiny-MoE autocast dtype mismatch and MPS histogram op error are unrelated." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Liger Kernel cross-entropy call and Phi-tiny-MoE autocast dtype mismatch are different failures." + "reason": "Liger Kernel cross-entropy path and Phi-tiny-MoE autocast mismatch are different issues." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and custom trainer loss signature change are unrelated." + "reason": "DeepSpeed collective mismatch and trainer loss-signature breakage are unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Activation checkpointing/FSDP recompute mismatch and GenerationMixin import error are unrelated." + "reason": "FSDP activation-checkpointing mismatch and GenerationMixin import error are unrelated." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "GPT-OSS AMD GPU failure and outdated RAG examples are not the same issue." + "reason": "GPT-OSS AMD GPU failure and RAG docs issue are unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross-entropy invocation and mask shape mismatch are unrelated." + "reason": "Liger Kernel cross-entropy call and mask shape mismatch are different failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "The clustering summary is grounded in the packet: the three cited subclusters are supported by the issue titles, and the non-duplicate judgments are conservative. The accepted pairs also look defensible from the titles alone (especially the two DeepSpeed ZeRO-3/BERT index errors and the torch 2.4.1 incompatibility pair)." + "feedback": "The summary is grounded in the packet and the soft-edge calls stay conservative. The two accepted pairs are well supported by the titles, and the remaining rejections avoid overclaiming duplicate status." }, "evaluator_used": true, "retried": false, @@ -645,8 +645,8 @@ "issue:43541|issue:46869", "issue:44292|issue:45901", "issue:39290|issue:45823", - "issue:45464|issue:45685", "issue:36010|issue:39290", + "issue:45464|issue:45685", "issue:43541|issue:45685", "issue:43541|issue:43828", "issue:44841|issue:45084", @@ -657,190 +657,190 @@ "issue:43572|issue:45070", "issue:45800|issue:45823" ], - "prepared_review_unit_hash": "d17e5b7d5f8ec6cb06d218b895270b3cbaa6efc0d7e353bd6432b235cf6afba2", + "prepared_review_unit_hash": "8cf457eb55708644dd316cc1c5fc21b8c60d1dd6ff4e5336dfa29679d6b15cf5", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 42, - "serialized_chars": 40046, - "estimated_input_tokens": 10012, - "estimated_eval_tokens": 20280 + "serialized_chars": 40043, + "estimated_input_tokens": 10011, + "estimated_eval_tokens": 20278 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. I only see two true duplicate-style matches: one ZeRO-3 model-loading IndexError and one broad AutoTokenizer token-ID regression. The remaining pairs are related at a subsystem level but appear to be distinct bugs or changes.", - "confidence": 0.76, - "canonical_issue_reason": "issue:45920 is the best canonical issue: it describes the broad v5 AutoTokenizer token-ID regression across multiple model families, which is the most central and representative bug in this set.", + "summary": "This set is mostly a noisy mix of unrelated issue reports. I only see two believable duplicate pairs: the v5 AutoTokenizer wrong-token-id regression, and the DeepSpeed ZeRO-3 index-out-of-bounds loading failure. The rest look like separate bugs, model-specific failures, or docs/version-compatibility issues rather than the same underlying change.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 45920 is the best canonical anchor for the clearest duplicate subcluster: it describes the v5 AutoTokenizer token-ID regression broadly across multiple model families, so it subsumes the narrower Granite-only report (45812) better than the reverse.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the strongest overall representative because it is broad, well-scoped, and explicitly frames a cross-model regression rather than a model-specific failure.", + "best_issue_reason": "45920 is the strongest representative issue in this set because it captures the widest shared failure mode with explicit regression framing and multiple affected models. The zero3 pair is separate and should not be merged into the same canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Both are model-loading failures, but they affect different models and likely different root causes (corrupted checkpoint vs missing text-encoder weights)." + "reason": "Both are model loading/weight issues, but one is a corrupted checkpoint claim and the other is missing text-encoder weights; different model families and likely different root causes." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Both mention version incompatibility, but one is torch/transformers 5.8.0 and the other is a specific v5.9.0 distributed build break; not the same bug." + "reason": "Both mention version compatibility, but one is a torch 2.4.1 / transformers 5.8.0 incompatibility and the other is a transformers 5.9.0 distributed-build regression; not the same bug." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "Unrelated failures: vLLM/PaddleOCR-VL load problem versus a PyTorch USE_DISTRIBUTED=0 regression." + "reason": "PaddleOCR-VL loading in vLLM and USE_DISTRIBUTED=0 support regression are unrelated failure modes with different stacks and symptoms." }, { "left": "issue:41762", "right": "issue:43596", "accept": true, - "reason": "Same underlying ZeRO-3 loading failure pattern: an IndexError from empty-shard/model init during checkpoint loading." + "reason": "These describe the same ZeRO-3 out-of-bounds index failure during model loading/init; the model names differ, but the underlying bug appears to be the shared zero-sized tensor/indexing path." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Different models and failure modes: a missing Gemma3 config attribute versus a Qwen3 auto-device-map CUDA assert." + "reason": "Gemma3TextConfig missing sliding_window_pattern in vLLM and Qwen3 auto-device-map cudaErrorAssert are different model/runtime failures." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are docs issues, but they concern different documentation remnants and are not the same change." + "reason": "One is a pipeline docs mismatch and the other is outdated docs about a removed command; same documentation area, but not the same issue." }, { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Storage-type lookup error versus distributed-support breakage; no shared concrete code path." + "reason": "Float8 storage lookup failure and distributed support breakage are separate runtime incompatibilities." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "Completely different areas: AMD GPU support for GPT-OSS versus table-question-answering crashes." + "reason": "AMD GPU execution failure for GPT-OSS-20B is unrelated to the table-question-answering crash." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Different bug classes: detokenization regression versus outdated RAG examples." + "reason": "Incorrect detokenization and outdated RAG examples are different problem types with no shared code path." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross-entropy dispatch and Qwen3-VL video StopIteration are unrelated." + "reason": "Liger Kernel cross_entropy dispatch and Qwen3-vl video StopIteration are unrelated issues." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Both involve model/input handling, but one is device-map mismatch and the other is video feature/token mismatch; different problems." + "reason": "Device-mapping mismatch in Gemma4 PLE and Qwen3.5 video token/feature mismatch are different bugs." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Different model families and failures: processor bug versus device-mapping mismatch." + "reason": "Processor failure for Voxtral and Gemma4 PLE device mismatch are not the same underlying defect." }, { "left": "issue:43866", "right": "issue:46772", "accept": false, - "reason": "Both are model compatibility issues, but corrupted Ovis2 checkpoint and DiffusionGemma ModelOpt compatibility are distinct." + "reason": "Corrupted Ovis2 checkpoint and DiffusionGemma ModelOpt compatibility are unrelated." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and RAG docs are unrelated." + "reason": "NemotronH checkpoint loading and outdated RAG docs are unrelated." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Both are regression reports, but one targets remote_code models and the other PyTorch distributed support; not the same bug." + "reason": "Remote-code breakage in recent transformers and distributed-build support regression are different compatibility problems." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same tokenizer regression: both report wrong token IDs, with 45920 being the broader multi-model description." + "reason": "These are the same AutoTokenizer wrong-token-ID regression in v5; 45920 is just the broader multi-model report." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Both are load failures, but they involve different models and different failure causes." + "reason": "Corrupted checkpoint versus inability to load checkpoints are superficially similar but different models and failure mechanisms." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen NVFP4 runtime failure and RAG documentation issues are unrelated." + "reason": "Qwen-3-8B-NVFP4 runtime error and outdated RAG examples have no shared code path." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes Int8Params compatibility is a different code path from the distributed-support regression." + "reason": "bitsandbytes Int8Params constructor incompatibility is unrelated to the distributed support regression." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window, but one is a missing config field for Gemma3 and the other is a Qwen3-MoE sliding_window issue." + "reason": "Both mention sliding_window, but one is a missing config attribute in vLLM and the other is a Qwen3-MoE model issue; not clearly the same bug." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexErrors, but the reported shapes and affected setup differ; not clearly the same defect." + "reason": "Both are index errors under DeepSpeed ZeRO-3, but the reported failures are different (mask shape mismatch vs zero-sized Bert init path)." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper loading/running failure and table-question-answering crashes are unrelated." + "reason": "Whisper-large-v2 failing to run and table-question-answering crashing are unrelated issues." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Different model-specific load problems: GLM-4.7-Flash versus Ovis2 checkpoint corruption." + "reason": "GLM-4.7-Flash unit-test loading failure and Ovis2 corrupted checkpoint are different model-loading problems." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version incompatibility and a pydantic PretrainedConfig field regression are separate issues." + "reason": "Torch 2.10 compatibility and pydantic PretrainedConfig field breakage are unrelated versioning issues." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Detokenization regression and table-question-answering crash do not share a concrete bug." + "reason": "Incorrect detokenization and table-question-answering crash are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5/chat, but one is a general chat crash and the other is a streaming inference API failure." + "reason": "Qwen3.5 chat crash and streaming chat/completions API failure are different Qwen-related symptoms with different surfaces." }, { "left": "issue:43792", @@ -852,121 +852,97 @@ "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Different failure classes: ZeRO-3 index error versus torch/transformers version incompatibility." + "reason": "ZeRO-3 index failure and torch/transformers incompatibility are different bug classes." }, { "left": "issue:43541", "right": "issue:46869", "accept": false, - "reason": "Grouped_mm tracing failure for Mixtral and multi-GPU device-map failure for YOLO are unrelated." + "reason": "Mixtral grouped_mm tracing error and YOLO multi-GPU device-map failure are unrelated model/runtime issues." }, { "left": "issue:44292", - "right": "issue:46015", + "right": "issue:45901", "accept": false, - "reason": "Duplicate of an unrelated pair: NVFP4 runtime failure does not match RAG docs." + "reason": "Qwen NVFP4 execution error and table QA crash do not share an underlying code path." }, { - "left": "issue:43872", - "right": "issue:46304", + "left": "issue:39290", + "right": "issue:45823", "accept": false, - "reason": "Same as above: bitsandbytes Int8Params issue is not the distributed-support regression." + "reason": "Gemma3 sliding_window_pattern missing under vLLM and Gemma4 device mismatch during forward are different failures." }, { - "left": "issue:39290", - "right": "issue:43531", + "left": "issue:36010", + "right": "issue:39290", "accept": false, - "reason": "Repeated sliding_window-themed issues, but the models and root causes differ." + "reason": "GenerationMixin import error and Gemma3 sliding_window_pattern error are unrelated API breakages." }, { - "left": "issue:41093", - "right": "issue:41762", + "left": "issue:45464", + "right": "issue:45685", "accept": false, - "reason": "Both are loading-time errors, but one is a mask shape mismatch and the other is a ZeRO-3 empty-shard IndexError." + "reason": "Streaming chat/completions failure and MPS histogram_not_implemented error are different runtime issues." }, { - "left": "issue:43792", - "right": "issue:45070", + "left": "issue:43541", + "right": "issue:45685", "accept": false, - "reason": "Repeated unrelated pair: Whisper runtime failure is not a PretrainedConfig/pydantic regression." + "reason": "Mixtral grouped_mm tracing and MPS histogram failure are unrelated." }, { - "left": "issue:43854", - "right": "issue:43866", + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different code paths and symptoms." + }, + { + "left": "issue:44841", + "right": "issue:45084", "accept": false, - "reason": "Repeated unrelated pair: GLM-4.7-Flash load failure is not Ovis2 checkpoint corruption." + "reason": "Voxtral processor failure and non-template-node compile error are unrelated." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "lm_head weight tying in Mistral3 is a different bug from Ovis2 checkpoint corruption." + "reason": "Mistral3 weight-tying bug and Ovis2 checkpoint corruption are different model-specific loading problems." }, { "left": "issue:45507", "right": "issue:46869", "accept": false, - "reason": "GraniteMoEHybrid invalid-method bug and YOLO device-map failure are different code paths." + "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU device-map failure are separate issues." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Repeated pair; the errors are similar in being IndexErrors, but the concrete failure modes are different." + "reason": "Both are erroring under different loading scenarios, but the mask-shape mismatch and zero-sized index error are not the same defect." }, { "left": "issue:45399", "right": "issue:46661", "accept": false, - "reason": "Flash-attn fallback gating and remote-code config_class errors are distinct issues." + "reason": "Flash-attn fallback gating and remote-code config_class AttributeError are different compatibility checks." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and pydantic PretrainedConfig field breakage are different configuration issues." + "reason": "StableLmConfig pad_token_idx regression and pydantic PretrainedConfig field breakage are unrelated." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers compatibility and Gemma4 device-mapping mismatch are unrelated." - }, - { - "left": "issue:43828", - "right": "issue:45800", - "accept": false, - "reason": "Autocast dtype mismatch and version incompatibility are different failure modes." - }, - { - "left": "issue:36010", - "right": "issue:39290", - "accept": false, - "reason": "Importing GenerationMixin and Gemma3TextConfig sliding_window_pattern are unrelated." - }, - { - "left": "issue:43541", - "right": "issue:45685", - "accept": false, - "reason": "Torch Dynamo grouped_mm tracing and an MPS histogram kernel missing for Int are separate backend bugs." - }, - { - "left": "issue:43541", - "right": "issue:43828", - "accept": false, - "reason": "Both are dtype/backend-sensitive, but one is a grouped_mm tracing error and the other is an autocast mismatch; not the same underlying bug." - }, - { - "left": "issue:44841", - "right": "issue:45084", - "accept": false, - "reason": "Processor loading for Voxtral and a template-node compilation error are unrelated." + "reason": "Torch/transformers version incompatibility and Gemma4 device mismatch are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and the conservative stance on soft-edge pairs is appropriate. The two accepted matches are well-supported by the titles and repeated failure pattern, and the canonical issue choice is reasonable for the broad v5 AutoTokenizer regression." + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The two accepted soft pairs are well supported by the issue titles and descriptions: the v5 AutoTokenizer token-ID regression pair and the DeepSpeed ZeRO-3 out-of-bounds loading failures. The other soft-edge rejections are cautious and do not overclaim duplication from superficial similarity alone." }, "evaluator_used": true, "retried": false, @@ -1062,233 +1038,233 @@ "issue:43404|issue:45005", "issue:42915|issue:45399", "issue:43873|issue:45005", - "issue:44530|issue:45005", "issue:44360|issue:44512", + "issue:44530|issue:45005", "issue:44492|issue:44512", "issue:44863|issue:45005", "issue:45183|issue:46463", "issue:43975|issue:44292" ], - "prepared_review_unit_hash": "7c0f0ef2b6434300b2cacd612b3c3c70ce3e76c9b8a5f9d3ac0f1e1f51fbc894", + "prepared_review_unit_hash": "e6e882216b63a48d342786db6f5852ef03b02bc125eaa543fd9c0a6e1d8f509d", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36796, - "estimated_input_tokens": 9199, - "estimated_eval_tokens": 18654 + "serialized_chars": 36792, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous, with no single obvious cluster-wide canonical issue. I only see a few plausible duplicate-like pairs around the same underlying bug, especially Qwen video-input regressions, trust_remote_code custom-module loading, and input-validation error messages.", + "summary": "This set is mostly a grab bag of unrelated transformers issues, with a few real duplicate pockets around input validation, Qwen video handling, and SigLIP2 regressions.", "confidence": 0.71, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 45183 is the best umbrella representative: it describes the generic input_ids/inputs_embeds validation bug that the T5-specific report (46463) appears to instantiate.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "45183 is the broadest, most reusable issue in the set and captures a concrete validation-path bug rather than a model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different failures: PaddleOCR-VL load failure vs GGUF gpt-oss architecture support." + "reason": "Different failures: one is a model load problem, the other is GGUF architecture support for gpt-oss." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both are shape-related, but one is a generic mask/tensor mismatch and the other is a Qwen3.5/TRL embedding unpacking issue." + "reason": "Both mention indexing problems, but the titles point to different models/code paths and the same root cause is not clear." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Different code paths: torch dynamo grouped_mm tracing vs tensor-parallel support for GPT-OSS MoE." + "reason": "Both are MoE-related, but one is a dynamo grouped_mm tracing failure and the other is a tensor-parallel support issue." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Unrelated bugs: int4 OOM regression vs tied-weights behavior on translation models." + "reason": "One is an int4 quantization OOM regression; the other is a tied-weights issue on translation models." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Different symptoms and subsystems: autocast dtype mismatch vs streaming chat/completions failure." + "reason": "Different modalities and symptoms: autocast dtype mismatch versus streaming chat/completions failure." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "Architecture support for gpt-oss GGUF is unrelated to Qwen3.5 streaming inference." + "reason": "No shared bug path; gpt-oss GGUF support and Qwen streaming inference are unrelated." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Different breakages: init_empty_weights argument handling vs pydantic PretrainedConfig field regression." + "reason": "One is about init_empty_weights argument handling, the other about a pydantic PretrainedConfig field regression." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is not the same bug as the pydantic field regression." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the pydantic model breakage." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but the reported code paths and triggers are different." + "reason": "Both are index errors, but the specific failure modes and contexts differ too much to treat as the same bug." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Different modalities and failure modes: image preprocessor dtype mismatch vs video chat-template token/feature mismatch." + "reason": "Image-preprocessor dtype mismatch and video token/feature mismatch are different multimodal bugs." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code compatibility regressions are too broad here; one is model breakage, the other a torch/transformers version incompatibility." + "reason": "Remote_code breakage and torch/version incompatibility are too broad and not the same concrete failure." }, { "left": "issue:43054", "right": "issue:43994", - "accept": false, - "reason": "Both mention SigLIP2, but one is quality/perplexity style feedback and the other is a pipeline/AutoModel correctness issue." + "accept": true, + "reason": "Same SigLIP2 regression: both reports describe broken/nonsensical outputs from the same model family and likely the same text-embedding path." }, { "left": "issue:44479", "right": "issue:46294", "accept": true, - "reason": "Same Qwen3.5/Qwen video-input regression family; 46294 looks like a specific instance of the broader video token/feature mismatch bug." + "reason": "46294 is a specific instance of the broader Qwen video-input regression described in 44479, with the same video feature/token mismatch symptom." }, { "left": "issue:46468", "right": "issue:46869", "accept": false, - "reason": "Nemotron example breakage and YOLO multi-GPU device-map failure are unrelated." + "reason": "Completely different areas: a Nemotron example issue versus YOLO multi-GPU device-map behavior." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Composite config unwrapping is a different issue from uninitialized rotary embedding memory after materialization." + "reason": "One is about unwrapping text_config for composite configs; the other is an initialization/memory-materialization bug in Qwen3_5MoeTextRotaryEmbedding." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Deepspeed ZeRO3/BertModel initialization and trust_remote_code resume behavior are different failures." + "reason": "Different remote-code/deepspeed resume problems; not the same underlying bug." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 fallback behavior is unrelated to remote-code model_class/config_class attribute errors." + "reason": "Flash Attention 2 support and remote-code model_class/config_class handling are unrelated." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Image-preprocessor dtype mismatch and Qwen3.5 embedding unpacking are different bugs." + "reason": "Different failures: dtype mismatch in a multimodal processor versus embedding unpacking in TRL SFT training." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP 2-D weight error is unrelated to Qwen3.5/TRL embedding unpacking." + "reason": "FSDP weight shape error and Qwen3.5 embedding unpacking are not the same code-path problem." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "High perplexity on gpt-oss-20b and GLM5 are not the same bug." + "reason": "High perplexity on gpt-oss and a GLM5 issue are unrelated model-specific reports." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom-model notebook initialization and flash-attn2 fallback checks are unrelated." + "reason": "Notebook custom-model initialization and flash-attn fallback gating are different regressions." }, { "left": "issue:45632", "right": "issue:45698", - "accept": true, - "reason": "Both point to trust_remote_code/from_pretrained resolving the wrong custom module due to cache/path handling." + "accept": false, + "reason": "Both touch trust_remote_code caching, but one is a leaf-directory cache collision while the other is wrong custom-module loading after save_pretrained." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Tied lm_head weights in Mistral3 conditional generation is unrelated to translation-model tied-weight issues in v5." + "reason": "A missing lm_head weight tie in Mistral3 is not the same bug as translation-model tied weights in v5." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3-MoE FP8 failure and flash-attn2 fallback gating are different subsystems." + "reason": "Qwen3Moe FP8 loading failure and flash-attn fallback gating are separate issues." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization offloading behavior is not the same as tied-weight regressions." + "reason": "Quantization/offloading behavior is not the same as tied-weights regressions." }, { - "left": "issue:44530", - "right": "issue:45005", + "left": "issue:44360", + "right": "issue:44512", "accept": false, - "reason": "PagedAttentionCache linear_attention crash and tied-weight translation-model bug are unrelated." + "reason": "A DSA indexer discussion is unrelated to a docs typo about the removed run command." }, { - "left": "issue:44360", - "right": "issue:44512", + "left": "issue:44530", + "right": "issue:45005", "accept": false, - "reason": "A DSA indexer discussion is unrelated to a docs-only regression." + "reason": "PagedAttentionCache linear_attention failure and tied-weights regressions are different concrete bugs." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache-strategy typo and removed run-command docs are separate documentation issues." + "reason": "A cache-strategy typo does not match a documentation command removal issue." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading and tied weights on translation models are different problems." + "reason": "NemotronH checkpoint loading and translation-model tied weights are unrelated." }, { "left": "issue:45183", "right": "issue:46463", "accept": true, - "reason": "Both are input-validation error-message bugs; 46463 looks like a T5-family manifestation of the same XOR/decoder-input validation logic." + "reason": "46463 is a specific model-family manifestation of the broader input_ids/inputs_embeds XOR validation bug described in 45183." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Incorrect detokenization in v5 and Qwen-3-8B-NVFP4 runtime failure are unrelated." + "reason": "Incorrect detokenization and an unexpected _is_hf_initialized argument error are different failure modes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the few accepted soft-edge pairs are reasonably conservative given the issue titles. The rejected pairs are also cautious rather than overassertive, and the overall claim that there is no single cluster-wide canonical issue is supported by the heterogeneous set." + "feedback": "The summary is grounded in the packet, and the accepted soft edges are conservative. The 45183/46463 duplicate call is well supported by the titles, and the other pair rejections avoid overstating weak similarity." }, "evaluator_used": true, "retried": false, @@ -1380,24 +1356,24 @@ "issue:45663|issue:45902", "issue:44403|issue:44451", "issue:43975|issue:45520", - "issue:41553|issue:43881", "issue:24643|issue:28530", "issue:42947|issue:43856", "issue:44403|issue:44488", "issue:43643|issue:45020", + "issue:41553|issue:43881", "issue:41762|issue:42915", "issue:44991|issue:45399", "issue:43782|issue:44960", "issue:43824|issue:43872" ], - "prepared_review_unit_hash": "ba588fc98e4024332f23d567d0ea7e432429542bd3e251ece7804bed9a991507", + "prepared_review_unit_hash": "714728aac67122dfa6f46e4a7dbc12c5b6d342ada7adf97e41098d58868bb8a3", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, @@ -1412,201 +1388,201 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a mix of unrelated transformer regressions; the strongest duplicate set is the Transformers v5 non-persistent-buffer corruption reports. The SigLIP2 pair is related in theme but not strong enough from titles alone to merge as a duplicate.", - "confidence": 0.92, - "canonical_issue_reason": "issue:43950 is the clearest, most detailed report of the Transformers 5.x non-persistent-buffer corruption regression and best represents that duplicate set.", + "summary": "This set is mostly heterogeneous issue reports, with only a few clear duplicate-style pairs. The strongest duplicate cluster is around the transformers 5.x non-persistent buffer regression; a Gemma4 quantized-embedder pair also looks identical. Most other pairs are related only at a broad subsystem level and should stay separate.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest, most concrete regression report in the set: a transformers 5.x non-persistent-buffer corruption bug with a very similar neighboring report (#44534) and another close buffer-related report (#43644).", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the best cluster representative because it names the shared regression precisely and covers the broadest manifestation of the bug.", + "best_issue_reason": "issue:43950 is the best overall representative because it is specific, actionable, and sits at the center of the only obvious duplicate-like issue cluster here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and failure modes: Qwen3.5 SFT unpacking vs GraniteMoEHybrid invalid method call." + "reason": "Different failures and code paths: Qwen3.5/TRL embedding unpacking vs a GraniteMoEHybrid invalid method call." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models with different concrete symptoms." + "reason": "Both are tokenizer regressions, but for different models and with different symptoms/root causes." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same Transformers v5 non-persistent-buffer corruption regression; both describe the junk/uninitialized buffer problem." + "reason": "Same underlying transformers 5.x non-persistent-buffer regression; both describe junk/corruption in buffer materialization/loading." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Doc-example model/processor mismatch and quantization failure is unrelated to SigLIP2 text-embedding quality." + "reason": "Docs-example model/processor mismatch and quantization failure is unrelated to a general SigLIP2 text-embedding quality issue." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Both describe the same v5 regression where non-persistent buffers are loaded with garbage/uninitialized data." + "reason": "Same underlying bug class: transformers 5.x mishandling of non-persistent buffers leading to junk/corruption." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Unexpected compute_loss kwarg and Gemma3 ZeRO-3 loading failure are different code paths." + "reason": "CustomTrainer API keyword breakage is unrelated to Gemma3 DeepSpeed ZeRO-3 loading failure." }, { "left": "issue:46899", "right": "issue:46932", "accept": true, - "reason": "Same Gemma4 quantized vision/audio embedder dtype-casting bug; the second issue explicitly says it is the same." + "reason": "These appear to be the same Gemma4 quantized vision/audio embedder dtype-cast bug; #46932 explicitly points to the same fix." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different incompatibilities: bitsandbytes `_is_hf_initialized` vs missing Float8 storage." + "reason": "Unrelated quantization/backend failures: bitsandbytes Int8Params kwargs vs missing Float8 storage lookup." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different symptoms and mechanisms: autocast dtype mismatch vs Qwen3.5 position_ids/flash-attn crash." + "reason": "Both involve dtype/precision problems, but on different models and different execution paths." }, { "left": "issue:45920", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer-related, but one is a broad multi-model token-ID regression and the other is DeepSeek-Coder-specific." + "reason": "A broad multi-model tokenizer regression is not the same as a DeepSeek-Coder v1-specific detokenization bug." }, { "left": "issue:45987", "right": "issue:46858", - "accept": true, - "reason": "Same cache.get_seq_length contract regression: a tensor is returned instead of an int, breaking downstream arithmetic." + "accept": false, + "reason": "Both mention get_seq_length, but one is a return-type contract issue and the other is a backend-specific tensor/device arithmetic failure." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality regression is unrelated to a generic tokenization-change report." + "reason": "Text-embedding quality regression is unrelated to a version-induced tokenization change." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Same model family, but zero probabilities and worse text embeddings are different bugs/symptoms." + "reason": "Unexpected zero probabilities and worse text embeddings are different symptoms with different implied causes." }, { "left": "issue:43950", "right": "issue:45902", - "accept": true, - "reason": "Same non-persistent-buffer initialization problem; 45902 is a specific manifestation on `inv_freq`." + "accept": false, + "reason": "Both are buffer-related, but one is a from_pretrained regression while the other is a specific meta\u2192to_empty(cuda) uninitialized-memory issue." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error handling for Voxtral is unrelated to Molmo missing tied-weights metadata." + "reason": "Bad error message loading Voxtral is unrelated to MolmoForCausalLM missing all_tied_weights_keys." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Related SigLIP2 topic, but the titles alone do not establish the same concrete bug; needs body-level confirmation." + "reason": "Model quality degradation is not the same as an HF-vs-JAX implementation discrepancy." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Flash-attn illegal access from 3D position_ids is unrelated to CPU-offload incompatibility in rotary embedding." + "reason": "Flash-Attention illegal-memory crash from 3D position_ids is unrelated to CPU-offload incompatibility in rotary embedding." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Both involve memory/perf, but gradient checkpointing ineffectiveness and int4 OOM are different problems." + "reason": "LoRA/gradient-checkpointing effectiveness and int4 quantization memory growth are different bugs." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states bug is a different code path from the Qwen3.5 rotary-embedding memory issue." + "reason": "Gemma4 FSDP2 shared_kv_states handling is unrelated to Qwen3_5MoeTextRotaryEmbedding buffer initialization." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise is not the same as a specific model load failure for ScandiBERT." + "reason": "Generic loading noise is not the same underlying bug as a specific ScandiBERT load failure." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "DeepSeek-Coder detokenization regression and a Python 3.13 flash_attn import error are unrelated." - }, - { - "left": "issue:41553", - "right": "issue:43881", - "accept": false, - "reason": "Different model-loading problems with no shared concrete failure mode." + "reason": "Tokenizer detokenization regression and a flash_attn import KeyError are unrelated." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "DeepSpeed weight shape error and early-stopping metric mismatch are unrelated training issues." + "reason": "DeepSpeed training shape error and early-stopping metric configuration are separate concerns." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "LoRA checkpointing inefficiency and Qwen3 MoE memory usage are different issues." + "reason": "Different memory/training issues in different models and setups; not the same concrete bug." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Generic loading noise is not the same as a model-specific tokenizer/load failure." + "reason": "Both involve loading, but one is generic noise and the other is a specific model load failure." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "`AutoConfig` missing fields with `trust_remote_code=True` is only one possible symptom of broader remote_code breakage, not the same concrete bug." + "reason": "Remote_code breakage is too broad to treat as the same bug as missing fields in AutoConfig.from_pretrained without stronger evidence." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different models and different failures: Voxtral tokenizer error messaging vs glm-4v-9b loading failure." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and Qwen3Moe FineGrainedFP8 failure are different model/config paths." + "reason": "Gemma3 ZeRO-3 load failure and Qwen3Moe FP8 config failure are unrelated." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure and flash-attn fallback gating are unrelated." + "reason": "Tokenizer loading failure and flash-attn fallback gating are separate issues." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only error and generic GLM5 issue are not the same bug." + "reason": "Qwen3VL weight-only loading error is unrelated to a GLM5 issue with no evidence of the same root cause." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Missing export/import symbol and bitsandbytes constructor incompatibility are different breakages." + "reason": "Importing a missing VL class and a bitsandbytes Int8Params incompatibility are unrelated." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, but one soft-edge verdict is too strong for the packet: issue:43054 vs issue:43493 is not clearly a duplicate from the titles alone. \u201ctext embedding of siglip2 is much worse than siglip\u201d and \u201cSigLIP2 discrepancy between HF implementation and original JAX implementation\u201d may be related, but the evidence here only supports a cautious \u2018possibly related\u2019 / \u2018needs body-level confirmation,\u2019 not a firm accept. Because the instructions require conservative soft-edge judgments, I can\u2019t accept this as written." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only true soft-edge pairs are the ones with strong title/body alignment around the non-persistent-buffer regression and the Gemma4 quantized-embedder dtype-cast bug; the rest are appropriately rejected as only loosely related. No overclaiming stands out." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1703,208 +1679,208 @@ "issue:44479|issue:44560", "issue:24643|issue:30064" ], - "prepared_review_unit_hash": "54a3116f450a24cf558af75570796ef2f6b9d123d093a9c1d2b2648a12a22cae", + "prepared_review_unit_hash": "c4c2ca8040072c029c9ce3aa63d5f63c1b35bcc641863424db22cba2a2c45fc8", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36752, - "estimated_input_tokens": 9188, - "estimated_eval_tokens": 18632 + "serialized_chars": 36753, + "estimated_input_tokens": 9189, + "estimated_eval_tokens": 18634 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is really several separate duplicate pairs, not one single underlying bug. Clear duplicates appear for the docs-removal notices, the `_is_hf_initialized` loading regression, the Qwen3 DeepSpeed weight-loading reports, the identical mask-shape IndexErrors, and the Florence2/CohereASR double-shift reports; most other pairs are only broad topical matches and should stay separate.", - "confidence": 0.79, - "canonical_issue_reason": "#43872 is the cleanest concrete bug report in the set: it names a specific incompatibility (`Int8Params.__new__` getting `_is_hf_initialized`) and has a near-identical follow-up report in #44291, making it a strong canonical issue for that subcluster.", + "summary": "Most of these items are only loosely related by subsystem; only a few pairs are true duplicates with the same concrete failure mode. The clearest duplicate clusters are the v5 docs cleanup, the `_is_hf_initialized`/`init_empty_weights` incompatibility, the mask-shape mismatch, and the training-loss double-shift bug pattern.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43872 is the best canonical anchor for the `_is_hf_initialized` / `init_empty_weights` incompatibility cluster because it names the concrete bitsandbytes failure and has the clearest external compatibility context.", "canonical_pr_reason": null, - "best_issue_reason": "#43872 is the best issue target overall because it is precise, regression-focused, and clearly duplicated by another report with the same failure signature; it is more actionable than the broader, model-specific reports.", + "best_issue_reason": "issue:44291 is the most actionable issue in that same cluster because it states the exact trigger (`init_empty_weights`) and the version scope (`transformers >= 5.0.0rc0`).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "Different model/runtime failures: MPS histogram dtype error vs Gemma4 device mismatch." + "reason": "Both are runtime device/backend problems, but they describe different code paths and symptoms." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Unrelated code paths: Mistral regex patch crash vs flash-attn fallback gating." + "reason": "Tokenizer regex crash vs flash-attn fallback gating are different bugs." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Broad remote_code breakage vs a specific Gemma4 forward device-mapping bug." + "reason": "Remote-code breakage and device-map mismatch are unrelated failures." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer mapping/new-model registration bug is distinct from remote_code model-loading regressions." + "reason": "Different problems in model/tokenizer registration and remote-code loading." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Generic GLM-4V loading failure is not the same as missing `all_tied_weights_keys` on Molmo." + "reason": "Different model-loading failures with different missing attributes and code paths." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Both involve Qwen/LoRA/weight tying, but they describe different symptoms and likely different fixes." + "reason": "Both mention Qwen/LoRA/tied embeddings, but one is a model usability regression and the other is a warning/flag issue, not the same bug." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Different mismatch types: autocast dtype issue vs device-map placement issue." + "reason": "Different failures: autocast dtype mismatch vs device placement mismatch." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote-code model breakage is unrelated to the Kimi-K2.5 tokenizer codec regression." + "reason": "Remote-code regression and tokenizer codec handling regression are separate issues." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but on different conversion directions and code paths." + "reason": "Both are tokenizer performance regressions, but they affect different conversion directions and likely different hot paths." }, { "left": "issue:45488", "right": "issue:46489", "accept": false, - "reason": "Both concern DeepSeek tokenizer regressions, but the reported root causes and model paths differ enough to keep separate." + "reason": "Both are tokenizer regressions, but the titles point to different underlying bugs and model families." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs bug: both say removed pipeline tasks are still referenced in documentation." + "reason": "Same docs cleanup: both refer to removed v5 pipeline tasks still mentioned in documentation." }, { "left": "issue:43872", "right": "issue:46869", "accept": false, - "reason": "Bitsandbytes/_is_hf_initialized incompatibility is unrelated to YOLO multi-GPU device-map failure." + "reason": "Bitsandbytes/init_empty_weights incompatibility is unrelated to YOLO multi-GPU device-map behavior." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Composite-config unwrapping bug is different from remote-code `config_class` lookup failure." + "reason": "Different auto-model loading issues: composite config handling vs remote-code config_class lookup." }, { "left": "issue:46894", "right": "issue:46897", "accept": true, - "reason": "Same training-loss double-shift bug pattern, reported on different models but the same underlying loss-path problem." + "reason": "Same training-loss double-shift pattern reported for two models; a single fix in the shared loss path could plausibly cover both." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Weight-tying bug in Mistral3 is unrelated to silent corruption of non-persistent buffers." + "reason": "Weight tying bug and non-persistent buffer corruption are different regression classes." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same regression signature: loading fails because `_is_hf_initialized` is passed into code that does not accept it." + "reason": "Same concrete incompatibility: loading with `init_empty_weights` passes unexpected `_is_hf_initialized` through the same code path." }, { "left": "issue:45020", "right": "issue:46661", "accept": false, - "reason": "General remote_code breakage and remote-code `config_class` handling are different problems." + "reason": "Both involve remote-code models, but the failures are different and would not necessarily share one fix." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Too little evidence that the generic Whisper failure and the fbank helper issue are the same concrete bug." + "reason": "Whisper model runtime failure and `_torch_extract_fbank_features()` issue are too broad to confirm as the same underlying bug from the titles alone." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both involve DeepSpeed/Qwen, but one is missing MoE weight conversion and the other is gradient averaging with GAS." + "reason": "Both mention DeepSpeed, but one is weight loading and the other is gradient averaging; different bugs." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing ineffective with LoRA is not the same as a missing `supports_gradient_checkpointing` flag on a trust-remote-code variant." + "reason": "Both touch gradient checkpointing, but one is ineffective LoRA checkpointing and the other is a missing capability flag." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Unexpected `num_items_in_batch` kwarg is a different failure from the mask/tensor shape mismatch." + "reason": "Different errors and different training-time failure modes." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading complaints, but one is a bad error message for Voxtral and the other is a specific tokenizer regression." + "reason": "Both are tokenizer-loading issues, but for different models and different root causes." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "GLM-4.7-Flash unit-test loading issues do not match NemotronH checkpoint incompatibility." + "reason": "Different model families and different checkpoint-loading failures." }, { "left": "issue:43940", "right": "issue:45313", - "accept": true, - "reason": "Same DeepSpeed ZeRO-3 weight-loading failure for Qwen3-family models; the missing-weights symptoms align closely." + "accept": false, + "reason": "Both are ZeRO-3 weight-loading problems, but the reported symptoms and model layouts differ too much to treat as one bug." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "RLHF gradient explosion from SDPA/BF16 is unrelated to flash-attention sink handling." + "reason": "RLHF NaN/SDPA collapse and flash-attn sink-less AttributeError are unrelated." }, { "left": "issue:41093", "right": "issue:44805", "accept": true, - "reason": "Same IndexError shape-mismatch failure pattern; the titles describe the same underlying mask/indexing bug." + "reason": "Same mask/tensor shape mismatch error pattern; these look like duplicate reports of the same indexing bug." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "Video-input regression across multiple models is not the same as the Qwen3-vl-embedding StopIteration error." + "reason": "Both concern video input regressions, but the titles point to different model-specific failures and not one clearly shared bug." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training `weight must be 2-D` is unrelated to void segmentation map processing." + "reason": "Training with DeepSpeed and void segmentation map processing are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet. The accepted duplicate pairs are supported by closely matching titles/signatures, and the rejected pairs are phrased conservatively without overstating equivalence. The summary stays appropriately high-level and does not claim unsupported merges beyond the evident duplicate subclusters." + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The accepted pairs are supported by the titles, especially the docs cleanup, `_is_hf_initialized`/`init_empty_weights`, and the mask-shape mismatch cluster. No soft-edge verdict seems to overstate identity beyond what the packet supports." }, "evaluator_used": true, "retried": false, @@ -1975,8 +1951,8 @@ "issue:46899" ], "soft_pairs": [ - "issue:41553|issue:44488", "issue:44530|issue:44863", + "issue:41553|issue:44488", "issue:44863|issue:46402", "issue:43366|issue:43828", "issue:45325|issue:45902", @@ -1996,8 +1972,8 @@ "issue:44560|issue:45685", "issue:42617|issue:43366", "issue:36010|issue:42915", - "issue:44683|issue:46515", "issue:44560|issue:44918", + "issue:44683|issue:46515", "issue:45812|issue:46489", "issue:45823|issue:46515", "issue:44661|issue:45823", @@ -2010,244 +1986,244 @@ "issue:43638|issue:44661", "issue:43596|issue:44661" ], - "prepared_review_unit_hash": "48a10a7d9baa20e4a66005990f9c4253685766f66dfb929126190785aa1bbf62", + "prepared_review_unit_hash": "dad94451779cb27e8ade854c5b9763296695f17633f25337dc8a4b95fad81d1f", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38059, - "estimated_input_tokens": 9515, - "estimated_eval_tokens": 19286 + "serialized_chars": 38061, + "estimated_input_tokens": 9516, + "estimated_eval_tokens": 19288 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a heterogeneous mix of unrelated issue reports. I only see two likely duplicate pairs: the Siglip2 bad-output reports and the Transformers 5 custom-model-init reports. Everything else looks like separate bugs or regressions despite superficial subsystem/model overlap.", - "confidence": 0.61, - "canonical_issue_reason": "issue:43994 is the broadest, cleanest report for the Siglip2 regression; it captures the user-visible failure mode more generally than issue:38175, which is a narrower symptom report.", + "summary": "This cluster is mostly a grab-bag of Transformers v5/model-loading regressions, but only one soft pair looks like the same underlying bug: custom model initialization broken in 5.0.0. The rest are different model-specific failures or unrelated backend/device issues.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the most general and representative report of the shared regression: Transformers 5.0.0 breaking custom model initialization. It subsumes the notebook-specific variant in 43645.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43994 is the best representative issue in this set because it describes the underlying Siglip2 model-output regression clearly and broadly, making it the best duplicate target for the accepted Siglip2 pair.", + "best_issue_reason": "issue:43646 is the best single issue to anchor the cluster because it states the core bug broadly and is most reusable for deduplication.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41553", - "right": "issue:44488", + "left": "issue:44530", + "right": "issue:44863", "accept": false, - "reason": "Different models and failures: Voxtral AutoTokenizer import/error messaging vs cjvt/sleng-bert loading." + "reason": "Both are model-loading-related at a high level, but one is a Qwen3.5 cache/runtime error and the other is a NemotronH checkpoint-loading failure. Different underlying bugs." }, { - "left": "issue:44530", - "right": "issue:44863", + "left": "issue:41553", + "right": "issue:44488", "accept": false, - "reason": "Different bugs: Qwen3.5 PagedAttentionCache crash vs NemotronH checkpoint loading failure." + "reason": "Both mention loading problems, but they affect different models and likely different code paths; the titles do not indicate the same concrete bug." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Both are load failures, but on different architectures and code paths (NemotronH vs Sam3Model text encoder weights)." + "reason": "Different models and failure modes: NemotronH checkpoint loading versus Sam3Model text encoder weights not being loaded." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Unrelated: gpt-oss GGUF support request vs Phi-tiny-MoE autocast dtype mismatch." + "reason": "Unrelated issues: GGUF architecture support versus an autocast dtype mismatch in Phi-tiny-MoE." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both involve rotary/position handling, but on different models and different failure mechanisms." + "reason": "Both involve rotary/position embedding internals, but they are different models and different failure mechanisms." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different issues: gpt-oss support vs Qwen3-MoE sliding_window behavior." + "reason": "One is GGUF gpt-oss support, the other is a Qwen3-MoE sliding_window bug; not the same bug." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "Different root causes: DeepSpeed ZeRO-3/BertModel index error vs a torch 2.4.1 compatibility complaint." + "reason": "One is a DeepSpeed ZeRO3 Bert index error; the other is a generic Torch compatibility issue. Too different to dedupe." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Different pipelines and symptoms: AudioFlamingo3 batched-inference leakage vs Whisper empty-transcription generation failure." + "reason": "Both are incorrect-generation style bugs, but one is AudioFlamingo3 batching leakage and the other is Whisper empty-transcription handling." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same family (Gemma4) but different bugs: per-layer shared_kv_states rebuild vs device_map auto device mismatch." + "reason": "Gemma4 FSDP2 shared_kv_states bug versus Gemma4 PLE device mismatch; same model family, different concrete failure paths." }, { "left": "issue:43638", "right": "issue:46515", "accept": false, - "reason": "Different problems: non-pretrained Bert/ZeRO-3 index error vs torch version incompatibility." + "reason": "Both involve loading/runtime regressions, but one is Bert/ZeRO3 initialization and the other is a Torch version incompatibility. Not the same bug." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "No clear match: gpt-oss support request vs Float8 storage/type error." + "reason": "GGUF gpt-oss support and missing Float8 storage are unrelated." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Different bugs: non-persistent buffer corruption vs Kosmos2.5 long-OCR index error." + "reason": "Persistent-buffer corruption in from_pretrained is a different class of loading bug from Kosmos2.5 long OCR index errors." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "ImportError for a missing class vs meta-device loading breakage; related loading area but not the same bug." + "reason": "Both are loading regressions, but one is a missing class import and the other is meta-device loading breakage. Different concrete problems." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Different Qwen3Moe FP8 failure vs a generic meta-device loading regression." + "reason": "Qwen3Moe FP8 failure is not the same as generic meta-device loading breakage." }, { "left": "issue:44910", "right": "issue:46858", "accept": false, - "reason": "Different cache/position-id bugs on different execution paths and backends." + "reason": "Flash-Attention 3D position_ids corruption on Qwen3.5 is different from static-cache seq-length tensor arithmetic on non-CUDA backends." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different models and symptoms: Mistral-3 dtype mismatch vs Qwen3-vl video StopIteration." + "reason": "Mistral-3 image preprocessor dtype mismatch and Qwen3-vl StopIteration are different multimodal bugs." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "These look like the same custom-model initialization regression in Transformers 5.0; the notebook-specific title is just a narrower variant." + "reason": "Both describe the same Transformers 5.0.0 regression breaking custom model initialization; 43645 is just the Jupyter-specific manifestation." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different Qwen3.5-related bugs: torch.compile+SDPA failure vs Flash Attention illegal memory access from 3D position_ids handling." + "reason": "Torch.compile/SDPA on Bamba is unrelated to Qwen3.5 Flash Attention position_ids handling." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Unrelated: Qwen3-vl video embedding failure vs an MPS histogram/int tensor limitation in MoE code." + "reason": "Qwen3-vl video StopIteration and MPS histogram on Int are unrelated failures." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "Different scope and failure mode: running 3d_parallel.py vs GGUF gpt-oss architecture support." + "reason": "3d_parallel.py runtime failure is not the same as GGUF gpt-oss support." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Different issues: GenerationMixin import breakage vs Qwen3Moe FP8 config failure." + "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 failure are completely different issues." }, { - "left": "issue:44683", - "right": "issue:46515", + "left": "issue:44560", + "right": "issue:44805", "accept": false, - "reason": "Both mention torch versions, but one is a compiled flex_attention regression on torch>=2.9 and the other is a torch 2.4.1 incompatibility report." + "reason": "Qwen3-vl video StopIteration and generic mask shape mismatch are different bugs." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:44683", + "right": "issue:46515", "accept": false, - "reason": "Different Qwen3-related failures: video StopIteration vs TRL SFT input-embedding unpacking." + "reason": "Compiled flex_attention on torch >= 2.9 is a different backend-specific issue from torch 2.4.1 incompatibility." }, { "left": "issue:45812", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and appear to be distinct bugs." + "reason": "Granite AutoTokenizer token-ID regression and DeepSeek-Coder tokenizer regression are separate tokenizer bugs." }, { "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "Gemma4 device mismatch is unrelated to the generic torch 2.4.1 incompatibility report." + "reason": "Gemma4 PLE device mismatch is not the same as the torch 2.4.1 incompatibility report." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Different bugs: add-new-model-like TOKENIZER_MAPPING_NAMES failure vs Gemma4 PLE device mismatch." + "reason": "Tokenizer mapping failure in add-new-model-like and Gemma4 PLE device mismatch are unrelated." }, { "left": "issue:38175", "right": "issue:43994", - "accept": true, - "reason": "Same underlying Siglip2 regression: one report describes zero probabilities, the other describes nonsensical output from the same model on AutoModel/pipeline." + "accept": false, + "reason": "Both mention SigLIP2 outputs, but one is zero probabilities and the other is nonsensical AutoModel/pipeline results; not clearly the same concrete bug." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Different symptoms and likely different code paths: video embedding StopIteration vs a mask/tensor shape mismatch error." + "reason": "Qwen3-vl StopIteration and mask/index shape mismatch are different failures." }, { "left": "issue:45672", "right": "issue:46899", "accept": false, - "reason": "Both are Gemma4 multimodal issues, but one is an audio uint8/torch.finfo bug and the other is a broader 4-bit dtype-cast problem." + "reason": "Both concern Gemma4 and quantization, but one is uint8 torch.finfo in audio modules and the other is unsafe dtype casts in multimodal embedders; different code paths." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "Different failures: `_is_hf_initialized` flag regression vs Kosmos2.5 long OCR indexing." + "reason": "_is_hf_initialized flag regression and Kosmos2.5 long OCR index error are unrelated." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Different validation/loading regressions affecting different configs." + "reason": "Strict config rejection for granite_speech and glm4v tiny-random config validation are different config-validation issues." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different breakages: hf4.x/hf5.x LoRA compatibility vs BitNet packed-weight loading in CI." + "reason": "Qwen3 MoE LoRA compatibility and BitNet packed-weight loading are unrelated." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Different bugs: ZeRO-3/BertModel indexing vs add-new-model-like tokenizer mapping failure." + "reason": "DeepSpeed ZeRO3 Bert index error and tokenizer-mapping-name handling are different bugs." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different issues and code paths: DeepSpeed/BertModel index error vs tokenizer-mapping helper failure." + "reason": "A ZeRO3/Bert initialization index error is not the same as add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The two accepted duplicate pairs are plausible from the titles/body context, and the rejection calls are generally cautious rather than overstated. No mergeability issues appear because there are no PR pairs here." + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted soft pair, issue:43645 \u2194 issue:43646, is well-supported by the titles and cluster context, and the remaining rejections do not overstate similarity. The choice of issue:43646 as the canonical anchor is reasonable because it is the broader report of the same Transformers 5.0.0 custom model initialization regression." }, "evaluator_used": true, "retried": false, @@ -2325,20 +2301,20 @@ "issue:41720|issue:42915", "issue:42491|issue:45276", "issue:42915|issue:44756", - "issue:44451|issue:44488", "issue:44960|issue:44977", + "issue:44451|issue:44488", "issue:44863|issue:46846", "issue:46459|issue:46512", "issue:45313|issue:45357", "issue:45507|issue:46468", "issue:44488|issue:44991", "issue:43644|issue:43645", - "issue:43257|issue:43854", "issue:39692|issue:43994", + "issue:43257|issue:43854", "issue:43856|issue:43873", "issue:42491|issue:45292", - "issue:43856|issue:44387", "issue:43854|issue:43931", + "issue:43856|issue:44387", "issue:43708|issue:44075", "issue:42491|issue:45335", "issue:44960|issue:46050", @@ -2349,190 +2325,190 @@ "issue:45276|issue:45335", "issue:44910|issue:45588" ], - "prepared_review_unit_hash": "2a3bd771f122b29819bfee4c83224a0d7dc6019cddcdc793fe2dfc1ac88b1190", + "prepared_review_unit_hash": "6ea19ae15da147eb16df3e2e87d9f0df3163757cb3e82575f70735120c6f895a", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36659, - "estimated_input_tokens": 9165, - "estimated_eval_tokens": 18586 + "serialized_chars": 36654, + "estimated_input_tokens": 9164, + "estimated_eval_tokens": 18584 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue families. The strongest canonical candidate is issue:43644, but its relationship to issue:44534 should be phrased as a plausible near-duplicate signal based on title/reference similarity, not an explicit duplicate link; issue:46897 and issue:46901 are the other clear duplicate-like pair.", - "confidence": 0.82, - "canonical_issue_reason": "issue:43644 is the cleanest canonical issue for the cluster because it describes a specific Transformers v5 regression and, among the listed items, has the strongest near-duplicate signal with issue:44534. The packet supports that as a plausible duplicate via title similarity and reference activity, but not as an explicit linked duplicate.", + "summary": "This cluster is mostly heterogeneous, but it contains tight duplicate families around tokenizer-loading regressions in v5, non-persistent buffer initialization, resize_token_embeddings propagation, and training-loss double-shift. The SigLIP2 items are only loosely related and should not be treated as a tight duplicate family.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 44991 is the clearest canonical representative because it is a specific, well-scoped transformers>=5.0.0 tokenizer-loading regression with a reproducible model impact and strong fit to the tokenizer-family duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the best representative issue because it is self-contained, broadly framed, and the closest thing in the set to a canonical regression report. Its pairing with issue:44534 looks duplicate-like, though only tentatively from the packet evidence.", + "best_issue_reason": "Issue 44991 is the best global issue representative: it is concrete, narrowly framed, and anchors the strongest duplicate subgroup in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both are offload/memory-related, but they describe different failures and likely different code paths." + "reason": "Both are offload/memory related, but they describe different failures and likely different fixes: dequantized model loading on gpu+cpu offload vs disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping failure vs mask/index shape mismatch; these are unrelated bugs." + "reason": "TOKENIZER_MAPPING_NAMES lookup failure and mask/tensor shape mismatch are unrelated code paths." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but one is zero-probability output behavior and the other is docs/example plus quantization issues." + "reason": "Same model family, but one is inference output corruption and the other is a documentation example/model-processor mismatch with quantization failure; not the same underlying bug." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "These look like the same Transformers v5 non-persistent-buffer regression; the titles and surrounding references are strongly aligned." + "reason": "Both report the same Transformers v5 non-persistent buffer regression; the titles are essentially the same bug phrased differently." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Different Qwen3 failures with different symptoms and likely different root causes." + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3Moe FineGrainedFP8Config failure are different concrete failures." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Qwen3 MoE HF4.x/HF5.x incompatibility versus a Gemma4 resize_token_embeddings propagation bug." + "reason": "LoRA hf4.x to hf5.x incompatibility and resize_token_embeddings not updating output embeddings are unrelated bugs." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "FP8 loading failure versus a Strix Halo mmap/OOM workaround; not the same bug." - }, - { - "left": "issue:44451", - "right": "issue:44488", - "accept": false, - "reason": "Both are model-loading complaints, but the affected models and reported symptoms differ." + "reason": "FineGrainedFP8Config on Qwen3Moe and Strix Halo mmap OOM are different problems." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Unrelated model families and failure modes." + "reason": "GLM5 and Qwen3.5 flash-attention issues do not share the same underlying failure." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Both describe the same tokenizer-loading regression in v5, expressed as model-specific load failures." }, { "left": "issue:44863", "right": "issue:46846", "accept": false, - "reason": "NemotronH checkpoint loading vs Qwen3.5 TP-plan/OOM-shape issues; different code paths." + "reason": "NemotronH checkpoint loading failure and missing linear_attn TP-plan entries are different bugs." }, { "left": "issue:46459", "right": "issue:46512", "accept": false, - "reason": "Both involve dtype handling, but one is AutoModelForCausalLM loading and the other is from_config/config staleness." + "reason": "Ignoring dtype during multimodal checkpoint loading and stale config.dtype after from_config are related but distinct issues." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5-related, but they affect different surfaces: ZeRO-3 weight loading vs save_pretrained serialization." + "reason": "ZeRO-3 language_model weight loading and incorrect visual-encoder keys in save_pretrained are separate serialization/loading bugs." }, { "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "Different models and different failure mechanisms." + "reason": "GraniteMoEHybrid invalid method call and broken Nemotron example are unrelated." }, { "left": "issue:44488", "right": "issue:44991", - "accept": false, - "reason": "Tokenizer-loading regressions, but on different model families and with different concrete bugs." + "accept": true, + "reason": "Both describe the same tokenizer-loading regression in v5, expressed as model-specific load failures." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Same release family, but buffer initialization and custom-model notebook initialization are different bugs." + "reason": "Non-persistent buffer junk and Jupyter custom-model initialization breakage are different regressions." }, { - "left": "issue:43257", - "right": "issue:43854", + "left": "issue:39692", + "right": "issue:43994", "accept": false, - "reason": "Qwen3 MoE accelerate/deepspeed conversion issue versus GLM-4.7-Flash unit-test load failure." + "reason": "The first is a doc example/model-processor mismatch plus quantization failure; the second is inference corruption on AutoModel/pipeline. Similar model, different bug." }, { - "left": "issue:39692", - "right": "issue:43994", + "left": "issue:43257", + "right": "issue:43854", "accept": false, - "reason": "Both touch SigLIP2, but one is docs/example mismatch and the other is nonsensical inference output." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and GLM-4.7-Flash unit-test load failure are unrelated." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both are memory-related, but the training-memory issue and the quantization offloading issue are distinct." + "reason": "Qwen3 MoE training memory inefficiency and quantization offload not working are different concrete issues." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Different model family and different embedding-propagation surface." + "reason": "LoRA hf4.x/hf5.x compatibility is unrelated to resize_token_embeddings not updating output embeddings." }, { - "left": "issue:43856", - "right": "issue:44387", + "left": "issue:43854", + "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE training memory versus int4 reserved-memory/OOM regression." + "reason": "GLM-4.7-Flash test loading and Qwen3-VL weight-shape mismatch are different model-loading failures." }, { - "left": "issue:43854", - "right": "issue:43931", + "left": "issue:43856", + "right": "issue:44387", "accept": false, - "reason": "Different model-load failures with different model families and symptoms." + "reason": "Training memory inefficiency and increased reserved memory under int4 quantization are not the same bug." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume step-count bug versus SGD args not being used." + "reason": "Trainer checkpoint-resume max_steps bug and SGD args ignored are unrelated training issues." }, { "left": "issue:42491", "right": "issue:45335", "accept": false, - "reason": "Different model families and different resize_token_embeddings behavior." + "reason": "LoRA model incompatibility and t5gemma resize_token_embeddings propagation are different problems." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "GLM5 issue versus sinq save/load corruption; unrelated." + "reason": "GLM5 issue and sinq save/load breakage are unrelated." }, { "left": "issue:46897", "right": "issue:46901", "accept": true, - "reason": "Same underlying training-loss double-shift bug: both describe labels being shifted incorrectly during loss computation." + "reason": "Both are training-loss double-shift bugs: labels are shifted one token too far in the loss path, just on different models." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Extra_state handling versus init_empty_weights/_is_hf_initialized constructor incompatibility." + "reason": "extra_state handling and _is_hf_initialized argument handling are different save/load regressions." }, { "left": "issue:45362", @@ -2544,19 +2520,25 @@ "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they are different concrete issues." + "reason": "Both are tokenizer regressions, but one overrides ByteLevel with Metaspace for DeepSeek V3/R1 and the other produces wrong token IDs for Granite models; not the same code-path bug." }, { "left": "issue:45276", "right": "issue:45335", + "accept": true, + "reason": "These are the same resize_token_embeddings propagation bug on different model variants: output embeddings are not updated." + }, + { + "left": "issue:44910", + "right": "issue:45588", "accept": false, - "reason": "Very similar API area, but the affected architectures and embedding surfaces differ enough to treat them as separate bugs." + "reason": "Illegal memory access from misread 3D position_ids and AttributeError on s_aux=None are different flash-attention failures." } ] }, "evaluator_result": { "accept": false, - "feedback": "The clustering is mostly conservative, but the canonical/best-issue rationale overstates the packet evidence. In particular, calling issue:44534 an \"explicit near-identical duplicate\" of issue:43644 is not supported by the packet: there are no explicit issue links, only highly similar titles and some reference activity. That makes the duplicate claim plausible, but not explicit. The summary should be phrased more tentatively. The other soft-edge verdicts look suitably conservative." + "feedback": "The summary is slightly overstated. The tokenizer-loading, non-persistent buffer, resize_token_embeddings, and double-shift families are grounded, but the SigLIP2 portion is too strong: the packet supports related SigLIP2 reports, not a clearly established tight duplicate family, and the analyst explicitly rejected the relevant pairings. Please downgrade that to a looser similarity claim or omit it." }, "evaluator_used": true, "retried": true, @@ -2659,20 +2641,20 @@ "issue:43646|issue:43708", "issue:43645|issue:44164" ], - "prepared_review_unit_hash": "b5efba6b72fab8ed02c4d47d5285197ed32a42f7159f9da77f6f863603879ed5", + "prepared_review_unit_hash": "e8f12c843f84a80ec57d5bdcda13b119039503eb2971a3b668964bceb8ed857a", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37061, + "serialized_chars": 37064, "estimated_input_tokens": 9266, "estimated_eval_tokens": 18788 }, @@ -2681,204 +2663,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous Transformers issues; only the resize_token_embeddings/output-embeddings reports form a clear duplicate cluster. No PRs are present in the input.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45292 is the best canonical issue because it states the bug in the most generic form ('resize_token_embeddings does not effect to output_embeddings') and is the central match for the closely related Gemma4 and t5gemma variants.", + "summary": "This cluster is mostly heterogeneous, but there are two small duplicate-like subclusters: the resize_token_embeddings reports (45276/45292/45335) and the untied lm_head reports (43404/43454). I\u2019d use 45292 as the canonical issue because it is the most generic description in the strongest duplicate set.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45292 is the most generic and central report in the clearest duplicate subcluster, covering the shared resize_token_embeddings propagation problem across model variants.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the strongest representative of the only clear same-bug cluster in this set; the other items are different model- or subsystem-specific regressions.", + "best_issue_reason": "Best representative of the cluster\u2019s strongest duplicate family: it states the shared symptom broadly, while the nearby issues are model-specific variants of the same resize_token_embeddings bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but one is a vocab-size mismatch and the other is inability to load a specific model; too different to treat as the same bug." + "reason": "Related BERT loading complaints, but one is a tokenizer vocab-size mismatch and the other is a model-specific load failure; not enough to call the same bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 docs/example errors and GLM5 loading are unrelated models and failure modes." + "reason": "Different topics entirely: SigLIP2 doc example errors vs a GLM5 model issue." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "CUDA assert during Qwen3 auto device mapping is unrelated to AudioFlamingo3 batched inference token/embedding leakage." + "reason": "Unrelated models and failure modes (Qwen3 device-map crash vs AudioFlamingo3 batched inference token leak)." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states KeyError and Qwen multimodal text_config unwrapping are different code paths and different failure modes." + "reason": "Different code paths and symptoms: FSDP/shared_kv_states KeyError vs AutoModelForCausalLM text_config unwrapping." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch 2.10 compatibility and meta-device loading failures are distinct regression categories." + "reason": "One is a Torch-version incompatibility, the other is a meta-device loading regression; not the same bug." }, { "left": "issue:43404", "right": "issue:43454", - "accept": false, - "reason": "Both mention tied lm_head weights, but they affect different model families and are not the same concrete bug." + "accept": true, + "reason": "Both report untied lm_head weights causing bad generation in related multimodal generation models; same underlying weight-tying bug." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing with PEFT LoRA is unrelated to a T5/apex RMSNorm memory leak." + "reason": "Gradient checkpointing/LoRA behavior is unrelated to the apex.FusedRMSNorm memory leak." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected _is_hf_initialized argument in init_empty_weights is unrelated to mmap OOM on Strix Halo." + "reason": "Unexpected init_empty_weights argument and Strix Halo mmap OOM are separate loading/runtime problems." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both are quantization-related OOMs, but one is offloading behavior and the other is increased reserved memory under int4; not the same bug." + "reason": "Both mention quantization, but one is offloading behavior and the other is increased reserved memory/OOM; different root causes." }, { "left": "issue:44079", "right": "issue:46829", "accept": false, - "reason": "ModelOutput key assignment bug and VideoPrism hidden_states misassignment are different output-shaping issues." + "reason": "ModelOutput key assignment bug is unrelated to VideoPrism output field mapping." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and Qwen3_5Moe rotary embedding CPU-offload incompatibility are separate issues." + "reason": "FSDP2 shared_kv_states KeyError and CPU offload incompatibility in Qwen3_5Moe are different issues." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Qwen-3-8B-NVFP4 loading failure and git-oss-20b MXFP4 fallback are different models and different quantization problems." + "reason": "Different quantization formats and symptoms: NVFP4 load error vs MXFP4 fallback to bf16." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "An import error for GenerationMixin is unrelated to Whisper pipeline return_language behavior." + "reason": "ImportError for GenerationMixin is unrelated to Whisper return_language pipeline regression." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization memory regression and mmap OOM on Strix Halo are unrelated failure causes." + "reason": "Quantization OOM and disabling mmap on Strix Halo are distinct problems." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 track leakage is unrelated to Gemma4 PLE device mismatch under device_map=auto." + "reason": "AudioFlamingo3 track leakage and Gemma4 device mismatch are not the same code-path problem." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "PEFT gradient checkpointing ineffectiveness and eval-mode checkpointing prohibition are different checkpointing bugs." + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode incompatibility." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading regressions and Strix Halo mmap OOM are not the same underlying problem." + "reason": "Meta-device loading breakage is unrelated to the Strix Halo mmap OOM report." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and Gemma4 FSDP2 KeyError are unrelated." + "reason": "SwitchTransformers sparse-layer creation and Gemma4 FSDP2 shared_kv_states are unrelated model bugs." }, { "left": "issue:44863", "right": "issue:46468", "accept": false, - "reason": "NemotronH checkpoint loading and a broken official Nemotron example are too broad and not clearly the same concrete bug." + "reason": "Nemotron checkpoint loading failure and an official example being broken are too vague to treat as the same bug." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "These describe the same resize_token_embeddings bug: updated embeddings/output heads are not propagated correctly, with 45276 being a model-specific manifestation." + "reason": "Same resize_token_embeddings bug: 45276 is a Gemma4-specific manifestation, while 45292 states the generic output_embeddings failure." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video input regression in Qwen VL models and strict config blocking granite_speech are unrelated." + "reason": "Video-input regression and strict config loading are unrelated." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Gradient checkpointing eval-mode handling is unrelated to RLHF NaNs from 3D position_ids and SDPA fallback." + "reason": "Eval-mode gradient checkpointing restriction is unrelated to RLHF NaN explosions from 3D position_ids." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling and mmap OOM on Strix Halo are unrelated." + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "flash_attention.py crashing on s_aux=None is a different code path from Gemma4 FSDP2 shared_kv_states rebuilding." + "reason": "Flash-attention sinkless-model crash is unrelated to the Gemma4 FSDP2 attention KeyError." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Tokenizer BPE marker handling and continue_final_message behavior are both mistral_common-related, but they are distinct tokenizer bugs." + "reason": "Tokenizer BPE-marker handling and mistral_common continue_final_message are different tokenizer bugs." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Read-only permission propagation during save_pretrained and loading the wrong custom module after save_pretrained are different save/load issues." + "reason": "Custom-model file permissions and loading the wrong custom module after save_pretrained are separate save/load issues." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Unit-test model loading failure for GLM-4.7-Flash and Qwen3.5 checkpoint saving regression are different bugs." + "reason": "GLM-4.7-Flash unit-test loading and Qwen3.5 save_pretrained regression are different model-specific problems." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Both report the same underlying resize_token_embeddings propagation failure, with 45335 as a model-specific decoder.embed_tokens example." + "reason": "Both describe the same resize_token_embeddings omission; 45335 is the T5Gemma-specific instance of the generic issue in 45292." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Vocab-size mismatch for a specific model and inability to load a different model are not the same defect." + "reason": "Both concern model loading, but the tokenizer vocab-size mismatch and the ScandiBERT load failure are not clearly the same root cause." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization breakage in Jupyter and Trainer max_steps miscalculation on resume are unrelated." + "reason": "Custom model initialization regression and Trainer resume_from_checkpoint max_steps calculation are unrelated." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Jupyter custom model initialization and save/from_pretrained extra_state handling are separate problems." + "reason": "Jupyter custom-model initialization and save/from_pretrained extra_state handling are different bugs despite both touching model persistence/loading." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet: the only clearly supported duplicate cluster is the resize_token_embeddings/output_embeddings set centered on issue:45292, with issue:45276 and issue:45335 as conservative matches. The non-matches are appropriately cautious and do not overstate evidence. No PRs are present, so that part is also consistent." + "feedback": "Grounded overall. The summary stays conservative about the cluster being mostly heterogeneous and only calls out the two clearly supported duplicate families. The canonical choice of 45292 is reasonable because it is the most generic title in the resize_token_embeddings subcluster. The accepted soft-edge pairs (43404/43454, 45276/45292, 45292/45335) are defensible from the packet titles, and the rejected pairs do not overreach." }, "evaluator_used": true, "retried": false, @@ -2961,8 +2943,8 @@ "issue:39404|issue:46665", "issue:43425|issue:46869", "issue:44683|issue:45823", - "issue:43425|issue:45823", "issue:43425|issue:43716", + "issue:43425|issue:45823", "issue:44479|issue:44805", "issue:44589|issue:45084", "issue:43906|issue:44530", @@ -2984,20 +2966,20 @@ "issue:45479|issue:46897", "issue:43596|issue:44683" ], - "prepared_review_unit_hash": "ab3a4c664a98d07a63ee763ce63b5bce9fd2a3ecf1cc58d9b5aaf50fc8a03f44", + "prepared_review_unit_hash": "4946c1c6c06a3d52f69de3b146df984a7c71c9cc86be7098cdd67b92cb01d3f3", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38134, + "serialized_chars": 38135, "estimated_input_tokens": 9534, "estimated_eval_tokens": 19324 }, @@ -3006,138 +2988,138 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of Transformers v5 regression reports, with a noticeable tokenizer/loading subtheme but many unrelated model, training, and docs bugs. The soft pairs mostly overlap only at a high level (same subsystem or similar symptom), not the same concrete defect.", - "confidence": 0.84, - "canonical_issue_reason": "issue:45701 is the broadest umbrella for the shared tokenizer-regression theme and best matches the cluster\u2019s main recurring pattern.", + "summary": "This cluster is mostly a collection of distinct Transformers v5 regression reports spanning tokenization, model loading, dtype/device mismatches, and quantization. The soft pairs are largely topical similarities rather than the same concrete bug.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43611 is the best representative because it is a broad, concrete v5 model-loading regression and sits near the center of the cluster\u2019s recurring loading/config failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest representative because it is concrete, user-facing, and describes a clear v4\u2192v5 tokenizer regression with wide model impact.", + "best_issue_reason": "issue:43611 is the clearest anchor for triage: it is specific enough to be actionable, but broad enough to represent the dominant loading-regression theme in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46894", "right": "issue:46901", "accept": false, - "reason": "Both describe training-loss double-shift bugs, but on different model families with no evidence of the same code path." + "reason": "Same high-level pattern (training-loss double shift), but these are different model reports with no evidence of the same underlying code path." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Both mention dtype/device-related failures, but one is autocast mismatch during inference and the other is meta-device loading; too different to merge." + "reason": "Both mention dtype mismatches, but one is an autocast forward issue and the other is a meta-device loading issue for different models." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage deserialization and mmap/OOM are unrelated failure modes." + "reason": "Unrelated failures: missing Float8 storage vs a Strix Halo mmap OOM workaround." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "Tokenizer performance regression vs CodeLlama space-stripping regression; different bugs." + "reason": "One is a performance regression in convert_tokens_to_ids; the other is a round-trip tokenization correctness bug." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Wrong token IDs and slower convert_tokens_to_ids are separate tokenizer issues." + "reason": "Tokenizer wrong-ID regression and tokenizer performance regression are different bugs and fixes." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is not the same defect as the Phi autocast dtype mismatch." + "reason": "Torch version incompatibility is not the same bug as a model-specific dtype mismatch under autocast." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language behavior is distinct from the broader tokenization regression theme." + "reason": "Whisper pipeline return_language breakage and generic tokenization-change reports are not the same underlying issue." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache group-type crash and template-node compilation failure are different code paths." + "reason": "Different code paths and symptoms: PagedAttentionCache group-type crash vs template-node compilation error." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Base_model_prefix loading failure and Jupyter custom-model initialization issues are different root causes." + "reason": "Both involve loading/custom model behavior, but one is base_model_prefix handling and the other is notebook-defined custom model initialization; not the same bug." }, { "left": "issue:39404", "right": "issue:46665", "accept": false, - "reason": "Pipeline return_language regression is unrelated to from_pretrained behavior inside classes." + "reason": "Whisper pipeline behavior and from_pretrained inside Python classes are unrelated loading/pipeline issues." }, { "left": "issue:43425", "right": "issue:46869", "accept": false, - "reason": "Torch 2.10 incompatibility and YOLO multi-GPU device_map failure are not the same bug." + "reason": "Torch incompatibility and YOLO multi-GPU device-map failure are different problems." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are unrelated." + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to Gemma4 device mismatch during forward." }, { "left": "issue:43425", - "right": "issue:45823", + "right": "issue:43716", "accept": false, - "reason": "General torch incompatibility does not match the specific Gemma4 PLE device mismatch." + "reason": "No shared bug: torch version compatibility vs Mistral-3 image preprocessor/model dtype mismatch." }, { "left": "issue:43425", - "right": "issue:43716", + "right": "issue:45823", "accept": false, - "reason": "Torch version incompatibility is not the same as Mistral-3 image dtype mismatch." + "reason": "Different failures with different triggers and fix scopes." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression and mask-shape IndexError are different model failures." + "reason": "A video-input regression and a mask-shape IndexError are not the same code-path problem." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to template compilation errors." + "reason": "Completely different failure modes and subsystems." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction of another issue is not the same as the PagedAttentionCache crash." + "reason": "An isolated reproduction of an older issue is not the same as a Qwen3.5 cache/group-type crash." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs issues, but they concern different pipeline/deprecation topics." + "reason": "Both are docs-related, but they cover different pipeline/API removals and are not duplicate documentation bugs." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Siglip2 output quality regression and ModelOutput key assignment bug are different." + "reason": "Incorrect Siglip2 outputs and ModelOutput key assignment are unrelated issues." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "T5Tokenizer add_special_tokens conflict and gguf_file loading breakage are separate tokenizer/model-loading failures." + "reason": "Tokenizer special-token conflict and gguf loading failures are distinct loading bugs." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export/torch_compilable_check and Perceiver interpolation failure are different export/inference issues." + "reason": "torch.export failures with torch_compilable_check and Perceiver resolution failures are different problems." }, { "left": "issue:39404", @@ -3149,73 +3131,73 @@ "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized save/load breakage and mistral_common continue_final_message behavior are not the same defect." + "reason": "Quantized save/load breakage and mistral_common continue_final_message behavior do not share the same underlying bug." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "Extra_state save/load handling and composite-config unwrapping are different loading bugs." + "reason": "save/from_pretrained extra_state handling and Qwen composite config unwrapping are separate loading issues." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "These reference different issues and different behaviors; no shared concrete bug." + "reason": "No concrete overlap beyond both involving Transformers APIs." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and tokenizer-loading failure for est-roberta are separate problems." + "reason": "CamemBERT MLM regression and tokenizer load failure for a different model are not duplicates." }, { "left": "issue:45823", "right": "issue:46899", "accept": false, - "reason": "Both are Gemma4-related, but one is a device mismatch and the other is a 4-bit dtype-cast bug in multimodal embedders." + "reason": "Both involve Gemma4, but one is a device-map forward mismatch and the other is an unsafe dtype cast in multimodal embedders." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading failure for one model and mistral regex patch crash for another are different code paths." + "reason": "Different tokenizers, different failure points, different fixes." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both involve Qwen3-VL loading, but one is a broad v5.dev0 breakage and the other is a weight-shape mismatch; not enough to unify." + "reason": "Same model family, but one is a general loading breakage and the other is a specific weight-shape mismatch; not enough to treat as the same bug." }, { "left": "issue:45479", "right": "issue:46894", "accept": false, - "reason": "Single-label zero-loss classification bug and double-shift training-loss bug are not the same root cause." + "reason": "A classification-loss degeneration and an ASR training-loss shift bug are unrelated." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "MoE training memory inefficiency and Strix Halo mmap OOM are unrelated memory issues." + "reason": "Training memory inefficiency and mmap OOM avoidance are different issues." }, { "left": "issue:44164", "right": "issue:46821", "accept": false, - "reason": "extra_state save/load handling and TimesFM window_size AttributeError are different API failures." + "reason": "extra_state save/load failure and a TimesFM window_size AttributeError are unrelated." }, { "left": "issue:45479", "right": "issue:46897", "accept": false, - "reason": "Different training-loss bugs affecting different models; only a superficial pattern match." + "reason": "Different model families and different loss bugs." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "Deepspeed ZeRO3/BertModel IndexError and compiled flex_attention on torch>=2.9 are unrelated." + "reason": "Deepspeed zero3/BertModel index error and compiled flex_attention on torch>=2.9 are unrelated." } ] }, @@ -3298,8 +3280,8 @@ "issue:46285|issue:46402", "issue:39404|issue:44671", "issue:43716|issue:45084", - "issue:43643|issue:44530", "issue:44737|issue:45859", + "issue:43643|issue:44530", "issue:44661|issue:44683", "issue:40990|issue:43421", "issue:42915|issue:44530", @@ -3320,224 +3302,215 @@ "issue:43317|issue:43482", "issue:43262|issue:43377" ], - "prepared_review_unit_hash": "39d04e6acdaa1222657d7d5eb84e1c11e40edc3c670e83343def4dea122478e0", + "prepared_review_unit_hash": "3a7f7a8258fbef0a2f905cf57de4d3b2c4b26ebc6ce99565177ba9ef0fb7d9f3", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36728, - "estimated_input_tokens": 9182, - "estimated_eval_tokens": 18620 + "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 is a loose cluster of Transformers/tokenizer/model-loading regressions, with one clearest duplicate candidate in the audio feature-extraction crash pair (44206/44220). There are also a few same-subsystem pairs that look superficially related from titles alone (for example 43611/43644, 44488/44987, 44987/44991, and 44811/45584), but the available evidence still points to separate bugs or regressions rather than true duplicates.", - "confidence": 0.63, - "canonical_issue_reason": "Issue 43611 is the broadest representative of the cluster\u2019s dominant Transformers v5 loading/initialization regression theme.", + "summary": "These items are mostly unrelated closed issues. The soft-similarity pairs share broad themes like tokenizer loading, Whisper, or Transformers v5 regressions, but they do not look like the same underlying bug and should not be merged as duplicates.", + "confidence": 0.87, + "canonical_issue_reason": "issue:44488 is the most central tokenizer-loading report in this set: it is broad, well-discussed, has inbound references, and sits closest to several other tokenizer-related reports by symptom.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the best single issue representative because it is broad, central to the cluster, and captures the main v5 regression pattern better than the more model-specific reports.", + "best_issue_reason": "issue:44488 is the best representative issue for this cluster because it is the broadest and most connected loading/tokenizer regression among the candidates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are tokenizer-loading failures, but they involve different models and no clear shared code path." + "reason": "Both are tokenizer/model loading failures, but they affect different repos and likely different code paths." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Tokenizer/model-loading failure vs a Mistral regex patch crash; different bugs." + "reason": "One is a generic tokenizer loading issue; the other is a Mistral regex patch crash with a different failure mode." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer corruption." + "reason": "Both are Transformers v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers." }, { "left": "issue:44462", "right": "issue:46489", "accept": false, - "reason": "Both concern tokenizers, but one ignores tokenizer.json while the other is a model-specific output regression." + "reason": "Both mention tokenizer loading/output on specific models, but the symptoms and likely fixes differ." }, { "left": "issue:42796", "right": "issue:43404", "accept": false, - "reason": "Tokenizer BPE marker handling and Mistral lm_head tying are unrelated issues." + "reason": "Tokenizer BPE marker handling and Mistral lm_head tying are unrelated bugs." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer method conflict vs FineGrainedFP8Config failure; different subsystems." + "reason": "Tokenizer init conflict vs FineGrainedFP8Config failure; different subsystems and root causes." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "CLIP structural mismatch vs Sam3 text-encoder weight loading; unrelated." + "reason": "Both are model-structure issues, but CLIP structure and Sam3 weight loading are separate problems." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline behavior vs CamemBERT masked-LM predictions; different models and bugs." + "reason": "Whisper pipeline return_language behavior and CamemBERT masked-LM predictions are unrelated." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor dtype mismatch vs template compilation error; unrelated code paths." + "reason": "Image preprocessor dtype mismatch vs template compilation error are different bugs." }, { - "left": "issue:43643", - "right": "issue:44530", + "left": "issue:44737", + "right": "issue:45859", "accept": false, - "reason": "AutoConfig/trust_remote_code missing fields vs PagedAttentionCache linear_attention crash; different failures." + "reason": "Both are device/offload related, but they hit different model components and failure modes." }, { - "left": "issue:44737", - "right": "issue:45859", + "left": "issue:43643", + "right": "issue:44530", "accept": false, - "reason": "Both are device/offload-related, but they affect different models and mechanisms." + "reason": "AutoConfig trust_remote_code field loss and PagedAttention linear_attention crash are unrelated." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer mapping registration and flex_attention compilation are unrelated." + "reason": "TOKENIZER_MAPPING_NAMES handling and flex_attention compilation are different code paths." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity regression vs runtime special-token post-processor update; not the same bug." + "reason": "Perplexity regression and runtime post-processor refresh are different tokenizer issues." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Same model family, but FP8 config failure and PagedAttention linear_attention crash are different code paths." + "reason": "Both involve Qwen variants, but one is FP8 config loading and the other is a PagedAttention cache crash." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are loading failures, but for different models with no concrete shared defect." - }, - { - "left": "issue:34567", - "right": "issue:40444", - "accept": false, - "reason": "Token counting metric bug vs multi-image IterableDataset fine-tuning failure; unrelated." + "reason": "Both are model-loading failures, but they target different repositories and likely different causes." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are initialization regressions, but buffer junk and custom model init failure are not the same underlying bug." + "reason": "Both touch Transformers 5 initialization behavior, but one is about non-persistent buffers and the other custom model init." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Grouped_mm tracing error vs non-template-node compilation error; unrelated." + "reason": "Torch Dynamo grouped_mm tracing and non-template-node compilation are unrelated." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "Multi-image training failure vs mistral_common continue_final_message bug; different areas." + "reason": "VL iterable-dataset batching and continue_final_message are different multimodal tokenizer flows." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "AppleSilicon warmup TypeError and _is_hf_initialized flag regression are distinct issues." + "reason": "AppleSilicon allocator warmup and the _is_hf_initialized flag are unrelated." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Both backend-related, but one is an install-extra omission and the other is a backend-mapping validation error." + "reason": "Both are backend/dependency issues, but they concern different backends and installation paths." }, { "left": "issue:44206", "right": "issue:44220", - "accept": true, - "reason": "Both point to the same audio feature-extraction path and the same crash in _torch_extract_fbank_features()/LasrFeatureExtractor." + "accept": false, + "reason": "FeatureExtractor center-arg handling and _torch_extract_fbank_features are different audio preprocessing bugs." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method conflict vs Jupyter/custom-model initialization breakage; unrelated." + "reason": "Tokenizer add_special_tokens conflict and notebook custom initialization are unrelated." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both are Whisper/tokenization-related, but decode skip_special_tokens and empty-transcription generation are different bugs." + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens and empty-transcription generation are different issues." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "GGUF loading breakage and Qwen2.5-GGUF v5 loading failure are model-specific and not clearly the same defect." + "reason": "GGUF loading via AutoTokenizer/AutoModel and Qwen2.5-GGUF v5 loading are distinct failures." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are loading regressions in v5, but they affect different tokenizers/models and lack a shared concrete code path." + "reason": "Both are v5+ tokenizer/loading regressions, but they affect different repositories and likely need different fixes." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing tied-weights metadata vs tokenizer loading failure; unrelated." + "reason": "Molmo tied-weights attribute error and tokenizer repository loading are unrelated." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error vs GraniteMoEHybrid invalid method call; different bugs." + "reason": "Template compilation error and GraniteMoEHybrid invalid method are not the same bug." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Dequantized model offload failure vs GGUF loading failure; unrelated." + "reason": "Dequantized model offload failure and GGUF loading failure are distinct problems." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Chat-template sampling-rate default bug vs padding-mask discrepancy in MIMI encoder; different issues." + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask batching are different bugs." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The soft-edge verdicts are mostly conservative, but the summary overstates the evidence by saying the audio feature-extraction crash is the *only* plausible duplicate pair from titles alone. The packet also contains other plausible same-subsystem/same-regression pairings (for example 43611/43644, 44488/44987, 44987/44991, and 44811/45584), even if they should still be rejected as duplicates absent stronger evidence. Please soften that claim and describe the cluster as having one clearest duplicate candidate rather than the only plausible one." - }, - "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 }, @@ -3609,15 +3582,15 @@ "issue:43645|issue:46661", "issue:44403|issue:45081", "issue:42915|issue:45084", - "issue:44849|issue:45325", "issue:43792|issue:46869", + "issue:44849|issue:45325", "issue:43377|issue:43819", "issue:43404|issue:44466", "issue:44464|issue:45987", - "issue:43883|issue:44991", "issue:28530|issue:34567", - "issue:43994|issue:44610", "issue:44987|issue:45081", + "issue:43994|issue:44610", + "issue:43883|issue:44991", "issue:44008|issue:46899", "issue:42915|issue:44291", "issue:44451|issue:45081", @@ -3627,224 +3600,212 @@ "issue:43089|issue:45987", "issue:43322|issue:43540", "issue:44479|issue:45325", - "issue:42915|issue:44589", "issue:43632|issue:46821", + "issue:42915|issue:44589", "issue:36032|issue:41720", "issue:43531|issue:44877", "issue:44991|issue:46315", "issue:45216|issue:45357" ], - "prepared_review_unit_hash": "4a80aa978ce750051ac12fd546341d9607914b943a78dbdc549bc536c679badf", + "prepared_review_unit_hash": "a695015a51ad6377c8b28ef4d878cf12b598b5efa0ad27c23836702140300518", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36618, - "estimated_input_tokens": 9155, - "estimated_eval_tokens": 18566 + "serialized_chars": 36615, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous model-loading and regression reports. Only one soft pair looks like a plausible duplicate in the same Qwen VL temporal-position/rope-index path; the rest are distinct bugs or symptoms in different subsystems.", - "confidence": 0.71, - "canonical_issue_reason": "Issue 44479 is the broadest and clearest representative: it describes a concrete v5.3.0 regression affecting multiple Qwen VL variants, which likely covers the narrower Qwen2.5-VL temporal position-id bug as a specific manifestation.", + "summary": "The packet supports one clear duplicate-like pair: issue:45216 and issue:45357, both about Qwen3.5 `save_pretrained` producing incorrect checkpoint/visual-encoder keys. The other soft pairs are related at a broad subsystem level but describe different defects, so they should remain rejected.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45216 is the best canonical issue because it is a concrete, specific regression report with a clear failure mode and version window, and 45357 reads like a later follow-up on the same Qwen3.5 `save_pretrained` bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the best single issue to anchor the cluster because it names the affected family, version, and regression class clearly and has the widest apparent scope among the items.", + "best_issue_reason": "45216 is the strongest representative issue: it names the exact API, affected model family, and regression behavior, and it has the only soft edge that looks like a true duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "Different failures: PEFT/local checkpoint loading vs mistral_common chat-template handling." + "reason": "Different failures: PEFT/local checkpoint loading vs mistral_common tokenizer/chat-template behavior." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Custom model init breakage and non-persistent buffer corruption are unrelated code paths." + "reason": "Both are v5 regressions, but one is custom model init and the other is non-persistent buffer corruption; not the same bug." }, { "left": "issue:43645", "right": "issue:46661", "accept": false, - "reason": "Both involve custom/remote-code models, but one is Jupyter initialization while the other is AutoModel config_class handling." + "reason": "Notebook custom-model initialization and remote-code `config_class` handling are separate code paths and symptoms." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Unrelated issues: noisy loading logs vs a tokenizer regex patch crash." + "reason": "Logging/noise during loading is unrelated to the Mistral regex patch crash." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "FP8/Qwen3Moe loading failure and template-node compilation error are different bugs." + "reason": "Qwen3Moe FP8 failure and template compilation TypeError are different defects." }, { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:43792", + "right": "issue:46869", "accept": false, - "reason": "Different Qwen3.5 hidden-state bug vs Qwen2.5-VL rope-index scaling bug." + "reason": "Whisper loading failure and YOLO multi-GPU device-map failure are different models and failure modes." }, { - "left": "issue:43792", - "right": "issue:46869", + "left": "issue:44849", + "right": "issue:45325", "accept": false, - "reason": "Whisper loading/runtime failure and YOLO multi-GPU device-map failure are unrelated." + "reason": "Both are Qwen-family issues, but one is hidden-state output behavior and the other is rope/position-id scaling." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI padding-mask batching issue and DAC latent/forward mismatch are different model bugs." + "reason": "MIMI batching/padding-mask mismatch and DAC STE mismatch are unrelated model-specific logic bugs." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both touch tied weights, but one is a model-specific weight-tying bug and the other is serialization inconsistency; not the same concrete fix." + "reason": "Both involve tied weights, but one is untied weights in a model class and the other is device-dependent serialization." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to StaticCache sequence-length return type." + "reason": "Chunked generation inconsistency with compiled forward is not the same as `get_seq_length()` returning the wrong type." }, { - "left": "issue:43883", - "right": "issue:44991", + "left": "issue:28530", + "right": "issue:34567", "accept": false, - "reason": "Molmo tied-weights attribute error and tokenizer loading failure are unrelated." + "reason": "Early stopping metric selection and `num_input_tokens_seen` tracking are unrelated trainer issues." }, { - "left": "issue:28530", - "right": "issue:34567", + "left": "issue:44987", + "right": "issue:45081", "accept": false, - "reason": "Early stopping metric lookup and num_input_tokens_seen tracking are different trainer issues." + "reason": "A specific model load failure and a tokenizer regex patch crash are different problems." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Siglip2 wrong outputs and OmDet-Turbo processor/model size mismatch are different problems." + "reason": "Siglip2 inference/pipeline failure and OmDet-Turbo processor size mismatch are separate issues." }, { - "left": "issue:44987", - "right": "issue:45081", + "left": "issue:43883", + "right": "issue:44991", "accept": false, - "reason": "Both are loading regressions, but they affect different tokenizer/model code paths and different failure modes." + "reason": "Molmo tied-weights attribute error and tokenizer load failure for est-roberta are unrelated." }, { "left": "issue:44008", "right": "issue:46899", "accept": false, - "reason": "Gemma3n variable collision and Gemma4 4-bit dtype casting are unrelated multimodal bugs." + "reason": "Gemma3n variable collision and Gemma4 quantized multimodal dtype-cast failure are different bugs." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Qwen3Moe FP8 failure and init_empty_weights/_is_hf_initialized regression are different initialization issues." + "reason": "Qwen3Moe FP8 config failure and `init_empty_weights` `_is_hf_initialized` handling are distinct initialization issues." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT loading failure and Mistral regex patch crash are not the same bug." + "reason": "ScandiBERT loading failure and Mistral regex patch crash are unrelated tokenizer/model loading bugs." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer special-token conflict and MobileLLM key error are distinct load-time errors." + "reason": "T5 tokenizer special-token conflict and MobileLLM key error are different tokenizer/model issues." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken argument conflict and physical-intelligence/fast loading regression are unrelated." + "reason": "AddedToken construction error and physical-intelligence/fast loading failure are not the same bug." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export failure with torch_compilable_check is unrelated to processor image-size mismatch." + "reason": "torch.export exportability via `torch_compilable_check` is unrelated to processor input-size mismatch." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead and StaticCache return-type bug are different concerns." + "reason": "Generation performance overhead and cache-length return-type correctness are different issues." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-input ValueError are separate multimodal failures." + "reason": "Llava Next load segfault and Qwen3OmniMoe video-processing ValueError are separate multimodal failures." }, { "left": "issue:44479", "right": "issue:45325", - "accept": true, - "reason": "Both point to the same Qwen VL temporal-position/rope-index regression in v5.3.0; the latter is a narrower manifestation of the broader family bug." - }, - { - "left": "issue:42915", - "right": "issue:44589", "accept": false, - "reason": "Qwen3Moe FP8 failure and Float8 storage lookup error are different low-level issues." + "reason": "Video-input regression across Qwen models is not the same as still-image temporal position scaling in Qwen2.5-VL." }, { "left": "issue:43632", "right": "issue:46821", "accept": false, - "reason": "_is_hf_initialized regression and TimesFM window_size AttributeError do not share the same code path." - }, - { - "left": "issue:36032", - "right": "issue:41720", - "accept": false, - "reason": "T5 tokenizer special-token conflict and Qwen3 auto device mapping CUDA assert are unrelated." + "reason": "`_is_hf_initialized` breakage and TimesFM `window_size` AttributeError are unrelated." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding-window bug and granite_speech config strictness are unrelated." + "reason": "Qwen3-MoE sliding-window behavior and granite_speech config loading are different code paths." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer loading failure and convert_tokens_to_ids performance regression are not the same bug." + "reason": "Tokenizer load failure and `convert_tokens_to_ids` slowdown are unrelated." }, { "left": "issue:45216", "right": "issue:45357", - "accept": false, - "reason": "Both are Qwen3.5 save_pretrained regressions, but one is about checkpoint correctness generally and the other about visual encoder key serialization; too broad to merge." + "accept": true, + "reason": "Both describe the same Qwen3.5 `save_pretrained` regression producing incorrect checkpoint/visual-encoder keys; 45357 is a later follow-up on the same underlying bug." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative. The cluster is clearly heterogeneous, and the only accepted soft edge (44479 \u2194 45325) is well supported by the titles as the same Qwen VL temporal-position/regression family. No pair is overstated as a duplicate, and the summary does not claim stronger evidence than the packet shows." + "accept": false, + "feedback": "Reject: the response overstates the evidence in the summary. The packet only clearly supports one duplicate-like pair, issue:45216 \u2194 issue:45357. The claim that there are \"two soft matches\" that look like true duplicates is not grounded, because the other v5/custom-model pair is explicitly marked non-duplicate in the soft verdicts and the titles point to different defects. The rest of the pairwise rejections are conservative enough." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -3925,8 +3886,8 @@ "issue:43632|issue:44756", "issue:33666|issue:40444", "issue:40990|issue:44811", - "issue:43792|issue:44589", "issue:44291|issue:44589", + "issue:43792|issue:44589", "issue:42796|issue:46710", "issue:44079|issue:45663", "issue:35707|issue:40444", @@ -3938,226 +3899,226 @@ "issue:42175|issue:45542", "issue:41720|issue:45081", "issue:44079|issue:44534", - "issue:41553|issue:43582", "issue:43317|issue:43856", "issue:44462|issue:45812", + "issue:41553|issue:43582", "issue:43377|issue:45491", "issue:43611|issue:43646" ], - "prepared_review_unit_hash": "cdebe123f305d9bd2e0bd7e1d65e6934aea3bfae14d3aee4f2ee9a3c4b8641f7", + "prepared_review_unit_hash": "95b4d6ab86e3a94a107f98b9d4f813fdbb8faad1f048019034aaf2a3cd07d1a6", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36932, - "estimated_input_tokens": 9233, - "estimated_eval_tokens": 18722 + "serialized_chars": 36940, + "estimated_input_tokens": 9235, + "estimated_eval_tokens": 18726 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is heterogeneous: mostly unrelated model/tokenizer/runtime regressions across different model families and versions, with no clear duplicate hub and no PRs present.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "This cluster is mostly a loose set of unrelated issues, with a smaller subtheme around Transformers v5 loading/tokenizer regressions. I did not find any true duplicate pairs; the soft links are mostly superficial similarity.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43611 is the broadest and most central representative of the v5-loading regression subgroup, but the cluster overall does not contain a clear single duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "43611 is the best umbrella issue for the cluster\u2019s main theme because it describes a general Transformers v5 model-loading breakage rather than a narrow, model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen2.5-VL-related, but one is finetuning with iterable datasets and multi-image prompts while the other is a rope-index scaling bug for still images; different code paths and symptoms." + "reason": "Both are Qwen2.5-VL training/position-id issues, but one is about iterable datasets with multiple images and the other is about still-image temporal scaling; different bugs." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "Both involve Transformers v5 regressions, but one is model loading with base_model_prefix and the other is non-persistent buffers being initialized incorrectly; not the same failure." + "reason": "Both are Transformers v5 regressions, but one is about loading models with base_model_prefix and the other about non-persistent buffer initialization; distinct code paths." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "These are different compatibility regressions: Torch 2.10 incompatibility versus noisy load-time logging. Same release window, but not the same bug." + "reason": "One is a hard Torch 2.10 compatibility problem, the other is noisy loading output; not the same underlying bug." }, { "left": "issue:43716", "right": "issue:46899", "accept": false, - "reason": "Both concern Gemma/Mistral multimodal processing, but one is image-preprocessor dtype mismatch and the other is 4-bit quantization cast issues in multimodal embedders; distinct defects." + "reason": "Both involve multimodal Gemma/Mistral-style processing and dtype issues, but one is an image preprocessor/model dtype mismatch and the other is a 4-bit quantization cast bug in multimodal embedders." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "One is PaddleOCR-VL model loading in vLLM, the other is a sliding_window issue in Qwen3-MoE; unrelated models and code paths." + "reason": "Different model families and failure modes: PaddleOCR-VL loading in vLLM versus Qwen3-MoE sliding_window behavior." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "The first is a feature/extraction bug in _torch_extract_fbank_features(); the second is compiled flex_attention breaking on newer torch. Different subsystems." + "reason": "Audio feature extraction bug vs compiled flex_attention failure on newer Torch; no shared code-path evidence." }, { "left": "issue:45812", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but one affects Granite token IDs globally and the other is CodeLlama leading-space round-trip behavior; not the same tokenizer bug." + "reason": "Both are tokenizer regressions, but one is a silent wrong-ID issue for Granite models and the other is a leading-space round-trip bug in CodeLlama; different tokenizer logic." }, { "left": "issue:42796", "right": "issue:43746", "accept": false, - "reason": "Ministral tokenizer BPE-marker handling and GraniteSpeech PEFT-local-checkpoint loading are separate issues in different components." + "reason": "Tokenizer BPE-marker handling and PEFT adapter checkpoint loading are unrelated bugs." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing in eval mode is a general training-mode misuse bug, while the other is a missing supports_gradient_checkpointing flag on a specific trust_remote_code model." + "reason": "Gradient checkpointing in eval mode is not the same as a missing supports_gradient_checkpointing flag, even if both touch checkpointing APIs." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Unnecessary load-time noise is a logging/UX issue; the other is tokenizer loading failing for a specific repo on Transformers >=5.0. Different symptoms and fixes." + "reason": "One is general loading noise, the other is tokenizer.json being ignored; not the same failure." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "One is the _is_hf_initialized flag regression in Transformers v5; the other is mmap causing OOM on Strix Halo. Different root causes." + "reason": "Both are loading/runtime regressions, but one concerns _is_hf_initialized and the other mmap OOM on Strix Halo; different problems." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Qwen2-VL multi-GPU training and Qwen2.5-VL multi-image iterable-dataset finetuning are both training topics but not the same bug or code path." + "reason": "Both mention Qwen2-VL/Qwen2.5-VL fine-tuning, but one is multi-GPU training and the other is an iterable dataset/multi-image failure." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "WikiText-2 perplexity on gpt-oss-20b and Whisper processor.batch_decode skipping special-tokens are unrelated evaluation/decoding bugs." + "reason": "Perplexity evaluation issue vs Whisper batch_decode skip_special_tokens handling; unrelated." }, { - "left": "issue:43792", + "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "openai/whisper-large-v2 runtime failure and Float8 storage lookup failure are different backend/runtime issues." + "reason": "Both are Transformers v5-related instantiation failures, but one is about an unexpected _is_hf_initialized argument and the other about missing Float8 storage support; not mergeable as one fix." }, { - "left": "issue:44291", + "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "init_empty_weights with unexpected _is_hf_initialized argument is a model-init regression; Float8 storage object lookup is a storage/backend type issue." + "reason": "Whisper model loading failure and Float8 storage lookup error are different code paths and symptoms." }, { "left": "issue:42796", "right": "issue:46710", "accept": false, - "reason": "Tokenizer BPE-marker handling for Ministral is not the same as the DeepSeek-R1-Distill-Llama accuracy regression after upgrading Transformers." + "reason": "Tokenizer marker handling for Ministral vs accuracy regression on DeepSeek after upgrade; no evidence of the same bug." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated model-internals bugs." + "reason": "ModelOutput key assignment bug and Gemma4 FSDP2 shared_kv_states issue are unrelated." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values and multi-image iterable-dataset finetuning are different generation vs training problems." + "reason": "Both are generation/training-related Qwen issues only at a very broad level; one is inputs_embeds/past_key_values generation, the other is multi-image finetuning." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Qwen2.5-GGUF loading failure on Transformers v5 and Llama4Config missing pad_token_id are separate model/config loading regressions." + "reason": "Qwen2.5-GGUF loading failure and missing pad_token_id on Llama4Config are separate model/config issues." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Both are Transformers v5 regressions, but one concerns initialization flags and the other non-persistent buffers; not mergeable into one fix." + "reason": "Both are Transformers v5 regressions, but the failing mechanisms are different: initialization flag handling versus non-persistent buffer filling." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights/_is_hf_initialized errors affect different model/runtime paths." + "reason": "Qwen3-MoE sliding_window behavior and _is_hf_initialized constructor failure are unrelated." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime special-token post-processor updates and add_special_tokens failing for mdeberta-v3 are tokenizer issues, but the specific bug mechanics differ." + "reason": "Runtime special-token post-processing updates and add_special_tokens BOS/EOS insertion are different tokenizer behaviors." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading with Transformers v5 and CLIPTokenizer max_length misconfiguration are unrelated loading/tokenizer issues." + "reason": "Qwen2.5-GGUF loading failure and CLIPTokenizer model_max_length behavior are unrelated." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "TensorFlow missing from a torch-only install and undefined tf backend after installing tensorboard are packaging/dependency issues, but not the same packaging bug." + "reason": "Backend installation packaging issue vs missing TensorFlow backend detection; not the same bug." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and Mistral regex patch crashing with backend_tokenizer missing are different model-specific failures." + "reason": "Auto device mapping cudaErrorAssert and Mistral regex patch crashing are different errors." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and v5 buffer initialization bugs are separate core-library regressions." - }, - { - "left": "issue:41553", - "right": "issue:43582", - "accept": false, - "reason": "AutoTokenizer error message for Voxtral and AppleSilicon caching_allocator_warmup TypeError are unrelated; one is messaging, the other is platform/runtime." + "reason": "ModelOutput key handling and v5 buffer initialization are distinct framework regressions." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "gpu+cpu offload dequantized-model loading and Qwen3 MoE training memory inefficiency are both memory-related, but one is a loading/offload bug and the other is a training-memory optimization issue." + "reason": "device_map/offload loading failure and Qwen3 MoE memory inefficiency are related only loosely by model placement/memory, not the same bug." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Ignoring tokenizer.json in a repo and Granite token-ID regression are different AutoTokenizer failures with distinct consequences." + "reason": "Ignoring tokenizer.json in AutoTokenizer and wrong token IDs for Granite are different tokenizer-loading problems." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and Apple Silicon TypeError in caching_allocator_warmup are unrelated." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI Encoder batched-vs-single mismatch from missing padding masks and Gemma3 NaNs from sliding-window/all-padding windows are different attention/masking bugs." + "reason": "MIMI encoder padding-mask batching bug and Gemma3 NaN embeddings from sliding-window all-padding windows are both batching issues but not the same code path." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers v5 breakages, but one is base_model_prefix loading and the other is custom model initialization; similar era, different concrete code paths." + "reason": "Both are broad Transformers v5 breakages, but one is about loading models with base_model_prefix and the other about custom model initialization; separate fixes." } ] }, @@ -4233,8 +4194,8 @@ "issue:43296|issue:46821", "issue:44509|issue:46909", "issue:43644|issue:44079", - "issue:44829|issue:45479", "issue:42175|issue:43296", + "issue:44829|issue:45479", "issue:43818|issue:46402", "issue:43818|issue:45357", "issue:44265|issue:45072", @@ -4262,14 +4223,14 @@ "issue:41720|issue:45823", "issue:43377|issue:44610" ], - "prepared_review_unit_hash": "a210d5efa66a3ed4d3c91412d9626ca08eedb0d1555d71e5d929f5a93b64b086", + "prepared_review_unit_hash": "c62d8bfaca12a80034e825255b51ebc907ce37dc0ff6cbbe3fc44e2357e8c485", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, @@ -4284,175 +4245,208 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of unrelated bugs/regressions across different models, features, and docs. None of the soft pairs looks like the same underlying issue, so I reject all candidate duplicate links.", - "confidence": 0.89, - "canonical_issue_reason": "No single canonical issue: the items span distinct models and subsystems (audio processors, vision-language models, quantization/offloading, tokenization, docs) without one shared concrete code path.", + "summary": "Mostly unrelated issues; the only clear duplicate-looking pair is the SigLIP2 report pair, with 43493 as the better canonical report.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43493 is the more root-cause-oriented SigLIP2 report, describing the HF vs JAX implementation discrepancy directly.", "canonical_pr_reason": null, - "best_issue_reason": "No clear global best issue; the cluster is not a true duplicate set and has no single representative bug that subsumes the rest.", + "best_issue_reason": "issue:43493 best represents the duplicate pair because it states the underlying model mismatch more generally, while 43994 reads like a downstream symptom in AutoModel/pipeline.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:46821", "accept": false, - "reason": "Different failures in different areas: PaddleOCR-VL model loading in vLLM vs a TimesFM `window_size` attribute error." + "reason": "Audio processor chat-template sample-rate defaulting is unrelated to a TimesFM window_size AttributeError." }, { "left": "issue:44509", "right": "issue:46909", "accept": false, - "reason": "Both are docs-related, but they refer to different broken examples and different model families; not the same underlying bug." + "reason": "Both are documentation-related, but they concern different docs problems and different fixes." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Unrelated core bugs: non-persistent buffer initialization junk vs `ModelOutput` key assignment behavior." + "reason": "Non-persistent buffer junk in v5 is a different bug from ModelOutput key assignment handling." }, { - "left": "issue:44829", - "right": "issue:45479", + "left": "issue:42175", + "right": "issue:43296", "accept": false, - "reason": "Both mention degenerate training/loss, but the root causes are different: FlashAttention 3 regression vs single-label classification loss handling." + "reason": "TensorFlow backend packaging and audio processor sampling-rate behavior are unrelated." }, { - "left": "issue:42175", - "right": "issue:43296", + "left": "issue:44829", + "right": "issue:45479", "accept": false, - "reason": "Package/dependency install issue vs a model loading/runtime failure; no shared bug path." + "reason": "Both affect training, but one is about flash_attention_3 degeneration and the other about single-label classification loss logic." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different multimodal model bugs: missing temporal attention/weight sharing in Video-LLaVA vs text encoder weights not loading in Sam3Model." + "reason": "A missing temporal-attention video tower is unrelated to Sam3Model text encoder weight loading." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Both touch multimodal weights, but one is Video-LLaVA architecture/weights and the other is Qwen3.5 save_pretrained key serialization; not the same fix." + "reason": "Different multimodal model families and different failures: Video-LLaVA architecture vs Qwen3.5 save_pretrained key handling." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "`torch.export` failure on torch_compilable_check is unrelated to bfloat16 dtype mismatches in inference." + "reason": "torch.export incompatibility with torch_compilable_check is unrelated to bf16 dtype mismatches in inference." }, { "left": "issue:43493", "right": "issue:43994", - "accept": false, - "reason": "Both are SigLIP2-related, but one is HF-vs-JAX implementation discrepancy and the other is nonsensical outputs from an AutoModel/pipeline usage or processor mismatch." + "accept": true, + "reason": "Both describe the same SigLIP2 implementation mismatch causing incorrect/nonsensical outputs compared with the reference behavior." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Different Qwen2.5-VL input-path bugs: iterable dataset with multiple images per prompt vs video vision position IDs." + "reason": "IterableDataset multi-image finetuning failure is not the same as Qwen2.5-VL video position-id regression." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Qwen2.5-VL RoPE/temporal position scaling bug vs Kosmos2.5 OCR long-input index error; different code paths and symptoms." + "reason": "Qwen2.5-VL rope-index/position-id scaling and Kosmos2.5 OCR index errors are different bugs." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL vision-position issues, but one is still-image temporal scaling and the other is video position IDs; not the same defect." + "reason": "Both involve Qwen2.5-VL position IDs, but one is still-image temporal scaling and the other is video-input position-ID correctness; not clearly the same defect." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Shared tie-weight/lm_head theme, but one is failure to tie weights causing bad generation and the other is inconsistent serialization depending on device." + "reason": "Tied lm_head weight serialization and weight tying causing garbage generation are related only loosely, not the same bug report." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Different model/loading problems: FineGrainedFP8Config failure for Qwen3Moe vs Qwen2.5-GGUF loading in transformers v5." + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to Qwen2.5-GGUF loading with transformers v5." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to tokenization changes across transformer versions." + "reason": "Quantization/offloading behavior and version-dependent tokenization are different issues." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Different config/init failures: unexpected `_is_hf_initialized` argument vs strict config rejection for `granite_speech`." + "reason": "Unexpected _is_hf_initialized argument in init_empty_weights is not the same as strict config rejection for granite_speech." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to Qwen2.5-VL position-id scaling." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to Qwen2.5-VL video vision_position_ids." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on GPT-OSS and Whisper empty-transcription failure are different model-specific bugs." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Bark CPU offload device mismatch is unrelated to torch.export failures on torch_compilable_check models." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio processor sample-rate defaults and GraniteSpeech PEFT checkpoint loading are different code paths." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Both are regressions, but one is a Qwen video-input regression and the other is a Gemma4 device-map forward mismatch; distinct bugs." + "reason": "A video-input regression in Qwen models is unrelated to Gemma4 PLE device mismatch during forward." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to Qwen3 auto device mapping causing CUDA assert." + "reason": "inputs_embeds/past_key_values generation issues are unrelated to Qwen3 auto device mapping cudaErrorAssert." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice preset handling and progressive generation with cache are unrelated." + "reason": "Bark voice_preset failure is unrelated to progressive generation with inputs_embeds and past_key_values." }, { "left": "issue:43828", "right": "issue:46821", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to TimesFM's `window_size` attribute error." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE has nothing to do with TimesFM window_size." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors and Qwen2.5-VL multi-image dataset training failure are different problems." + "reason": "SigLIP2 documentation example errors and Qwen2.5-VL multi-image finetuning failure are separate problems." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer NER/padding crash is unrelated to `torch.export` failures involving `torch_compilable_check`." + "reason": "LayoutLMv2Tokenizer padding/truncation crashes are unrelated to torch.export and torch_compilable_check." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision tied-weight loading/generation bug and Bamba torch.compile SDPA failure are different code paths." + "reason": "lm_head weight tying/serialization and Bamba torch.compile SDPA failure are different bugs." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Both are audio-model bugs, but one is CPU-offload device mismatch and the other is DAC latent/STE mismatch; not the same fix." + "reason": "Bark CPU offload device mismatch is unrelated to DAC.from_latents vs forward-pass mismatch." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert and Gemma4 PLE device mismatch are separate runtime failures." + "reason": "Qwen3 cudaErrorAssert under auto device mapping is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI padding-mask batching bug and OmDet-Turbo processor/model input-size mismatch are unrelated." + "reason": "MIMI padding-mask batching and OmDet-Turbo input-size mismatch are different model/processor bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: most pairs are correctly rejected as unrelated, and 43493/43994 is the only pair that plausibly reads as the same SigLIP2 bug with one report describing the implementation discrepancy and the other its downstream symptom. No clear overstatement stood out." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -4519,11 +4513,11 @@ "issue:46821" ], "soft_pairs": [ - "issue:43881|issue:44488", "issue:43257|issue:45923", "issue:44568|issue:45920", "issue:46305|issue:46315", "issue:42796|issue:46209", + "issue:43881|issue:44488", "issue:43454|issue:45072", "issue:45823|issue:46821", "issue:40990|issue:42947", @@ -4540,231 +4534,231 @@ "issue:41720|issue:44977", "issue:36331|issue:42915", "issue:44466|issue:45357", - "issue:43450|issue:45701", "issue:40990|issue:45701", + "issue:43450|issue:45701", "issue:44877|issue:44991", - "issue:43526|issue:43883", "issue:43065|issue:46402", "issue:26413|issue:28530", "issue:41720|issue:42617", - "issue:43881|issue:44991", "issue:44186|issue:45072", - "issue:43873|issue:44756" + "issue:43526|issue:43883", + "issue:43873|issue:44756", + "issue:43881|issue:44991" ], - "prepared_review_unit_hash": "1b15cb0ba066086c6c9f5d5e1514bde88d3c8c6568937359fee424d3738396e0", + "prepared_review_unit_hash": "db0497217f628b466c5af3a2defbf85360d53e5bb249e3b00813640bc554a004", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36965, - "estimated_input_tokens": 9242, - "estimated_eval_tokens": 18740 + "serialized_chars": 36963, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No clear duplicate cluster emerges here: the soft edges mostly connect issues that share a broad subsystem (tokenization, loading/offloading, quantization, or model-specific regressions) but do not appear to describe the same concrete bug. I would not merge any of the proposed pairs.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43873 is the best representative of the only loosely related theme in this set (offloading/quantization loading behavior): it is broad, active, and has the most discussion/inbound references among the candidates. That said, the overall set is not a true duplicate cluster.", + "summary": "The set is mostly unrelated issue reports across different models and subsystems. None of the soft pairs look like the same underlying bug with sufficient specificity to merge as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:43873 is the broadest and most central candidate in this set: it is open, has multiple inbound references, and covers a general offloading/quantization failure mode that could subsume nearby reports. That said, the rest of the items are mostly distinct bugs.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the strongest standalone issue in this set: it is open, has the most follow-up activity, and its title describes a concrete user-visible failure with a reasonably general repro surface.", + "best_issue_reason": "issue:43873 is the strongest representative issue because it is broader than the other reports, already attracting references, and is phrased as a general regression rather than a single model-specific failure.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43881", - "right": "issue:44488", - "accept": false, - "reason": "Both are load failures, but they target different models and likely different root causes; this is too broad to treat as the same bug." - }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Different model paths and failures: one is MoE weight conversion with accelerate/deepspeed, the other is a missing gradient-checkpointing flag on a trust_remote_code variant." + "reason": "Different bugs: one is MoE weight conversion with accelerate+deepspeed, the other is a missing gradient-checkpointing flag on a trust_remote_code variant." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both are tokenizer regressions, but one is about add_special_tokens/BOS-EOS handling for a specific tokenizer, while the other is wrong token IDs across multiple models; not the same concrete issue." + "reason": "Both are tokenizer regressions, but they affect different code paths and model families; not the same concrete bug." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "Completely different code paths: streaming window masking off-by-one versus tokenizer conversion performance." + "reason": "One is a sliding-window masking off-by-one bug, the other is a tokenizer lookup performance regression; unrelated failure modes." }, { "left": "issue:42796", "right": "issue:46209", "accept": false, - "reason": "Tokenizer BPE-marker handling and FP8 inference on B200 are unrelated failures." + "reason": "Tokenizer BPE marker handling and FP8 inference on B200 are unrelated issues in different components." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are model loading failures, but for different models and likely different causes; too broad to treat as the same bug." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Different problems in different subsystems: weight tying vs dtype mismatches during bfloat16 inference." + "reason": "Tie-weight serialization/generation corruption is unrelated to dtype mismatches in bfloat16 inference." }, { "left": "issue:45823", "right": "issue:46821", "accept": false, - "reason": "One is a device-map forward mismatch in Gemma4 PLE; the other is an AttributeError on a window_size argument in TimesFM 2.5." + "reason": "Device-mapping forward mismatch and a TimesFM window_size AttributeError are different code paths and symptoms." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity degradation on one model/data pair is unrelated to gradient checkpointing being ineffective with PEFT LoRA." + "reason": "High perplexity on a model and ineffective gradient checkpointing with LoRA are not the same underlying bug." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Both involve batching/padding behavior, but they affect different models and different mechanisms; not the same underlying bug." + "reason": "Batch-vs-single padding mask inconsistency in MIMI is not the same as Perceiver image classification failure at non-default resolution." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "A script execution failure and a FineGrainedFP8Config load failure are not the same issue." + "reason": "A script execution failure and a Qwen3 MoE FP8 loading issue are unrelated." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Weight tying for AyaVision is unrelated to packed-weight unpacking in BitNet accelerate loading." + "reason": "Different components and behaviors: tied-weight serialization vs packed-weight unpacking during accelerate loading." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "Both mention loading/serialization, but one is FSDP CPU RAM efficient loading and the other is non-persistent buffer initialization; different bugs." + "reason": "FSDP CPU RAM-efficient loading breakage and non-persistent buffer junk values are different loading regressions, not the same concrete bug." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "These are both offload/quantization-adjacent, but one is a dequantized model load failure with gpu+cpu offload and the other is a broader quantization offloading problem; not specific enough to merge." + "reason": "Both concern offloading/quantization, but one is specific to dequantized model loading with device_map=auto and the other is a broader offloading complaint; insufficient evidence they are the same bug." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "A missing pad_token_idx in StableLmConfig and failure to unwrap text_config in composite Qwen multimodal configs are unrelated config-loading bugs." + "reason": "Missing pad_token_idx in StableLmConfig is unrelated to failing to unwrap text_config for composite Qwen multimodal configs." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Segmentation fault when loading Llava Next and sinq save/load breakage are different failure modes and model families." + "reason": "Segfault loading Llava Next models and sinq save/load corruption are different failure modes." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and Llava Next loading segfault are different models and different code paths." + "reason": "CUDA assert on Qwen3 with auto device mapping and a Llava Next load segfault are unrelated and model-specific." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "A trainer signature mismatch and a 3d_parallel.py execution failure are unrelated." + "reason": "Trainer compute_loss API breakage is unrelated to not being able to run a parallelism example script." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "CUDA reserved-memory OOM under int4 quantization is not the same as the T5/apex FusedRMSNorm memory leak." + "reason": "Int4 quantization memory regression and an apex RMSNorm memory leak are different memory issues with different causes." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Both are Qwen-related, but one is an auto device-mapping cudaErrorAssert and the other is flash-attention generation behavior; different bugs." + "reason": "Auto device mapping CUDA assert and flash-attention generation issues in Qwen3.5 are distinct bugs." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Unexpected compute_loss keyword arguments and Qwen3Moe FP8 loading failure are unrelated." + "reason": "Unexpected compute_loss keyword argument and Qwen3 MoE FP8 failure are unrelated." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both are serialization regressions, but one concerns lm_head.weight tied-weights behavior and the other incorrect visual encoder keys; not the same fix." + "reason": "Both involve save_pretrained regressions, but one is lm_head serialization/tied weights and the other is incorrect visual encoder keys; different code paths." }, { - "left": "issue:43450", + "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Video processor batch-shape issues and tokenization changes across versions are unrelated." + "reason": "A perplexity regression does not match a broad tokenization-change report; too weak to call duplicates." }, { - "left": "issue:40990", + "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Perplexity regression on one model is not the same as version-dependent tokenization changes." + "reason": "Video processor batch-shape bug and tokenization changes are unrelated subsystems." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech and tokenizer loading for EMBEDDIA/est-roberta are separate regressions." - }, - { - "left": "issue:43526", - "right": "issue:43883", - "accept": false, - "reason": "BeitImageProcessorFast label reduction and missing all_tied_weights_keys on MolmoForCausalLM are unrelated." + "reason": "Strict config rejection for granite_speech and tokenizer loading failure for est-roberta are unrelated issues." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder and missing text encoder weights in Sam3Model are different issues in the same family but not the same concrete bug." + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3Model text encoder weights not loading are separate model components and bugs." }, { "left": "issue:26413", "right": "issue:28530", "accept": false, - "reason": "Checkpoint resume epoch_iterator failure and early-stopping metric resolution are different trainer behaviors." + "reason": "resume_from_checkpoint epoch_iterator failure and early-stopping metric lookup are different Trainer issues." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device mapping and inability to run 3d_parallel.py are unrelated." + "reason": "Auto device mapping CUDA assert and an example script execution failure do not describe the same bug." }, { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:44186", + "right": "issue:45072", "accept": false, - "reason": "Both are load-related, but one is a model loading failure and the other is tokenizer loading regression; too broad to merge." + "reason": "Tokenizer crashes on NER/batched padding are unrelated to dtype mismatches in bfloat16 inference." }, { - "left": "issue:44186", - "right": "issue:45072", + "left": "issue:43526", + "right": "issue:43883", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/batching and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." + "reason": "BeitImageProcessorFast label reduction and missing all_tied_weights_keys on Molmo are unrelated." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization offloading problems and mmap OOM on Strix Halo are not the same underlying bug." + "reason": "Both are memory/offload themed, but one is a quantization/offloading regression and the other is a Strix Halo mmap OOM workaround; not the same concrete bug." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "A model loading failure and a tokenizer loading failure for a different model are unrelated." } ] }, @@ -4843,8 +4837,8 @@ "issue:43295|issue:45701", "issue:45685|issue:46821", "issue:44186|issue:44610", - "issue:43550|issue:43606", "issue:43646|issue:46665", + "issue:43550|issue:43606", "issue:43404|issue:45357", "issue:36010|issue:42617", "issue:44610|issue:44898", @@ -4862,153 +4856,153 @@ "issue:42915|issue:46291", "issue:36331|issue:41720", "issue:44186|issue:44361", - "issue:43316|issue:44913", "issue:43425|issue:46821", + "issue:43316|issue:44913", "issue:42915|issue:43322", "issue:28530|issue:30064" ], - "prepared_review_unit_hash": "28ba5c9207479ab5a1dc62396b83d00ea192e2699718bd43e01dff7c03daf0c2", + "prepared_review_unit_hash": "18fc38277d63ede365838fd87239f305c254bd9f420318b64d2950c063cb541f", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35918, - "estimated_input_tokens": 8980, - "estimated_eval_tokens": 18216 + "serialized_chars": 35916, + "estimated_input_tokens": 8979, + "estimated_eval_tokens": 18214 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No true duplicate cluster is present: the soft pairs span unrelated regressions across generation, tokenizers, model loading, training, and serving. Most overlaps are only superficial (same broad subsystem, version regression, or cache/device language) and do not indicate the same underlying bug.", - "confidence": 0.94, - "canonical_issue_reason": "issue:45701 is the broadest, most cross-cutting user-facing regression in the set and has the strongest external signal (inbound references), so it is the best umbrella representative if one issue must be chosen.", + "summary": "This cluster is highly heterogeneous: the issues span unrelated training, tokenizer/processor, model-loading, device-map, serving, and cache bugs. None of the proposed soft pairs look like the same underlying defect, so no soft edges are accepted. There are no PRs in the cluster.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45701 is the best representative because it is the most broadly phrased and has the highest external reference signal, making it the closest thing to a hub in an otherwise mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the most general and highest-signal issue here; it is the strongest candidate for a representative triage target, even though the set is not actually a duplicate cluster.", + "best_issue_reason": "issue:45701 is the best issue candidate for canonicalization, but only as a broad anchor; the cluster does not actually contain a tight duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46612", "right": "issue:46858", "accept": false, - "reason": "Different failures: beam-search cache reorder omission for legacy models vs static-cache slice arithmetic on non-CUDA backends." + "reason": "Both involve cache/generation, but one is a model-family beam-search reorder omission and the other is a static-cache seq-length arithmetic bug on non-CUDA backends." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "One is Bamba torch.compile + SDPA behavior; the other is BitNet packed-weight unpacking during accelerate loading." + "reason": "Different code paths: Bamba torch.compile + SDPA versus BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Unrelated issues: dtype propagation in Blip2 loading vs Trainer checkpoint-resume max_steps math." + "reason": "Blip2 dtype propagation and Trainer resume/max_steps calculation are unrelated bugs." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Both are version regressions, but one is processor.tokenizer/image handling and the other is tokenization behavior changes; not the same bug." + "reason": "One is a regression around processor.tokenizer/images passthrough; the other is a broad tokenization-version discrepancy. Related theme, not the same bug." }, { "left": "issue:45685", "right": "issue:46821", "accept": false, - "reason": "MPS histogram on Int in MoE code is unrelated to TimesFM's missing window_size attribute." + "reason": "Completely different models and failures: MoE MPS histogram support versus TimesFM window_size AttributeError." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer padding/NER crash in LayoutLMv2 is not the same as OmDet-Turbo processor/model image-size mismatch." + "reason": "Both are processor/tokenizer bugs, but they affect different models and different failure modes: NER/batched padding versus image size mismatch." }, { - "left": "issue:43550", - "right": "issue:43606", + "left": "issue:43646", + "right": "issue:46665", "accept": false, - "reason": "Different device-path problems: torch.compile/SDPA on Bamba vs CPU offload mismatch in bark-small." + "reason": "Custom model initialization breakage in v5 is not the same as inconsistent from_pretrained behavior inside Python classes." }, { - "left": "issue:43646", - "right": "issue:46665", + "left": "issue:43550", + "right": "issue:43606", "accept": false, - "reason": "Both concern loading custom models, but the concrete failures differ: v5 custom init breakage vs from_pretrained behavior inside Python classes." + "reason": "Different model-specific device/offload failures; no shared concrete code path." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "lm_head tying in Mistral3 is unrelated to Qwen3.5 save_pretrained writing wrong visual encoder keys." + "reason": "Weight tying in Mistral3 conditional generation is unrelated to Qwen3.5 visual encoder key saving." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import failure and 3d_parallel.py execution failure are unrelated." + "reason": "Importing GenerationMixin and running 3d_parallel.py are unrelated issues." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both mention resolution/size mismatches, but they affect different models and code paths; not one underlying defect." + "reason": "Both are vision input-shape problems, but they involve different models and different mechanisms: processor resize mismatch versus interpolate_pos_encoding handling." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "FineGrainedFP8Config failure in Qwen3Moe is separate from dequantized-model loading with gpu+cpu offload." + "reason": "Qwen3Moe FineGrainedFP8Config failure and dequantized model offload loading failure are different model-loading bugs." }, { "left": "issue:44734", "right": "issue:46858", "accept": false, - "reason": "Both involve cache/generation internals, but one is wrong tensor indexing in serve responses and the other is device-tensor seq length in static-cache prefill." + "reason": "Both touch cache logic, but one is serve-side continuation tensor indexing and the other is prefill static-cache length arithmetic; not the same bug." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both hit serving multimodal paths, but one is a string-to-device conversion crash and the other is a missing _tokenizer attribute on Gemma4Processor." + "reason": "Both are serve/multimodal-related, but the actual failures differ: string-to-device conversion versus missing _tokenizer on Gemma4Processor." }, { "left": "issue:46821", "right": "issue:46869", "accept": false, - "reason": "TimesFM window_size attribute error is unrelated to YOLO multi-GPU device-map failures." + "reason": "TimesFM window_size AttributeError and YOLO multi-GPU device_map failure are unrelated." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma3n variable-name collision in forward is unrelated to Gemma4 PLE device mismatch under device_map=auto." + "reason": "Gemma3n forward variable collision is unrelated to Gemma4 PLE device mismatch under device_map=auto." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and GPT-OSS MoE tensor-parallel failure are different model/runtime issues." + "reason": "Different models and symptoms: Qwen3 auto device mapping cudaErrorAssert versus GPT-OSS MoE tensor-parallel failure." }, { "left": "issue:28530", "right": "issue:33666", "accept": false, - "reason": "Early-stopping metric selection and Qwen2-VL multi-GPU training are unrelated." + "reason": "Early stopping/eval metric selection and Qwen2-VL multi-GPU training are not the same defect." }, { "left": "issue:44464", "right": "issue:46858", "accept": false, - "reason": "Both mention compiled/cache generation, but one is inconsistent chunked generation outputs and the other is a static-cache seq-length tensor bug." + "reason": "Chunked generation inconsistency with compiled forward is a different bug from static-cache seq-length arithmetic on non-CUDA backends." }, { "left": "issue:45632", @@ -5020,55 +5014,55 @@ "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are different model bugs." + "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are separate model-specific bugs." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE has nothing to do with StableLmConfig missing pad_token_idx." + "reason": "Sliding-window behavior in Qwen3-MoE and missing pad_token_idx in StableLmConfig are unrelated." }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "FineGrainedFP8Config failure is unrelated to kernels v0.15 import/revision requirement changes." + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to kernels package version/revision import requirements." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer compute_loss signature breakage is unrelated to Qwen3 device-map CUDA assertion." + "reason": "Trainer compute_loss signature change and Qwen3 device-map cudaErrorAssert are different layers of the stack." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Different tokenizer failures affecting different tasks and tokenizers; not the same root cause." + "reason": "Both tokenizer-related, but the models and concrete errors differ; they are not the same underlying bug." }, { - "left": "issue:43316", - "right": "issue:44913", + "left": "issue:43425", + "right": "issue:46821", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is unrelated to GPTNeoX rotary_pct reload behavior." + "reason": "Torch 2.10 compatibility and TimesFM window_size AttributeError do not share a concrete cause." }, { - "left": "issue:43425", - "right": "issue:46821", + "left": "issue:43316", + "right": "issue:44913", "accept": false, - "reason": "Torch 2.10 compatibility and TimesFM window_size AttributeError are unrelated." + "reason": "Both concern config API behavior, but for different configs and unrelated fields." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "FineGrainedFP8Config failure and Llava Next segmentation fault are separate loading problems." + "reason": "Qwen3Moe FP8 loading and Llava Next segmentation fault are distinct model-loading failures." }, { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early stopping metric lookup and void segmentation-map processing are unrelated." + "reason": "Early stopping metric resolution and void segmentation map processing are unrelated." } ] }, @@ -5131,7 +5125,6 @@ "issue:45357", "issue:45406", "issue:45464", - "issue:45479", "issue:45701", "issue:45704", "issue:45759", @@ -5143,16 +5136,16 @@ ], "soft_pairs": [ "issue:33357|issue:35707", - "issue:43526|issue:43881", "issue:43540|issue:43782", "issue:29159|issue:36032", "issue:44387|issue:45701", "issue:43873|issue:45704", "issue:30064|issue:34634", + "issue:43526|issue:43881", "issue:43550|issue:43653", "issue:44534|issue:46665", - "issue:43606|issue:43720", "issue:43531|issue:45759", + "issue:43606|issue:43720", "issue:41720|issue:43746", "issue:45406|issue:45464", "issue:44492|issue:44509", @@ -5170,217 +5163,210 @@ "issue:44077|issue:44625", "issue:43644|issue:46665", "issue:44062|issue:44451", - "issue:44898|issue:45072", - "issue:44625|issue:45479" + "issue:44898|issue:45072" ], - "prepared_review_unit_hash": "3f4ba4d861676d93a3d3d789fb8fc77bcae042bba577788671819053d6d076db", + "prepared_review_unit_hash": "5fb21a95aca3ffd9915d2ddf24c2ef7beb374675a9d8f9a4f2ff171a48246fb1", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36829, - "estimated_input_tokens": 9208, - "estimated_eval_tokens": 18672 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35883, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly over-grouped by title similarity: the soft pairs span unrelated bugs across tokenizers, processors, loading/offloading, multimodal models, and docs. I found no soft-edge duplicates to merge; the only obvious duplicate theme in the broader set is the non-persistent-buffer regression (43644/44534), which is not one of the provided soft pairs.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue for the whole cluster: the items are heterogeneous and do not share one underlying bug. If forced to anchor the only clear duplicate subtheme, 44534 is the broader wording of the non-persistent-buffer regression.", + "summary": "This cluster is mostly a loose similarity set of unrelated issue reports, with only a few near-duplicate regressions around tokenizer/model-loading/v5 behavior. None of the soft pairs look like the exact same underlying bug, so all are rejected.", + "confidence": 0.63, + "canonical_issue_reason": "Issue 44534 is the clearest broad regression report in the set and is the best representative of the closest duplicate-like theme here (the non-persistent-buffer v5 bug).", "canonical_pr_reason": null, - "best_issue_reason": "44534 is the best representative available because it states a concrete, regression-style bug clearly and broadly; it is also the cleaner umbrella for the buffer-corruption subtheme compared with 43644.", + "best_issue_reason": "Issue 44534 is the most suitable anchor for the cluster because its title is broad, specific, and representative of a concrete regression rather than a model- or task-specific edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "Different failure modes: MacOS CLIP bus error vs generation with inputs_embeds/past_key_values." - }, - { - "left": "issue:43526", - "right": "issue:43881", - "accept": false, - "reason": "Unrelated subsystems: image label reduction vs glm-4v loading failure." + "reason": "Different bugs: MacOS bus error in a CLIP model versus generation behavior with inputs_embeds/past_key_values." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Different bugs: Qwen3OmniMoe video processing error vs Qwen3VL weight_only load error." + "reason": "Both involve Qwen multimodal models, but one is a video-processing ValueError and the other is a from_pretrained weight_only loading error." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Both tokenizer-related, but one is slow/fast inconsistency and the other is a T5 add_special_tokens constructor conflict." + "reason": "Tokenizer-related, but not the same bug: slow/fast inconsistency versus a T5Tokenizer load-time AddedToken conflict." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization OOM is not the same as version-dependent tokenization changes." + "reason": "Unrelated symptoms: int4 quantization OOM versus tokenizer behavior changes across versions." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Offloading/quantization issue vs T5 apex RMSNorm memory leak; different code paths and symptoms." + "reason": "Different concrete problems: quantization/offloading behavior versus an apex RMSNorm memory leak." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Void segmentation-map handling is unrelated to BarkProcessor voice_preset loading." + "reason": "Both are processor bugs, but they affect different processors and failure modes." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated: BeitImageProcessorFast label reduction bug versus a GLM-4V model loading failure." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "SDPA/torch.compile failure is unrelated to BigBirdTokenizer special-token decode behavior." + "reason": "Different code paths: torch.compile/SDPA failure versus tokenizer special-token registration." }, { "left": "issue:44534", "right": "issue:46665", "accept": false, - "reason": "Both mention loading patterns, but one is a buffer initialization regression and the other is from_pretrained behavior inside Python classes." + "reason": "Both mention from_pretrained/v5-era behavior, but one is a buffer-initialization regression and the other is inconsistent behavior inside Python wrapper classes." }, { - "left": "issue:43606", - "right": "issue:43720", + "left": "issue:43531", + "right": "issue:45759", "accept": false, - "reason": "CPU-offload device mismatch is not the same as packed-weight unpacking during accelerate loading." + "reason": "Both are Qwen-family issues, but one is a sliding_window bug and the other is config unwrapping for composite multimodal models." }, { - "left": "issue:43531", - "right": "issue:45759", + "left": "issue:43606", + "right": "issue:43720", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and composite-config unwrap bug are different issues." + "reason": "Accelerate/loading-related, but one is CPU offload device mismatch and the other is packed-weight unpacking for BitNet." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 device-map CUDA assert is unrelated to GraniteSpeech PEFT local checkpoint loading." + "reason": "Completely different: Qwen3 device-mapping CUDA assert versus GraniteSpeech local checkpoint loading with PEFT adapters." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer is a different bug from Qwen3.5 streaming chat/completions failure." + "reason": "Both hit serving/API paths, but the reported failures are different: missing _tokenizer versus streaming inference failure." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A cache-strategy typo and docs mentioning removed pipeline tasks are separate documentation issues." + "reason": "A typo in cache strategy docs is unrelated to stale pipeline-task documentation." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash local model warning is unrelated to an undefined-name bug in the video branch." + "reason": "Different issues: local path trailing-slash import warning versus a video multimodal token-counting bug." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Last-batch label truncation in gather_for_metrics is not the same as SAM3 text/bounding-box behavior." + "reason": "Different subsystems: gather_for_metrics label truncation versus SAM3 text/box prompting behavior." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "CPU offload device mismatch is unrelated to tokenizer special-token registration and decode output." + "reason": "No shared underlying bug; device mismatch during offload is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token accounting and progressive generation are different features with different bugs." + "reason": "TrainerState token counting is unrelated to progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice preset failure is unrelated to Qwen3 auto device-mapping CUDA asserts." + "reason": "Different model families and failures: BarkProcessor voice preset versus Qwen3 auto device mapping CUDA assert." }, { "left": "issue:43531", "right": "issue:46821", "accept": false, - "reason": "Qwen3-MoE sliding_window and TimesFM window_size attribute errors are separate model-specific issues." + "reason": "Both involve model-specific arguments, but sliding_window on Qwen3-MoE is unrelated to TimesFM window_size AttributeError." }, { "left": "issue:43540", "right": "issue:46858", "accept": false, - "reason": "Video-input validation in Qwen3OmniMoe is not the same as cache sequence-length arithmetic on non-CUDA backends." + "reason": "Different code paths: Qwen3OmniMoe video input processing versus cache seq_length tensor arithmetic in _prefill." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Both involve Qwen3.5 components, but one is config unwrapping and the other is rotary embedding CPU-offload compatibility." + "reason": "Both are Qwen 3.5-related, but one is config unwrapping and the other is a CPU-offload incompatibility in rotary embedding." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer-loading failures for different models; no shared underlying bug is evident." + "reason": "Both are tokenizer/model-loading failures, but they are different model-specific regressions and not the same concrete bug." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Sliding_window behavior and missing Float8 storage object are unrelated." + "reason": "Sliding-window handling for Qwen3-MoE is unrelated to a Float8 storage lookup error." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Incorrect visual-encoder save keys and missing text-encoder weight loading are different model-load/save bugs." + "reason": "Both concern multimodal model save/load, but they affect different models and different directions (saving visual keys vs loading text encoder weights)." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Optional post_init policy for patchtsmixer is unrelated to num_labels propagation." + "reason": "Different issues: an optional post_init policy problem versus Qwen3.5 num_labels propagation." }, { "left": "issue:43644", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer initialization junk and from_pretrained behavior inside Python classes are different problems." + "reason": "Non-persistent-buffer junk values and inconsistent from_pretrained behavior inside Python classes are not the same bug." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "AddedToken keyword conflict is unrelated to loading EMBEDDIA/est-roberta." + "reason": "A tokenizer AddedToken argument conflict is not the same as ScandiBERT loading failure." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Perceiver non-default-resolution failure and bfloat16 dtype mismatch are distinct runtime issues." - }, - { - "left": "issue:44625", - "right": "issue:45479", - "accept": false, - "reason": "num_labels propagation and degenerate zero loss are separate classification bugs." + "reason": "Different model behaviors: Perceiver image classification with interpolation versus bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." } ] }, @@ -5409,7 +5395,6 @@ "issue:39692", "issue:41720", "issue:42886", - "issue:43295", "issue:43381", "issue:43421", "issue:43493", @@ -5427,6 +5412,7 @@ "issue:44514", "issue:44545", "issue:44610", + "issue:44625", "issue:44671", "issue:44843", "issue:44918", @@ -5437,10 +5423,10 @@ "issue:45356", "issue:45357", "issue:45381", + "issue:45479", "issue:45507", "issue:45704", "issue:45759", - "issue:45834", "issue:45907", "issue:45923", "issue:45987", @@ -5455,9 +5441,10 @@ ], "soft_pairs": [ "issue:44987|issue:45042", + "issue:44625|issue:45479", "issue:43381|issue:43844", - "issue:44368|issue:45357", "issue:39692|issue:44610", + "issue:44368|issue:45357", "issue:41720|issue:45704", "issue:43746|issue:46050", "issue:43381|issue:45704", @@ -5483,229 +5470,228 @@ "issue:43720|issue:43819", "issue:46829|issue:46832", "issue:43653|issue:43720", - "issue:45987|issue:46829", - "issue:43295|issue:45834" + "issue:45987|issue:46829" ], - "prepared_review_unit_hash": "1a08127df6847b0f48bf0b8aebb103b80a00d40fb803b09b9092b8bb7e368eb2", + "prepared_review_unit_hash": "6ec937c3f12a790d13aeb8df5eb0f6393853a04860f83867a577eb24cc0cd530", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37455, - "estimated_input_tokens": 9364, - "estimated_eval_tokens": 18984 + "serialized_chars": 37509, + "estimated_input_tokens": 9378, + "estimated_eval_tokens": 19012 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose collection of unrelated Transformers bugs, with two duplicate pockets standing out: the Qwen2_5_VL batched apply_chat_template crash and an offline tokenizer-loading regression tied to _patch_mistral_regex/model_info().", - "confidence": 0.83, - "canonical_issue_reason": "Issue 44545 is the clearest representative for the Qwen2_5_VL batched crash: it states the exact failure mode succinctly and matches 44514 one-to-one.", + "summary": "This cluster is mostly unrelated transformer/model bugs that should not be merged; only the Qwen2_5_VLProcessor batched `padding=False` reports are true duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:44545 is an exact duplicate of issue:44514: same `Qwen2_5_VLProcessor.apply_chat_template` crash on batched input with `padding=False`, and it has slightly more discussion/reference activity.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44545 is the best overall representative because it is a clear, reproducible, user-facing crash report and an exact duplicate of 44514.", + "best_issue_reason": "issue:44545 is the best representative issue because it is the clearest same-bug duplicate pair in the set and captures the concrete code-path failure precisely.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Both are model-loading problems, but one is about a specific model package and the other is a general PIL/torchvision requirement regression; not the same bug." + "reason": "Different bugs: one is loading `physical-intelligence/fast`, the other is PIL image processors incorrectly requiring `torchvision`." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:44625", + "right": "issue:45479", "accept": false, - "reason": "Different failure modes: eval-mode checkpointing misuse versus a ZeRO-3 gradient blow-up during training." + "reason": "Both concern classification configs, but one is `num_labels` propagation in Qwen3.5 and the other is a generic zero-loss bug for `num_labels=1`." }, { - "left": "issue:44368", - "right": "issue:45357", + "left": "issue:43381", + "right": "issue:43844", "accept": false, - "reason": "One is a warning about tie_word_embeddings in LoRA fine-tuning; the other is a save_pretrained regression for visual encoder keys." + "reason": "Different failure modes: gradient checkpointing in eval mode vs gradient explosion with HfDeepSpeedConfig + ZeRO-3." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Different multimodal model issues: SigLIP2 doc/example errors versus OmDet-Turbo image size mismatch." + "reason": "SigLIP2 docs/example errors are unrelated to OmDet-Turbo's processor/model image-size mismatch." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Unrelated: a `tie_word_embeddings` warning during LoRA fine-tuning vs incorrect visual encoder keys saved for Qwen3.5." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Unrelated subsystems and symptoms: Qwen3 device-mapping assert versus T5 RMSNorm memory leak." + "reason": "Different code paths: Qwen3 auto device mapping on A800 vs T5 silently using apex RMSNorm with a memory leak." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Both involve loading/saving, but one is PEFT local checkpoint loading and the other is quantized sinq round-tripping; not the same bug." + "reason": "One is PEFT adapters/local checkpoint loading for GraniteSpeech; the other is quantized model save/load corruption with sinq." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to the T5 apex RMSNorm leak." + "reason": "Eval-mode gradient checkpointing and T5 apex RMSNorm leak are unrelated bugs." }, { "left": "issue:46655", "right": "issue:46693", "accept": false, - "reason": "Video padding calculation in a processor is unrelated to FA2 long-context performance regression." + "reason": "SmolVLM video padding calculation bug is unrelated to a FA2 long-context performance regression." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Special-token post-processor updates and DiaConfig custom token IDs are different tokenizer/state bugs." + "reason": "Both touch tokenizer metadata, but one is runtime post-processor updates and the other is DiaConfig losing custom token IDs on save/load." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Different issues in different model families: RMSNorm leak versus missing gradient-checkpointing flag." + "reason": "T5 RMSNorm memory leak and missing `supports_gradient_checkpointing` on Nemotron are distinct issues." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "TRl SFT embedding unpacking and trainer.evaluate() after train() are distinct training bugs." + "reason": "Different failures: Qwen3.5 embedding unpacking with TRL SFT trainer vs `trainer.evaluate()` after `trainer.train()`." }, { "left": "issue:42886", "right": "issue:44843", - "accept": true, - "reason": "Same offline-tokenizer failure: 44843 pinpoints the unconditional model_info() call in _patch_mistral_regex that breaks HF_HUB_OFFLINE, matching 42886's cache-load failure." + "accept": false, + "reason": "Both involve offline/tokenizer behavior, but one is cache loading offline and the other is an unconditional `model_info()` call in `_patch_mistral_regex`." }, { "left": "issue:45759", "right": "issue:45923", "accept": false, - "reason": "Composite multimodal config unwrapping and a missing supports_gradient_checkpointing flag are different issues." + "reason": "Different problems: unwrapping `text_config` in composite Qwen configs vs a missing gradient-checkpointing capability flag." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT checkpoint loading and quantization/offloading behavior are not the same underlying bug." + "reason": "PEFT/local checkpoint loading is unrelated to quantization offloading behavior." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Exact duplicate: both report Qwen2_5_VLProcessor.apply_chat_template crashing on batched input when padding=False." + "reason": "Exact duplicate reports of the same `Qwen2_5_VLProcessor.apply_chat_template` batched `padding=False` crash." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Both concern model quality, but one is SigLIP2 inference nonsense and the other is CamemBERT masked-LM prediction regression." + "reason": "SigLIP2 nonsensical outputs and CamemBERT masked-LM prediction regressions are different models and bugs." }, { "left": "issue:46032", "right": "issue:46463", "accept": false, - "reason": "Different code paths: Mamba2Mixer use_cache correctness versus T5 decoder-input validation error messaging." + "reason": "Mamba2 `use_cache` correctness issue vs T5 decoder input validation error message are not the same bug." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "SigLIP2 documentation/example issues and HF-vs-JAX implementation discrepancy are related to the same model family but not the same concrete bug." + "reason": "Both mention SigLIP2, but one is a documentation example error and the other is an HF-vs-JAX implementation discrepancy." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Runtime special-token/post-processor updates are unrelated to BigBirdTokenizer's mask-token registration bug." + "reason": "Different tokenizer issues: post-processor not updating at runtime vs BigBirdTokenizer mask token not registered as special." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL video position IDs are separate multimodal issues." + "reason": "Qwen3VL `weight_only=True` loading error and Qwen2.5-VL video `vision_position_ids` bug are separate multimodal issues." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation-map processing and a GenerationMixin import error are unrelated." + "reason": "Completely different areas: void segmentation maps in image processing vs an ImportError for `GenerationMixin`." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both touch fix_mistral_regex, but one is a backend_tokenizer crash while the other is a Kimi-K2.5 codec-handling regression; not clearly the same bug." + "reason": "Mistral regex patch crash is unrelated to the Kimi-K2.5 tokenizer codec/regression report." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Tokenizer post-processor updates and LayoutLMv2 NER/padding crashes are different failures." + "reason": "Runtime special-token post-processor updates are unrelated to LayoutLMv2Tokenizer crashes on NER/batched padding." }, { "left": "issue:45507", "right": "issue:46858", "accept": false, - "reason": "GraniteMoEHybrid invalid method calls and static-cache seq-length tensor arithmetic are unrelated." + "reason": "GraniteMoEHybrid invalid method call and static-cache sequence-length tensor arithmetic are different code paths." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Kimi tokenizer codec/regression and list_repo_files revision forwarding are unrelated." + "reason": "Kimi tokenizer regression is unrelated to `list_repo_files` failing to forward `revision`." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 doc/example errors are not the same as quantization/offloading failures." + "reason": "SigLIP2 example/model mismatch is unrelated to quantization offloading behavior." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading and DAC.from_latents/STE mismatch are different model-specific bugs." - }, - { - "left": "issue:46829", - "right": "issue:46832", - "accept": false, - "reason": "Different model outputs and failure points: output dataclass mismatch versus RT-DETR feature-level indexing TypeError." + "reason": "BitNet packed-weights loading and DAC `from_latents`/STE mismatch are different model bugs." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading are unrelated." + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight unpacking are unrelated." }, { - "left": "issue:45987", - "right": "issue:46829", + "left": "issue:46829", + "right": "issue:46832", "accept": false, - "reason": "StaticCache seq-length return type and VideoPrism output fields are unrelated bugs." + "reason": "Different models and failures: VideoPrism hidden-state output bug vs RT-DETR feature-level mismatch." }, { - "left": "issue:43295", - "right": "issue:45834", + "left": "issue:45987", + "right": "issue:46829", "accept": false, - "reason": "Processor.tokenizer regression and Kosmos2.5 long-OCR index error are different multimodal failures." + "reason": "StaticCache sequence-length return type issue is unrelated to VideoPrism output wiring." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The two accepted duplicate pockets are supported by the titles and descriptions (42886/44843 and 44514/44545), and the rest of the pair verdicts avoid overclaiming by distinguishing different failure modes." + "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate pair is issue:44514/issue:44545, and the other soft-edge pairs are appropriately rejected as distinct bugs." }, "evaluator_used": true, "retried": false, @@ -5732,6 +5718,7 @@ "issue:36331", "issue:39692", "issue:41720", + "issue:43295", "issue:43335", "issue:43531", "issue:43540", @@ -5775,6 +5762,7 @@ "issue:46932" ], "soft_pairs": [ + "issue:43295|issue:45834", "issue:44062|issue:44488", "issue:43550|issue:43927", "issue:43550|issue:44361", @@ -5786,8 +5774,8 @@ "issue:44186|issue:44898", "issue:44008|issue:46858", "issue:43653|issue:43819", - "issue:45987|issue:46932", "issue:41720|issue:44991", + "issue:45987|issue:46932", "issue:30064|issue:36331", "issue:43643|issue:44877", "issue:44220|issue:44661", @@ -5804,202 +5792,208 @@ "issue:45183|issue:45479", "issue:45584|issue:45834" ], - "prepared_review_unit_hash": "f050bbfd9472518bcd82173a9b64c70d3687b9433aaf2696ac4415e04522f75a", + "prepared_review_unit_hash": "98435aeed41754a8c2b11f084f341b5e73986bafb7c498e474ced00e13944b6c", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35750, - "estimated_input_tokens": 8938, - "estimated_eval_tokens": 18132 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36647, + "estimated_input_tokens": 9162, + "estimated_eval_tokens": 18580 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated one-off bug reports. The only near-match is the Mistral tokenizer `_patch_mistral_regex` pair, but it still looks like two different failure modes rather than the same underlying bug.", - "confidence": 0.42, - "canonical_issue_reason": "Issue 44843 is the best anchor because it captures the shared `_patch_mistral_regex` Mistral-tokenizer path in the broadest way and describes the most central failure mode in that small subtopic.", + "summary": "This cluster is mostly false-positive similarity across unrelated Transformers bugs. None of the soft pairs look like the same underlying defect, so all soft edges are rejected. Issue 43550 is the most central item, but it is not a safe duplicate umbrella.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43550 is the most central node among the candidates (highest soft-pair connectivity) and is a concrete runtime/model failure report, so it is the best available issue representative for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "44843 is the strongest overall issue candidate because it is the most general and actionable report in the only closely related pair; the rest of the set is clearly heterogeneous.", + "best_issue_reason": "It is the strongest representative only in a graph-centrality sense; the cluster itself is too heterogeneous for a true duplicate canonical, but 43550 is the clearest, most specific issue to anchor on.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Different code paths: processor/tokenizer regression vs long OCR input indexing error." + }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "Different tokenizer bugs: AddedToken kwargs duplication vs a model-specific load failure." + "reason": "Both are tokenizer-related, but one is an AddedToken constructor bug and the other is a tokenizer loading failure for a specific model." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Unrelated models and paths: torch.compile/SDPA vs save/load config token ID loss." + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig losing custom token IDs on save/load." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Different subsystems and errors: model compile failure vs trainer method signature mismatch." + "reason": "Different models and failures: Bamba compile error vs MLuke tokenizer AttributeError." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both involve Qwen video input, but one is a preprocessing ValueError and the other is chat-template token/feature mismatch on a different model." + "reason": "Both involve video, but one is Qwen3OmniMoe input processing and the other is a chat-template feature/token count mismatch in Qwen3.5." }, { "left": "issue:45381", "right": "issue:46693", "accept": false, - "reason": "Video position IDs vs long-context FA2 performance regression are different bugs in different code paths." + "reason": "Video position ID bug in Qwen2.5-VL is not the same as the FA2 long-context performance regression." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to trust_remote_code config field loss." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to trust_remote_code AutoConfig field loss." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "ZeRO3/BertModel indexing and audio fbank extraction are unrelated failures." + "reason": "ZeRO3/BertModel indexing failure and audio fbank feature extraction are unrelated subsystems." }, { "left": "issue:43611", "right": "issue:46665", "accept": false, - "reason": "Both concern from_pretrained/loading behavior, but they describe different breakages and not the same concrete bug." + "reason": "Model loading with base_model_prefix is a different bug from inconsistent from_pretrained behavior inside Python classes." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer padding/NER crash is unrelated to Perceiver non-default-resolution image classification." + "reason": "Tokenizer crashes on NER padding/truncation are unrelated to Perceiver image classification at non-default resolution." }, { "left": "issue:44008", "right": "issue:46858", "accept": false, - "reason": "Gemma3n variable-name collision and StaticCache seq-length contract are different issues." + "reason": "Gemma3n variable collision in forward() is unrelated to StaticCache sequence-length tensor semantics." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration bug is unrelated to DAC latent decoding/STE mismatch." + "reason": "BigBirdTokenizer special-token registration and DAC.from_latents missing STE are different bugs." }, { - "left": "issue:45987", - "right": "issue:46932", + "left": "issue:41720", + "right": "issue:44991", "accept": false, - "reason": "Cache return-type bug and Gemma4 quantized embedder dtype casting are unrelated." + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { - "left": "issue:41720", - "right": "issue:44991", + "left": "issue:45987", + "right": "issue:46932", "accept": false, - "reason": "Both are loading-related, but one is a CUDA device-mapping assert and the other is tokenizer loading for a different model." + "reason": "StaticCache get_seq_length() returning a tensor is unrelated to Gemma4 quantized embedder dtype casting." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor segmentation-map handling and CustomTrainer loss signature are unrelated." + "reason": "Void segmentation-map processing and CustomTrainer.compute_loss signature mismatch are unrelated." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both mention config loading, but one is missing fields with trust_remote_code and the other is strict config blocking granite_speech; not the same bug." + "reason": "Both involve config loading, but missing fields from trust_remote_code and strict granite_speech config rejection are distinct failures." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction and TOKENIZER_MAPPING_NAMES model-addition logic are unrelated." + "reason": "Audio feature extraction and add-new-model-like failing in TOKENIZER_MAPPING_NAMES are unrelated." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Different model regressions with different symptoms; same broad theme of v5 output changes is too coarse to merge." + "reason": "CLIP attentions assignment regression and RTDetrV2 output drift are different model behaviors." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Sparse-layer config creation and Smollm3 RoPE-layer dropping are unrelated model-configuration bugs." + "reason": "Sparse-layer creation at zero layers and Smollm3 RoPE-layer dropping are separate configuration bugs." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "torch.compile/SDPA failure and DAC.from_latents STE mismatch are unrelated." + "reason": "Bamba compile/SDPA failure and DAC forward/from_latents mismatch are unrelated model issues." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Same `_patch_mistral_regex` area, but one is an offline network call bug and the other is a backend_tokenizer AttributeError; they are separate failure modes." + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; not the same defect." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Documentation example errors/quantization failure and long OCR index error are unrelated." + "reason": "SigLIP2 doc-example errors are unrelated to Kosmos2.5 long-OCR indexing." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both report v4/v5 output differences, but on different models and likely different code paths." + "reason": "Different models and symptoms: Pegasus output drift vs CamemBERT masked-LM prediction errors." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template compilation are unrelated." + "reason": "AutoImageProcessor URL loading and template-node compilation errors have no common code path." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Tokenizer NER/padding crash and AutoTokenizer loading FastSpeech2ConformerTokenizer are different problems." + "reason": "LayoutLMv2 tokenizer crashes and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are unrelated." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Different multimodal models and different attribute/masking bugs." + "reason": "Qwen3OmniMoe video input handling and Gemma3n audio-mask attribute collision are different model paths." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice preset and GenerationMixin import failure are unrelated." + "reason": "BarkProcessor voice_preset and GenerationMixin import errors are unrelated issues." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Input validation messaging and zero-loss classification behavior are separate issues." + "reason": "Validation XOR error messaging and degenerate zero-loss classification behavior are different bugs." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription handling and Kosmos2.5 long-OCR indexing are unrelated." + "reason": "Whisper empty-transcription generation and Kosmos2.5 long OCR index errors are unrelated." } ] }, @@ -6077,20 +6071,20 @@ "issue:43606|issue:45072", "issue:43388|issue:46032", "issue:43825|issue:43994", - "issue:43550|issue:45072", "issue:44479|issue:45834", + "issue:43550|issue:45072", "issue:43782|issue:44977", "issue:43421|issue:44811", - "issue:43577|issue:43818", "issue:44561|issue:45020", - "issue:46294|issue:46463", + "issue:43577|issue:43818", "issue:43582|issue:46512", + "issue:46294|issue:46463", "issue:44991|issue:46291", "issue:43716|issue:44008", "issue:43931|issue:46846", "issue:44805|issue:44936", - "issue:44610|issue:45072", "issue:43540|issue:45381", + "issue:44610|issue:45072", "issue:43749|issue:44291", "issue:43720|issue:45072", "issue:43825|issue:43827", @@ -6100,202 +6094,199 @@ "issue:43577|issue:43582", "issue:34634|issue:36331" ], - "prepared_review_unit_hash": "a2e26656d7122bddc23a037287093a4d51ca7c775efdc0c360928c2a00c9c10d", + "prepared_review_unit_hash": "e187f4702b6bdd1c4976ce6689395492397001692e4b2e6dd494914e03eb9b5e", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35455, - "estimated_input_tokens": 8864, - "estimated_eval_tokens": 17984 + "serialized_chars": 35448, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No pair looks like a true duplicate; the items are mostly unrelated model-specific or subsystem-specific bugs with only loose thematic overlap.", - "confidence": 0.89, - "canonical_issue_reason": "No single duplicate set emerges; if forced, issue 45020 is the broadest umbrella report.", + "summary": "The cluster is mostly a set of unrelated issue reports. All soft edges should be rejected; the tokenizer pair is related in theme but not supported strongly enough to treat as a duplicate from titles alone.", + "confidence": 0.66, + "canonical_issue_reason": "Issue #28218 is the most concrete and narrowly scoped report in the set, so it is the best issue anchor even though it should not be merged as a duplicate with #29159 on the available evidence.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the most representative because it describes a broad version-level regression rather than a niche model-specific failure.", + "best_issue_reason": "Issue #28218 is the clearest, most specific bug report here. It is more precise than the broader tokenizer inconsistency report in #29159, so it is the best representative issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer initialization junk vs FSDP CPU RAM-efficient loading; different failure modes." + "reason": "Similar loading/runtime wording, but they describe different failures and different code paths: junk non-persistent buffers vs FSDP CPU RAM-efficient loading." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next load segfault vs MPS int histogram error; unrelated code paths." + "reason": "Different symptoms and subsystems: Llava Next loading segfault vs an MPS histogram op type error in MoE code." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Bark CPU offload device mismatch vs dtype mismatch in other models; same broad area, not same bug." + "reason": "Both mention device/dtype mismatches, but they affect different models and different failure modes; not the same concrete bug." }, { "left": "issue:43388", "right": "issue:46032", "accept": false, - "reason": "Metric gathering label truncation vs Mamba cache correctness; distinct issues." + "reason": "One is a metrics-label truncation bug, the other is a use_cache correctness issue in Mamba2Mixer; no shared underlying defect." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline translation support message vs SigLIP2 bad outputs; no shared concrete bug." + "reason": "Pipeline error-message wording and a SigLIP2 bad-output bug are distinct; same API surface is not enough to merge them." }, { - "left": "issue:43550", - "right": "issue:45072", + "left": "issue:44479", + "right": "issue:45834", "accept": false, - "reason": "Bamba torch.compile/SDPA failure vs generic bfloat16 dtype mismatch; different root causes." + "reason": "Both involve multimodal/video/OCR inputs, but the models and failures are different; no evidence of the same root cause." }, { - "left": "issue:44479", - "right": "issue:45834", + "left": "issue:43550", + "right": "issue:45072", "accept": false, - "reason": "Qwen video-input regression vs Kosmos2.5 OCR index error; unrelated models and symptoms." + "reason": "torch.compile+SDPA failure in Bamba is unrelated to bfloat16 dtype mismatches in other models." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Qwen3VL weight_only load error vs Qwen3.5 flash-attention generation issue; separate failures." + "reason": "A from_pretrained weight_only loading error is not the same as a flash-attention generation bug." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Tokenizer backend post-processor refresh vs Whisper decode skipping specials; different tokenizer behaviors." - }, - { - "left": "issue:43577", - "right": "issue:43818", - "accept": false, - "reason": "BLIP2 dtype propagation bug vs Video-LLaVA architecture/weight issue; not the same bug." + "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens handling are separate tokenizer/decoding issues." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Specific is_torch_fx_available removal breakage vs broad remote_code regressions; related theme but not one concrete fix." + "reason": "Both reference remote_code regressions, but one is specifically removal of is_torch_fx_available and the other is a broad recent-version breakage report." }, { - "left": "issue:46294", - "right": "issue:46463", + "left": "issue:43577", + "right": "issue:43818", "accept": false, - "reason": "Qwen3.5 video template mismatch vs T5 decoder-input validation message; unrelated." + "reason": "Blip2 dtype persistence and Video-LLaVA architecture/weight-sharing issues are unrelated." }, { "left": "issue:43582", "right": "issue:46512", "accept": false, - "reason": "Apple Silicon warmup TypeError vs stale config.dtype after from_config; different API surfaces." + "reason": "Apple Silicon TypeError in caching_allocator_warmup and stale model.config.dtype after from_config are different bugs." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen3.5 video-chat template token/feature mismatch is unrelated to T5 decoder input validation messaging." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "Tokenizer loading failure for one model vs pipelines import failing on kernels dependency; unrelated." + "reason": "Tokenizer loading failure for a specific model and kernels package import-version requirements are unrelated." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Mistral-3 image preprocessor dtype mismatch vs Gemma3n variable name collision; distinct bugs." + "reason": "A dtype mismatch between image preprocessor and model is not the same as a Gemma3n variable-name collision AttributeError." }, { "left": "issue:43931", "right": "issue:46846", "accept": false, - "reason": "Qwen3-VL weight shape mismatch vs Qwen3.5 TP-plan missing entries; different model-loading problems." + "reason": "Weight shape mismatch during load and missing TP-plan entries causing OOM/shape errors are different issues, even if both involve Qwen models." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask shape IndexError vs evaluate-after-train failure; not the same underlying issue." + "reason": "Mask/tensor shape mismatch and trainer.evaluate failing after train do not share a concrete underlying bug." }, { - "left": "issue:44610", - "right": "issue:45072", + "left": "issue:43540", + "right": "issue:45381", "accept": false, - "reason": "OmDet processor/model image-size mismatch vs dtype mismatch in other models; unrelated." + "reason": "Both involve video inputs in Qwen-family models, but they target different model variants and different mistakes (processing error vs position-id regression)." }, { - "left": "issue:43540", - "right": "issue:45381", + "left": "issue:44610", + "right": "issue:45072", "accept": false, - "reason": "Video processing ValueError in Qwen3OmniMoe vs Qwen2.5-VL vision_position_ids regression; same domain, different bug." + "reason": "OmDet-Turbo processor image-size mismatch is unrelated to bfloat16 dtype mismatch failures." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP RAM-efficient loading breakage vs init_empty_weights unexpected argument; both loading-related but distinct root causes." + "reason": "FSDP CPU RAM-efficient loading being broken is not the same as init_empty_weights rejecting an unexpected argument." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading vs dtype mismatch; different code paths." + "reason": "Packed-weight unpacking during accelerate loading and dtype mismatches are distinct loading issues." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Error-message regression vs docs still referencing pipeline after removals; related topic but not same change." + "reason": "One is a runtime error-message bug, the other is stale docs referencing removed pipeline() APIs; related topic, but not the same issue." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "Extra space after added token vs general slow/fast tokenizer inconsistency; one symptom does not establish same bug." + "reason": "The title overlap is not enough: #28218 is a specific added-token spacing bug, while #29159 is a broader slow-vs-fast tokenizer inconsistency report without stronger body-level evidence of the same root cause." }, { "left": "issue:43645", "right": "issue:46665", "accept": false, - "reason": "Notebook custom-model initialization breakage vs from_pretrained inside Python classes; different context-specific behaviors." + "reason": "Both involve model initialization patterns, but the notebook custom-model breakage and from_pretrained-inside-classes inconsistency are not clearly the same bug." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT MLM prediction regression vs tokenizer version changing tokenization; too broad and not the same defect." + "reason": "CamemBERT masked-LM prediction regressions and a general tokenization-change report are too broad and not concrete duplicates." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "BLIP2 dtype state bug vs Apple Silicon warmup TypeError; unrelated." - }, - { - "left": "issue:34634", - "right": "issue:36331", - "accept": false, - "reason": "BarkProcessor voice_preset failure vs CustomTrainer compute_loss signature error; unrelated." + "reason": "Blip2 dtype persistence and Apple Silicon warmup TypeError are unrelated failures." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The analysis is mostly grounded, but it overstates the evidence for the only accepted soft edge. In the packet, `issue:28218` vs `issue:29159` is explicitly marked `deterministic_accept: false`, and the titles alone do not firmly establish a duplicate relationship\u2014`#28218` is a specific added-token spacing bug, while `#29159` is a broader slow-vs-fast tokenizer inconsistency report. Please keep that verdict conservative and reject it unless there is stronger body-level evidence. The rest of the soft-edge rejections look appropriately cautious." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -6366,8 +6357,8 @@ "issue:46175|issue:46932", "issue:46697|issue:46728", "issue:45003|issue:46285", - "issue:43653|issue:44912", "issue:43782|issue:45325", + "issue:43653|issue:44912", "issue:43262|issue:43819", "issue:45372|issue:46326", "issue:43582|issue:44857", @@ -6381,8 +6372,8 @@ "issue:44060|issue:44368", "issue:36010|issue:43645", "issue:36010|issue:36331", - "issue:43329|issue:45600", "issue:43906|issue:46209", + "issue:43329|issue:45600", "issue:43526|issue:44554", "issue:45084|issue:45538", "issue:43819|issue:43927", @@ -6391,14 +6382,14 @@ "issue:42175|issue:46291", "issue:42490|issue:43257" ], - "prepared_review_unit_hash": "2b2b799da1455c58938a631a6506ef1c2a1708423546e2508ef908e1f2d21259", + "prepared_review_unit_hash": "3c88361a6022902c4dc2d687e6ed4aec616defc60c62b664197c473b54d752cc", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, @@ -6413,180 +6404,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it spans unrelated loading, tokenizer, multimodal, quantization, and trainer regressions. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them.", - "confidence": 0.94, - "canonical_issue_reason": "Issue #45020 is the broadest umbrella-style report here, covering recent-version model-loading breakage via remote code; it is the closest thing to a representative issue, though the cluster is not truly duplicate-heavy.", + "summary": "No soft pair looks like the same underlying bug/change; the matches are mostly broad semantic similarity across unrelated regressions. I\u2019d keep the issues separate and use a broad, clearly-scoped issue as the canonical representative.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45020 has the clearest broad regression title and scope, making it the best representative issue in this set.", "canonical_pr_reason": null, - "best_issue_reason": "#45020 is the most general and cross-cutting of the listed issues, but only as a loose representative of the loading-regression theme rather than a true duplicate center.", + "best_issue_reason": "issue:45020 is the strongest standalone issue for tracking because it states a concrete, user-facing breakage with a clear subsystem and version regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Different failures: backend extra/dependency omission vs Mistral tokenizer regex crash." + "reason": "Different bugs: backend install/dependency selection vs Mistral tokenizer regex patch crash." }, { "left": "issue:46175", "right": "issue:46932", "accept": false, - "reason": "Gemma4-related but distinct bugs: huge one-hot embedding materialization vs quantized embedders dtype casting." + "reason": "Both Gemma4-related, but one is a huge position-embedding materialization issue and the other is dtype casting for quantized embedders." }, { "left": "issue:46697", "right": "issue:46728", "accept": false, - "reason": "Different image-processor bugs: over-padding vs incorrect patch count on odd resolution multiples." + "reason": "Different image processors and symptoms: over-padding vs wrong patch count." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unrelated code paths: unsafe sys.modules access vs CLIP text model structure inconsistency." + "reason": "Remote-code/sys.modules safety issue vs CLIP model structure inconsistency." }, { - "left": "issue:43653", - "right": "issue:44912", + "left": "issue:43782", + "right": "issue:45325", "accept": false, - "reason": "Special-token registration bug vs quantization fallback issue; not the same underlying defect." + "reason": "Different model families and failures: Qwen3-VL load error vs Qwen2.5-VL rope indexing bug." }, { - "left": "issue:43782", - "right": "issue:45325", + "left": "issue:43653", + "right": "issue:44912", "accept": false, - "reason": "Both VL models, but one is a weight_only load error and the other is a rope-index scaling bug." + "reason": "Tokenizer special-token registration bug vs quantization fallback/load problem." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate default vs DAC latent reconstruction mismatch are different bugs." + "reason": "Audio chat-template sample-rate default vs DAC forward/from_latents mismatch." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "Different tokenizer/processor failures in different integration points; not one fix." + "reason": "Both involve mistral_common, but one is an import dependency break and the other is a continue_final_message behavior bug." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Different platform/runtime failures: Apple Silicon TypeError vs CUDA AMP loss crash." + "reason": "Apple Silicon TypeError vs CUDA AMP crash; unrelated execution paths." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Both mention loading/processor code, but Mistral regex patch failure and Gemma4Processor missing _tokenizer are separate issues." + "reason": "Different missing-attribute failures in different code paths: backend_tokenizer vs _tokenizer." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "Both are loading-related, but a key error on one model is not the same concrete bug as remote_code regressions in general." + "reason": "Specific KeyError loading one model vs broad remote_code regression; not the same concrete defect." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage deserialization error vs torchvision dependency requirement are unrelated." + "reason": "Storage-object lookup failure vs PIL/torchvision dependency requirement." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "Torch.compile SDPA failure and MXFP4 quantization fallback affect different code paths." + "reason": "torch.compile/SDPA bug vs MXFP4 quantization load fallback." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both are from model loading, but gguf_file handling and a specific tokenizer regression are not the same bug." + "reason": "Different loading regressions: gguf_file handling across auto classes vs a specific tokenizer load failure." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE weight conversion problem vs undefined variables in multimodal token counting are unrelated." + "reason": "Accelerate/deepspeed weight conversion bug vs multimodal token counting/video branch undefined names." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Label reduction bug and RTDetr output divergence do not share the same root cause." + "reason": "Beit label reduction bug vs RTDetrV2 output drift in v5; different model behaviors." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Related model family, but one is an incorrect tied-weights warning and the other is a user-facing config warning." + "reason": "Related warning theme, but one is a model weight-tying bug and the other is a user-facing fine-tuning warning." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "ImportError for GenerationMixin and Jupyter custom-model initialization breakage are different failures." + "reason": "Import path regression vs notebook custom-model initialization breakage." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Import path regression vs custom trainer signature mismatch; no shared underlying fix." + "reason": "ImportError from generation module vs custom trainer signature mismatch." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:43906", + "right": "issue:46209", "accept": false, - "reason": "Multimodal token counting bug vs stale auto_mappings config references are unrelated." + "reason": "Repro of an older issue vs FP8 inference failure on B200; unrelated symptoms." }, { - "left": "issue:43906", - "right": "issue:46209", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "Isolated reproduction of another issue vs FP8 inference failure on B200; not the same bug." + "reason": "Undefined video-branch helper bug vs stale auto_mappings config reference." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Image label reduction bug and MPS attention correctness issue are separate." + "reason": "Image label reduction bug vs MPS attention correctness issue." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation error vs CLIPTokenizer max_length bug are unrelated." + "reason": "Template compilation error vs tokenizer model_max_length sentinel issue." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC forward/from_latents mismatch and DiaConfig token-ID loss are different problems." + "reason": "DAC latent reconstruction mismatch vs DiaConfig save/load token-ID loss." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Zero-3 index error in a non-pretrained Bert setup is not the same as evaluate-after-train failure." + "reason": "Zero3/non-pretrained Bert IndexError vs train-then-evaluate failure." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Sliding-window bug in Qwen3-MoE and torchvision dependency breakage have no shared code path." + "reason": "Qwen3-MoE sliding-window bug vs PIL backend torchvision requirement." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "TensorFlow backend packaging issue and kernels version/revision import failure are unrelated." + "reason": "Torch backend install issue vs kernels dependency version/revision requirement." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local-model warning vs Qwen3 MoE weight conversion are not the same defect." + "reason": "Trailing-slash local model import warning vs Qwen3-MoE weight conversion failure." } ] }, @@ -6661,9 +6652,9 @@ "soft_pairs": [ "issue:43746|issue:44811", "issue:43643|issue:45759", - "issue:43720|issue:44610", "issue:43854|issue:44977", "issue:39692|issue:44912", + "issue:43720|issue:44610", "issue:38175|issue:44610", "issue:43257|issue:43696", "issue:43262|issue:44811", @@ -6687,202 +6678,205 @@ "issue:43454|issue:43577", "issue:44568|issue:45356" ], - "prepared_review_unit_hash": "4461c4b8a993a82e943d260ceb1c801c3341409f1387893b2cf9ebb3b589148a", + "prepared_review_unit_hash": "6566ff3b06efc6684b946a3930e89afa07ffbb9ddf5f1b9a14ba2d5f80e543c5", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35444, - "estimated_input_tokens": 8861, - "estimated_eval_tokens": 17978 + "serialized_chars": 35447, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the issues span unrelated models, tokenizers, loading paths, and training/runtime bugs. None of the soft pairs look like true duplicates.", - "confidence": 0.88, - "canonical_issue_reason": "No safe canonical issue stands out; the cluster is too diverse to anchor on one report.", + "summary": "The cluster is still mostly heterogeneous. `issue:43329` is the strongest representative for the video/multimodal area, but the related video items are adjacent more than clearly duplicate; I would not merge any of the soft-edge pairs here.", + "confidence": 0.66, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a strong global representative because the cluster mixes many distinct bug classes and subsystems.", + "best_issue_reason": "`issue:43329` is the clearest representative because it is a concrete bug in the multimodal token-counting/video code path, but the surrounding video/VLM reports are only loosely adjacent rather than definite duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT local-checkpoint loading vs Whisper batch_decode special-token handling; different code paths." + "reason": "Different bugs: PEFT/local checkpoint loading vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "AutoConfig trust_remote_code field loss vs AutoModelForCausalLM.from_config text_config unwrapping; related area, different bug." - }, - { - "left": "issue:43720", - "right": "issue:44610", - "accept": false, - "reason": "BitNet packed-weight loading bug vs OmDet image-size mismatch; unrelated failures." + "reason": "Both involve config-loading behavior, but they target different code paths and different failures." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "GLM-4.7-Flash load issue vs Qwen3.5 flash-attention generation bug; different models and symptoms." + "reason": "Different models and failures; one is a unit-test load issue, the other is a flash-attention generation regression." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "SigLIP2 docs/example errors vs gpt-oss-20b MXFP4 quantization fallback; not the same bug." + "reason": "SigLIP2 docs/example errors are unrelated to GPT-OSS MXFP4 quantization fallback." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to OmDet-Turbo processor image-size mismatch." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero probabilities vs OmDet processor/model input-size mismatch; distinct defects." + "reason": "Different model-specific problems: zero probabilities vs wrong processor output resolution." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MoE weight conversion/loading bug vs GPT-oss OOM; different underlying problems." + "reason": "Qwen3 MoE weight conversion/loading is unrelated to a GPT-OSS out-of-memory failure." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio apply_chat_template sampling-rate default vs Whisper decode token skipping; unrelated processors." + "reason": "Both are processor-related, but one is chat-template sample-rate handling and the other is Whisper decoding flags." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE accelerate/deepspeed conversion bug vs Qwen3-VL weight-shape mismatch; different load failures." + "reason": "Different loading bugs: accelerate/deepspeed weight conversion vs Qwen3-VL shape mismatch." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Both are OOM-related, but one is int4 reserved-memory behavior and the other is a model-specific video OOM." + "reason": "One is an int4 memory regression and the other is a model-specific CUDA OOM; roots are not the same." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Undefined multimodal token helper in video branch vs batched video processor shape bug; separate code paths." + "reason": "Shared video theme, but token-counting undefined variables and batched-shape handling are different bugs." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Tokenizer error masking with missing protobuf vs bare except in Fuyu tensor conversion; different components." + "reason": "Both swallow errors, but they are different exception-handling bugs in different components." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "Video-token helper bug vs broad v5.3.0 video-input regression; overlapping area but not the same concrete issue." + "reason": "They are thematically close video/VLM regressions, but the packet does not clearly show the same underlying bug or a single code-path fix." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Qwen3Moe FP8 config failure vs MobileLLM-125M key error; unrelated loading bugs." + "reason": "Different model-loading failures: FineGrainedFP8Config vs a key error in MobileLLM." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token-count tracking vs GPT-oss perplexity regression; different subsystems." + "reason": "Trainer token-count tracking and GPT-OSS perplexity are unrelated problems." }, { "left": "issue:46421", "right": "issue:46424", "accept": false, - "reason": "Chunked-prefill recompilation vs max_cache_length handling; both cache-related but distinct behaviors." + "reason": "Both are cache-related, but one is lazy-init recompilation and the other is max_cache_length semantics." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM efficient loading vs mmap/Strix Halo OOM; separate memory-related issues." + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are different loading/memory issues." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights argument mismatch vs PIL image processors requiring torchvision; unrelated." + "reason": "init_empty_weights argument handling is unrelated to PIL backend torchvision requirements." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "Generic placeholder issue vs isolated reproduction of another issue; no shared bug." + "reason": "The first title is not a concrete bug report; the second is an isolated reproduction of another issue." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Bark CPU-offload device mismatch vs Qwen3_5Moe rotary embedding CPU-offload incompatibility; different models/code paths." + "reason": "Both mention CPU offload, but they affect different models and different code paths." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Qwen3Moe FP8 loading bug vs AutoImageProcessor-from-URL failure; unrelated." + "reason": "Qwen3Moe FP8 loading and AutoImageProcessor-from-URL are unrelated." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Slow/fast tokenizer inconsistency vs trainer token-count update bug; different areas." + "reason": "Tokenizer slow/fast inconsistencies are unrelated to TrainerState token counting." }, { "left": "issue:43540", "right": "issue:46701", "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError vs DiffusionGemma pipeline failure; different models and pipelines." + "reason": "Different multimodal pipelines and different failure modes." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig save/load token-ID loss vs gpt-oss MXFP4 quantization fallback; unrelated." + "reason": "DiaConfig token-ID persistence is unrelated to MXFP4 quantization fallback." }, { "left": "issue:28282", "right": "issue:29870", "accept": false, - "reason": "Missing PyTorch import error vs Bart generate decoder_attention_mask unbound local; different issues." + "reason": "ImportError for PyTorch and Bart generate() UnboundLocalError are unrelated." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "AyaVision lm_head tying bug vs Blip2 dtype propagation bug; not the same failure." + "reason": "Both affect multimodal model loading/generation, but the underlying bugs are different." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "mDeBERTa special-token regression vs Kimi-K2.5 tokenizer codec regression; both tokenizer bugs but distinct." + "reason": "Different tokenizer regressions on different models; same subsystem, not the same bug." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level summary is mostly conservative, but one soft-edge verdict is too strong: issue:43329 and issue:44479 are similar video/VLM regressions, yet the packet does not clearly show they are the same bug. Also, saying 43329 is the only item with a plausible duplicate partner slightly overstates the evidence, since other video/multimodal items are at least thematically adjacent even if not duplicates." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -6957,11 +6951,11 @@ "issue:33357|issue:34567", "issue:43404|issue:45127", "issue:43425|issue:46291", - "issue:45538|issue:46869", "issue:45137|issue:45405", - "issue:42915|issue:45538", + "issue:45538|issue:46869", "issue:42175|issue:43824", "issue:28282|issue:29127", + "issue:42915|issue:45538", "issue:44479|issue:46693", "issue:44062|issue:44991", "issue:43742|issue:44291", @@ -6972,228 +6966,228 @@ "issue:24643|issue:29159", "issue:44849|issue:45701", "issue:27957|issue:28282", - "issue:44987|issue:45907", "issue:41720|issue:43540", + "issue:44987|issue:45907", "issue:43540|issue:44560", "issue:44479|issue:45381", "issue:42175|issue:44488", "issue:34567|issue:36331", "issue:43540|issue:44479" ], - "prepared_review_unit_hash": "d2624d9e0b6f61a90ca2d9bddf7fefd25aac0f14facc447160c4ab6393adfd23", + "prepared_review_unit_hash": "77cbe9260bc4b344d2e4ad1a6ac21f279fc7854afdb74b1a540745f95a490254", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 35980, - "estimated_input_tokens": 8995, - "estimated_eval_tokens": 18246 + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the issues span unrelated model-loading, tokenizer, training, and backend regressions. Only one soft edge looks plausibly like the same underlying bug; the rest should stay separate.", - "confidence": 0.86, - "canonical_issue_reason": "If forced to anchor a small coherent slice, issue 44479 is the broadest Qwen video-input regression report and best represents that subcluster; however, the overall cluster is still too mixed to have a truly clean canonical issue.", + "summary": "This cluster is mostly heterogeneous, with many unrelated model-loading, trainer, tokenizer, and video-input issues. The only clearly plausible duplicate subcluster is the Qwen video regression pair 44479/45381; the rest should stay separate.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 44479 is the best canonical anchor because it is the broadest regression ticket in the only coherent subcluster and aligns with the more specific video-position-id symptom reported in 45381.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the strongest representative of the only obvious subtheme (Qwen video-input regressions). It is broad enough to subsume nearby reports like 45381, but it should not be treated as a canonical for the whole cluster.", + "best_issue_reason": "Issue 44479 is the best overall representative for triage: it names the regression, spans multiple affected Qwen video models, and is the strongest match for a mergeable duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Different problems: Llama4 pad_token_id config attribute vs kernels list_repo_files revision forwarding." + "reason": "Both are loading-related, but they describe different failures in different subsystems: Llama4 pad_token_id vs kernels revision forwarding." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "One is a model-loading regression on transformers>=5.1.0; the other is CLIPTokenizer's max_length constant." + "reason": "One is a model-loading failure for a specific repo; the other is a tokenizer max_length bug. No shared code-path." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Missing PyTorch in the environment is unrelated to the Llama 3.2 vision loading regression." + "reason": "Both mention model loading, but one is a missing PyTorch import error and the other is a Llama 3.2 vision-loading regression." }, { "left": "issue:44857", "right": "issue:46832", "accept": false, - "reason": "Different RT-DETR/LwDetr failures: AMP/CUDA loss crash vs forward shape/feature-level TypeError." + "reason": "Different RT-DETR bugs: AMP loss crash versus forward shape/feature-level mismatch." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output drift in v5 and Qwen3.5 save_pretrained key mapping are separate code paths." + "reason": "Different issues in different paths: output differences in RTDetrV2 vs incorrect visual encoder keys on save_pretrained." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token-count tracking." + "reason": "Completely unrelated: MacOS bus error with CLIP vs TrainerState token counting." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Untied lm_head weights in Mistral3 and LoRA merge collapse with extended vocab are distinct defects." + "reason": "Both involve weight tying, but they are different model families and different failure modes; not the same concrete bug." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility is a different compatibility issue than kernels revision forwarding." - }, - { - "left": "issue:45538", - "right": "issue:46869", - "accept": false, - "reason": "CLIPTokenizer max_length behavior and YOLO multi-GPU device-map failures do not share a code path." + "reason": "Torch compatibility and kernels revision handling are unrelated issues." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque underflow is unrelated to the unreleased PEFT version bump." + "reason": "DeepSpeed ZeRO3 deque error and PEFT version bump are unrelated." }, { - "left": "issue:42915", - "right": "issue:45538", + "left": "issue:45538", + "right": "issue:46869", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to CLIPTokenizer's max_length constant." + "reason": "Tokenizer max_length behavior and YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging missing TensorFlow backend and missing Qwen2_5_VL import are separate import/distribution problems." + "reason": "Backend packaging/import problem vs missing Qwen2.5-VL symbol; different causes." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "PyTorch-not-found import error is unrelated to LayoutLMv3's unclear box-content validation message." + "reason": "Generic missing-PyTorch import error vs LayoutLMv3 box-validation messaging; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FP8/Qwen3Moe failure and CLIPTokenizer max_length are unrelated." }, { "left": "issue:44479", "right": "issue:46693", "accept": false, - "reason": "A video-input regression and a FA2 long-context performance regression are different bugs, not one code-path fix." + "reason": "Video-input regression and FA2 long-context performance regression are different problems." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken special-arg duplication is not the same as a specific tokenizer loading failure for EMBEDDIA/est-roberta." + "reason": "AddedToken keyword collision vs tokenizer loading failure for a specific model repo; no shared code-path." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "MobileLLM-125M key loading and init_empty_weights' _is_hf_initialized argument mismatch are separate model-init issues." + "reason": "Both are model-loading errors, but one is a key lookup issue and the other is an init_empty_weights argument mismatch." }, { "left": "issue:44936", "right": "issue:46821", "accept": false, - "reason": "trainer.evaluate after train and TimesFM window_size AttributeError are unrelated trainer/model bugs." + "reason": "Trainer lifecycle bug vs TimesFM window_size attribute error; unrelated." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8Config failure and FSDP CPU RAM efficient loading are different regressions." + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are unrelated." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf_file loading breakage and AppleSilicon caching_allocator_warmup TypeError are unrelated." + "reason": "GGUF loading API break and Apple Silicon warmup TypeError are unrelated." }, { "left": "issue:43577", "right": "issue:46512", "accept": false, - "reason": "Both concern dtype handling, but one is Blip2 weight casting and the other is stale model.config.dtype from from_config; not the same bug." + "reason": "Both concern dtype handling, but one is a Blip2 loading issue and the other is a generic from_config config-staleness bug; not the same change." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed 'weight must be 2-D' training error and tokenizer slow/fast inconsistency are unrelated." + "reason": "DeepSpeed weight-dimension error and slow/fast tokenizer inconsistency are unrelated." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states behavior and generic version-dependent tokenization changes are not the same bug." + "reason": "Qwen3.5 hidden-states bug and version-dependent tokenization changes are unrelated." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta is unrelated to an environment missing PyTorch." + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to missing PyTorch import handling." }, { - "left": "issue:44987", - "right": "issue:45907", + "left": "issue:41720", + "right": "issue:43540", "accept": false, - "reason": "Model loading failure on transformers>=5.1.0 and list_repo_files not forwarding revision are separate issues." + "reason": "Both are Qwen-family runtime failures, but they are different code paths and symptoms: auto device mapping vs video input processing." }, { - "left": "issue:41720", - "right": "issue:43540", + "left": "issue:44987", + "right": "issue:45907", "accept": false, - "reason": "Qwen3 auto device-mapping cudaErrorAssert and Qwen3OmniMoe video-input ValueError are different failure modes." + "reason": "Loading a model repo and list_repo_files revision forwarding are different bugs." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Two Qwen video-input bugs, but one is Qwen3OmniMoe ValueError and the other is Qwen3-vl-embedding StopIteration; not clearly the same root cause." + "reason": "Both involve video inputs, but they affect different Qwen model families and do not clearly share the same concrete bug." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Same Qwen video-input regression family in 5.3.0; 45381 looks like the qwen2.5-vl symptom of the broader bug reported in 44479." + "reason": "These look like the same underlying Qwen video-input regression: 44479 reports the broader v5.3.0 regression and 45381 describes the specific vision_position_ids symptom." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "TensorFlow packaging omission is unrelated to a specific model-loading failure for cjvt/sleng-bert." + "reason": "Packaging/backend inclusion issue vs a specific model load failure; unrelated." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token accounting and CustomTrainer.compute_loss signature mismatch are separate training API issues." + "reason": "TrainerState token accounting and custom compute_loss signature mismatch are separate trainer API issues." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both mention video inputs, but one is a Qwen3OmniMoe ValueError and the other is a broader v5.3.0 video regression across several Qwen model families; not enough evidence to merge as the same bug." + "reason": "Both are video-input related, but the model families and reported failures differ enough that they do not clearly represent one mergeable bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about the cluster\u2019s heterogeneity. The only accepted soft edge, issue:44479 \u2194 issue:45381, is plausibly the same Qwen video-input regression family; the other denials are well-justified and do not overstate similarity. Canonical/representative wording is also appropriately cautious." + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft edge, 44479/45381, is plausibly the same Qwen video-input regression based on the titles, and the remaining pair verdicts are appropriately cautious. No clear overstatement stands out." }, "evaluator_used": true, "retried": false, @@ -7244,15 +7238,14 @@ "issue:45072", "issue:45081", "issue:45325", - "issue:45357", "issue:45405", "issue:45488", + "issue:45588", "issue:45600", "issue:45663", "issue:45672", "issue:45674", "issue:45823", - "issue:45920", "issue:46399", "issue:46463", "issue:46468", @@ -7273,14 +7266,14 @@ "issue:45042|issue:45081", "issue:43582|issue:46821", "issue:43756|issue:46399", - "issue:43756|issue:45325", "issue:43742|issue:44589", + "issue:43756|issue:45325", "issue:43716|issue:46701", "issue:44568|issue:46491", "issue:45405|issue:45600", + "issue:44737|issue:46463", "issue:45672|issue:45674", "issue:43577|issue:45072", - "issue:44737|issue:46463", "issue:43531|issue:44821", "issue:45823|issue:46463", "issue:43906|issue:46693", @@ -7288,207 +7281,207 @@ "issue:24643|issue:29870", "issue:43452|issue:44488", "issue:44625|issue:44877", - "issue:45663|issue:46846", "issue:43540|issue:46693", + "issue:45663|issue:46846", "issue:43697|issue:46829", - "issue:45357|issue:45920" + "issue:44737|issue:45588" ], - "prepared_review_unit_hash": "56c6d7a193e42fa6b0b6ed1d0cdb7cb98b9bc75f131537578056ffd0f0e0266e", + "prepared_review_unit_hash": "906e5ecab1fe092900c04bb9fbb65498bac9069dd47f2dda89781d797b663567", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36556, - "estimated_input_tokens": 9139, - "estimated_eval_tokens": 18534 + "serialized_chars": 35945, + "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 a loose collection of unrelated Transformers bugs, mostly model-specific regressions and loading errors. None of the soft pairs look like true duplicates; they share only broad subsystem similarity.", - "confidence": 0.96, - "canonical_issue_reason": "issue:45488 is the strongest representative issue here: it has a clear, concrete tokenizer regression, strong discussion activity, and broad impact across DeepSeek V3/R1-style models.", + "summary": "The cluster is highly heterogeneous: most pairs share only broad subsystem keywords (tokenizers, multimodal loading, dtype, config, training) but not the same concrete bug or code path, so I would not merge any of the soft edges.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the items span many unrelated bugs, and none of the soft pairs is close enough to serve as a single duplicate anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is the best standalone issue in this set because it is well-scoped, reproducible, and describes a user-visible regression with wider ecosystem impact.", + "best_issue_reason": "No clear best issue: there is no strong representative issue for the cluster because the candidate pairs are not true duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Different models and failure modes: unexpected zero probabilities vs image-classification resolution/interpolation failure." + "reason": "Both are vision-model inference bugs, but they involve different models and different failure modes (zero probabilities vs resize/interpolation behavior)." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both are multimodal-related, but one is a Janus generation test failure and the other is a Phi-4-multimodal loading issue; no shared code-path bug." + "reason": "Both are multimodal loading/generation reports, but they target different models and distinct breakages." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Separate tokenizer regressions: one is BOS/EOS not added for mDeBERTa, the other is LlamaTokenizer overriding ByteLevel pre-tokenizer behavior." + "reason": "Both concern tokenizer behavior, but one is missing BOS/EOS insertion and the other is a pre-tokenizer override regression in a different tokenizer path." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Completely different bugs: DeepSpeed/accelerate MoE weight conversion vs Smollm3 RoPE layer dropping." + "reason": "These are unrelated Qwen/SmolLM issues: one is MoE weight conversion under accelerate+deepspeed, the other is RoPE layer selection." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Both are tokenizer loading failures, but the root causes differ: AddedToken kwargs collision vs missing backend_tokenizer attribute in a regex patch." + "reason": "Both are tokenizer errors, but they come from different code paths and unrelated root causes." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Different API surfaces and symptoms: init_empty_weights argument handling vs loading AutoImageProcessor from a URL." + "reason": "These address completely different subsystems: init_empty_weights loading versus AutoImageProcessor URL loading." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Image processor torchvision dependency regression is unrelated to the Mistral tokenizer regex crash." + "reason": "Image processor torchvision dependency and Mistral regex patching are unrelated bugs." }, { "left": "issue:43582", "right": "issue:46821", "accept": false, - "reason": "AppleSilicon allocator TypeError and TimesFM window_size AttributeError are unrelated code paths." + "reason": "AppleSilicon allocator warmup and TimesFM window_size handling are unrelated." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Different Qwen/Smollm bugs: RoPE layer configuration vs missing decorator for apply_rotary_pos_emb registration." + "reason": "Different models and different failures: RoPE layer dropping versus kernel registration/decorator error." }, { - "left": "issue:43756", - "right": "issue:45325", + "left": "issue:43742", + "right": "issue:44589", "accept": false, - "reason": "Both touch RoPE-ish logic, but one drops layers in Smollm3 while the other mis-scales temporal positions in Qwen2.5-VL." + "reason": "Both are model-loading errors, but the specific causes and affected paths are different." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:43756", + "right": "issue:45325", "accept": false, - "reason": "Different load-time failures: MobileLLM key lookup vs Float8 storage lookup TypeError." + "reason": "Both mention RoPE/position handling, but they affect different models and different logic." }, { "left": "issue:43716", "right": "issue:46701", "accept": false, - "reason": "Different multimodal regressions with different models and symptoms; no shared underlying bug." + "reason": "Both involve multimodal pipelines, but one is dtype mismatch and the other is a broken image-text-to-text output path." }, { "left": "issue:44568", "right": "issue:46491", "accept": false, - "reason": "Separate tokenizer regressions affecting different model families and different round-trip/add_special_tokens behavior." + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different round-trip behaviors." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "One is a version pin/release packaging problem; the other is stale auto_mappings config references breaking CI." + "reason": "A PEFT version pin and a removed config reference are unrelated CI/config issues." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "XLNet CPU allocation bug and T5 decoder error-message bug are unrelated." }, { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both involve 4-bit inference, but one is a Gemma4 audio-module uint8/torch.finfo crash and the other is a BitsAndBytesConfig skip_modules exclusion bug." + "reason": "Both are 4-bit inference issues, but one is a Gemma4 audio-module uint8/finfo problem and the other is a BitsAndBytesConfig exclusion bug." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both are dtype mismatches, but they occur in different models and different inference paths." - }, - { - "left": "issue:44737", - "right": "issue:46463", - "accept": false, - "reason": "XLNet CPU device placement bug and T5 decoder-input error-message issue are unrelated." + "reason": "Both mention dtype mismatches, but they are different models and different inference paths." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoImageProcessor URL loading." + "reason": "Qwen3-MoE sliding_window and AutoImageProcessor URL loading are unrelated." }, { "left": "issue:45823", "right": "issue:46463", "accept": false, - "reason": "Gemma4 device-map forward mismatch is unrelated to T5 decoder validation messaging." + "reason": "Gemma4 device_map forward mismatch and T5 validation messaging are unrelated." }, { "left": "issue:43906", "right": "issue:46693", "accept": false, - "reason": "An isolated reproduction of a past issue is not the same as a new FA2 long-context performance regression." + "reason": "The first is an isolated reproduction of another issue; the second is a FA2 performance regression, so they do not describe the same bug." }, { "left": "issue:43906", "right": "issue:46468", "accept": false, - "reason": "No evidence these describe the same Nemotron/model-loading bug; one is just a reproduction of another issue." + "reason": "An isolated reproduction of a separate issue is not enough to match an official Nemotron example breakage." }, { "left": "issue:24643", "right": "issue:29870", "accept": false, - "reason": "DeepSpeed training weight shape error and Bart generate() decoder_attention_mask UnboundLocalError are distinct failures." + "reason": "DeepSpeed weight-dimension training error and Bart decoder_attention_mask UnboundLocalError are unrelated." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both are loading-related, but one is a gguf/from_pretrained breakage and the other is a specific model load failure with no clear shared root cause." + "reason": "gguf_file breaks from_pretrained APIs, while the other reports a model-specific loading failure." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Different config propagation/strictness issues affecting different model families." + "reason": "One is Qwen3.5 label propagation, the other is strict config loading for granite_speech; different fixes and code paths." }, { - "left": "issue:45663", - "right": "issue:46846", + "left": "issue:43540", + "right": "issue:46693", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states bug and Qwen3.5 TP-plan missing linear_attn entries are different distributed-training problems." + "reason": "Video input processing in Qwen3OmniMoe is unrelated to FA2 long-context performance regression." }, { - "left": "issue:43540", - "right": "issue:46693", + "left": "issue:45663", + "right": "issue:46846", "accept": false, - "reason": "Different models and failure classes: video input processing error vs FA2 generation performance regression." + "reason": "Both are parallelism/training issues, but they affect different models and different mechanisms (FSDP2 shared_kv_states vs TP plan entries)." }, { "left": "issue:43697", "right": "issue:46829", "accept": false, - "reason": "Different model outputs/heads regressions with no shared concrete code path." + "reason": "Different models and different symptoms: output drift versus an output field mis-assignment." }, { - "left": "issue:45357", - "right": "issue:45920", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "Qwen3.5 save_pretrained key regression and broad AutoTokenizer wrong-token-ID regression are separate issues." + "reason": "XLNet CPU tensor creation and flash_attention s_aux=None handling are unrelated." } ] }, @@ -7535,7 +7528,6 @@ "issue:44442", "issue:44466", "issue:44493", - "issue:44737", "issue:44811", "issue:44843", "issue:44898", @@ -7545,11 +7537,12 @@ "issue:45200", "issue:45216", "issue:45356", + "issue:45357", "issue:45372", "issue:45375", "issue:45479", "issue:45542", - "issue:45588", + "issue:45920", "issue:46050", "issue:46121", "issue:46170", @@ -7562,7 +7555,7 @@ "issue:46932" ], "soft_pairs": [ - "issue:44737|issue:45588", + "issue:45357|issue:45920", "issue:46463|issue:46899", "issue:28282|issue:32090", "issue:44913|issue:46121", @@ -7577,208 +7570,208 @@ "issue:43761|issue:46829", "issue:43452|issue:44843", "issue:44361|issue:44442", - "issue:43653|issue:44361", "issue:43688|issue:43749", + "issue:43653|issue:44361", "issue:38175|issue:43295", "issue:43819|issue:44811", "issue:43994|issue:44898", "issue:45372|issue:46170", - "issue:46832|issue:46932", "issue:43525|issue:46701", + "issue:46832|issue:46932", "issue:43531|issue:43742", - "issue:43295|issue:45356", - "issue:45542|issue:46178" + "issue:45542|issue:46178", + "issue:43295|issue:45356" ], - "prepared_review_unit_hash": "f7f4c3108b05fe5fd5d34ba4cf286d5df7573d1afc54c48f0d15f43b95d346d5", + "prepared_review_unit_hash": "740d7d9bc1e53f04cd54ce01bb7e7a7a2b120f25ef02afd8cf16843cf38d699e", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 36111, - "estimated_input_tokens": 9028, - "estimated_eval_tokens": 18312 + "serialized_chars": 36113, + "estimated_input_tokens": 9029, + "estimated_eval_tokens": 18314 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issues with one plausible duplicate pair around v5 tied-weights serialization. The rest appear to target distinct bugs or model families.", - "confidence": 0.78, - "canonical_issue_reason": "#45005 is the best canonical for the only plausible duplicate pair (#44466/#45005): it is broader and explicitly frames the shared v5 tied-weights problem.", + "summary": "Mostly heterogeneous issues with no single cluster-wide theme. The only plausible duplicate pair is the tied-weights/save_pretrained regression in v5 (#44466 and #45005); the rest are separate bugs across unrelated code paths.", + "confidence": 0.63, + "canonical_issue_reason": "#45005 is the broader umbrella report for the tied-weights regression in v5, and #44466 reads like a specific lm_head/device manifestation of the same save_pretrained problem.", "canonical_pr_reason": null, - "best_issue_reason": "#45005 is the most suitable representative because it is broader than #44466 while describing the same tied-weights regression more generally.", + "best_issue_reason": "#45005 is the best representative issue in this set because it captures the broader tied-weights failure mode that can subsume the more specific serialization symptom in #44466.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44737", - "right": "issue:45588", + "left": "issue:45357", + "right": "issue:45920", "accept": false, - "reason": "Different subsystems and failures: XLNet CPU device placement vs flash_attention s_aux=None crash." + "reason": "Different bugs: Qwen3.5 visual-encoder key serialization vs broad AutoTokenizer token-ID regression." }, { "left": "issue:46463", "right": "issue:46899", "accept": false, - "reason": "Unrelated bugs: T5 decoder-input validation message vs Gemma4 quantized multimodal dtype casting." + "reason": "Unrelated failures: T5 decoder-input validation message vs Gemma4 4-bit multimodal dtype casting." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Different errors and code paths: missing PyTorch import vs Trainer GPU broadcast NoneType handling." + "reason": "ImportError for missing PyTorch vs Trainer GPU broadcast NoneType error are unrelated." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "Both touch rope/config handling, but one is reload persistence and the other is JSON list type validation; not the same bug." + "reason": "Both touch RoPE-related config handling, but one is GPTNeoX config reload and the other is a validation helper TypeError; not the same concrete bug." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "Distinct regressions: RTDetrV2 output changes vs Qwen3.5 save_pretrained checkpoint corruption." + "reason": "RT-DETR output differences and Qwen3.5 checkpoint saving are separate regressions." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "Different model areas and failures: BeitImageProcessorFast label reduction vs zero-loss sequence classification." + "reason": "BEiT image label reduction bug and sequence-classification zero-loss bug are unrelated." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Packed-weight loading bug in BitNet vs Perceiver interpolation failure; not the same issue." + "reason": "BitNet packed-weight loading and Perceiver interpolation failure are different model/code paths." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Different config bugs: generic trust_remote_code field loss vs one specific Qwen3_5MoeVisionConfig field dropped by @strict." + "reason": "Config-field loss from trust_remote_code and Qwen3.5 vision config field dropping are different issues." }, { "left": "issue:45200", "right": "issue:46463", "accept": false, - "reason": "Different model families and symptoms: Gemma4 mm_token_type_ids defaulting vs T5 decoder validation error message." + "reason": "Gemma4 mm_token_type_ids defaulting and T5 decoder validation error are unrelated." }, { "left": "issue:44466", "right": "issue:45005", "accept": true, - "reason": "Both describe the same v5 tied-weights serialization regression; #45005 is the broader formulation." + "reason": "Both describe the same v5 tied-weights/save_pretrained regression; #44466 is a specific lm_head/device symptom, while #45005 is the broader tied-weights translation-model report." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Unrelated runtime failures: ZeRO-3 loading index error vs trainer.evaluate() after train." + "reason": "Gemma3 ZeRO-3 loading index error and trainer.evaluate() failing after train() are unrelated." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Different serialization/quantization problems: unexpected position-id keys vs sinq save/load breakage." + "reason": "Unexpected position-id key warnings and sinq quantized save/load breakage are different bugs." }, { "left": "issue:43761", "right": "issue:46829", "accept": false, - "reason": "Both mention hidden_states, but the affected models and output mistakes differ." + "reason": "Both concern output fields, but CLIPVision hidden_states regression and VideoPrism output misassignment are different model implementations and not the same bug." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Different loading regressions: gguf_file handling vs offline mistral regex hub lookup." + "reason": "gguf_file loading breakage and mistral regex offline failure are separate AutoTokenizer code paths." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Separate tokenizer-loading failures for different tokenizers and code paths." + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer load failure are unrelated tokenizer bugs." }, { - "left": "issue:43653", - "right": "issue:44361", + "left": "issue:43688", + "right": "issue:43749", "accept": false, - "reason": "BigBird special-token decode bug is unrelated to MLukeTokenizer task errors." + "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated." }, { - "left": "issue:43688", - "right": "issue:43749", + "left": "issue:43653", + "right": "issue:44361", "accept": false, - "reason": "Auxiliary-loss normalization vs FSDP CPU RAM efficient loading are different problems." + "reason": "BigBirdTokenizer special-token decode bug and MLukeTokenizer task AttributeError do not share the same underlying cause." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Both involve SigLIP2-ish usage, but one is zero probabilities while the other is processor.tokenizer regression." + "reason": "SigLIP2 zero-probabilities output and processor.tokenizer regression are different symptoms and code paths." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "Different functionality: DAC latent reconstruction vs Whisper batch_decode skip_special_tokens." + "reason": "DAC from_latents mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different model/pipeline failures: nonsensical SigLIP2 outputs vs Perceiver non-default-resolution error." + "reason": "SigLIP2 nonsensical outputs and Perceiver image-classification resolution handling are different bugs." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Separate dependency issues: mistral_common ReasoningEffort import vs Windows ROCm torch.distributed dependency." + "reason": "Gemma4 processor import failure from mistral_common and Windows ROCm torch.distributed dependency problems are unrelated." }, { - "left": "issue:46832", - "right": "issue:46932", + "left": "issue:43525", + "right": "issue:46701", "accept": false, - "reason": "RT-DETR feature-level TypeError is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "Llama4Config pad_token_id missing and DiffusionGemmaGenerationOutput pipeline failure are separate issues." }, { - "left": "issue:43525", - "right": "issue:46701", + "left": "issue:46832", + "right": "issue:46932", "accept": false, - "reason": "Llama4Config pad_token_id attribute bug is unrelated to DiffusionGemma image-text pipeline failure." + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedder dtype casting are unrelated." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window issue is not the same as MobileLLM-125M key error." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM load key error are different bugs." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:45542", + "right": "issue:46178", "accept": false, - "reason": "Both are regressions in tokenization/processing, but they affect different models and different root causes." + "reason": "TensorBoard-without-TensorFlow backend error and tensorflow_text backend mapping error are separate optional-dependency issues." }, { - "left": "issue:45542", - "right": "issue:46178", + "left": "issue:43295", + "right": "issue:45356", "accept": false, - "reason": "Both are backend/optional-dependency complaints, but they are about different backend names and failure modes." + "reason": "Processor/tokenizer regression in v4.57.5 and Kimi-K2.5 codec handling regression are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The judgment is appropriately conservative and grounded in the packet. The only accepted pair, issue:44466/issue:45005, is well-supported by both titles as the same v5 tied-weights regression, with #45005 reasonably broader. The remaining soft-edge rejections are cautious and do not overstate duplicate evidence." + "feedback": "The summary is well grounded in the packet titles, and the only accepted pair (#44466/#45005) is a conservative match on the tied-weights/save_pretrained regression. The other soft-edge rejections are appropriately cautious and do not overclaim shared causality." }, "evaluator_used": true, "retried": false, @@ -7855,15 +7848,15 @@ "issue:43688|issue:45704", "issue:42898|issue:46491", "issue:44843|issue:44991", - "issue:43756|issue:46402", "issue:43452|issue:43577", + "issue:43756|issue:46402", "issue:43688|issue:45701", "issue:44464|issue:44743", "issue:43653|issue:43927", "issue:44479|issue:44625", "issue:45491|issue:46002", - "issue:44484|issue:44704", "issue:44821|issue:46821", + "issue:44484|issue:44704", "issue:33666|issue:36010", "issue:43688|issue:45440", "issue:32090|issue:33666", @@ -7876,23 +7869,23 @@ "issue:43504|issue:43720", "issue:43577|issue:44554", "issue:43611|issue:46846", - "issue:43643|issue:44561", - "issue:43782|issue:46693" + "issue:43782|issue:46693", + "issue:43643|issue:44561" ], - "prepared_review_unit_hash": "54a762f3a676e8aaf7a309ede324b4752e13d1f72d8a4872e47db38aed44632f", + "prepared_review_unit_hash": "2a34ac50cd34fcc38da09f488353b9ca77f355d008c083f626df99f0e40795e5", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36487, + "serialized_chars": 36488, "estimated_input_tokens": 9122, "estimated_eval_tokens": 18500 }, @@ -7901,186 +7894,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a true duplicate cluster: the issues span unrelated tokenizer, loading, generation, multimodal, and distributed-training bugs. The soft pairs mostly share broad vocabulary only, so none should be merged as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "This set is mostly a grab bag of distinct regressions across tokenization, loading, generation, and model-specific code paths. The soft pairs only share broad subsystem similarity, not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.42, + "canonical_issue_reason": "issue:45701 is the broadest umbrella here ('transformers version changes the tokenization') and works best as a representative bucket for the tokenizer-regression cluster; the other tokenizer issues are narrower and appear to have different root causes.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45701 is the strongest global representative because it is generic enough to cover the shared tokenizer-regression theme without overcommitting to one specific model or failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both concern tokenizer special tokens, but one is BigBird mask-token registration and the other is mDeBERTa BOS/EOS handling; different code paths and failures." + "reason": "Both involve token handling, but one is BigBird special-token registration and the other is mDeBERTa add_special_tokens/BOS-EOS behavior; different model paths and symptoms." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "One is a Transformers 5.0 custom model-init regression; the other is broken FSDP CPU RAM-efficient loading. Different subsystems." + "reason": "Custom model init under v5 and FSDP CPU-efficient loading are unrelated failure modes with different code paths." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Tokenizer inconsistency vs a MacOS bus error on a pretrained CLIP model; no shared underlying bug." + "reason": "Tokenizer slow/fast inconsistency and a MacOS bus error on a CLIP model are not the same bug." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models is unrelated to T5 using apex FusedRMSNorm and leaking memory." + "reason": "Auxiliary-loss normalization in MoE models and a T5/apex RMSNorm memory leak are different issues." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Both are tokenization regressions, but one is a global v5 clean-up-space behavior change and the other is a CodeLlama leading-space round-trip bug." + "reason": "Both concern tokenization behavior changes, but one is clean_up_tokenization_spaces and the other is CodeLlama leading-space round-trip regression; distinct fixes." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Offline-mode hub lookup in _patch_mistral_regex is unrelated to EMBEDDIA/est-roberta tokenizer loading failure." + "reason": "Offline-mode breakage from an unconditional model_info call is a different root cause than EMBEDDIA/est-roberta tokenizer loading failure." }, { - "left": "issue:43756", - "right": "issue:46402", + "left": "issue:43452", + "right": "issue:43577", "accept": false, - "reason": "Smollm3 RoPE layer loading bug and Sam3Model text-encoder weight loading bug are different model-specific failures." + "reason": "gguf_file loading breakage and Blip2 dtype not propagating are separate model-loading bugs." }, { - "left": "issue:43452", - "right": "issue:43577", + "left": "issue:43756", + "right": "issue:46402", "accept": false, - "reason": "gguf_file loading breakage and Blip2 dtype propagation are unrelated loading issues." + "reason": "Smollm3 RoPE-layer dropping and Sam3Model text-encoder weights not loading are unrelated model-specific defects." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Auxiliary-loss normalization bug is not the same as a general tokenization-version regression." + "reason": "A MoE auxiliary-loss normalization bug is not the same as a generic tokenization-version regression." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different generation bugs in different paths." + "reason": "Compiled chunked generation inconsistency and recurrent-state reset with cache are both generation bugs, but they hit different mechanisms and code paths." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBirdTokenizer special-token registration and DiaConfig save/load token IDs are separate tokenization/config bugs." + "reason": "BigBird special-token registration and DiaConfig save/load token-id loss are different token/config persistence problems." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression in Qwen vision-language models is not the same as num_labels propagation in Qwen3.5 config." + "reason": "A video-input regression in Qwen VL models and missing num_labels propagation in Qwen3.5 are separate bugs." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Gemma3 NaNs in embeddings on GPU are not the same issue as NaNs in classification heads after checkout/init." + "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on GPU and the other is classification-head init/checkpoint NaNs; different root causes." }, { - "left": "issue:44484", - "right": "issue:44704", + "left": "issue:44821", + "right": "issue:46821", "accept": false, - "reason": "save_pretrained shard-size policy is unrelated to AutoProcessor kwargs not being forwarded to cached_file." + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated." }, { - "left": "issue:44821", - "right": "issue:46821", + "left": "issue:44484", + "right": "issue:44704", "accept": false, - "reason": "Loading AutoImageProcessor from URL and TimesFM window_size AttributeError are unrelated API failures." + "reason": "save_pretrained shard-size behavior and AutoProcessor.cached_file kwargs passing are separate serialization/loading concerns." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training and a GenerationMixin import error are different problems with no common code path." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are completely different issues." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "MoE auxiliary-loss normalization is unrelated to DeepseekV3MoE diverging from the remote implementation." + "reason": "MoE auxiliary-loss normalization and DeepseekV3MoE divergence from remote implementation are not the same bug." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error and Qwen2-VL multi-GPU training are both distributed-training themed but not the same bug." + "reason": "Trainer gpu_broadcast NoneType failure is unrelated to Qwen2-VL multi-GPU training." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "Inconsistent lm_head serialization by device is separate from sinq quantization save/load breakage." + "reason": "Device-dependent lm_head serialization and sinq quantization save/load breakage are different persistence bugs." }, { "left": "issue:44977", "right": "issue:46693", "accept": false, - "reason": "Qwen3.5 flash-attention correctness failure and FA2 long-context performance regression are different symptoms and likely different fixes." + "reason": "Flash-attention generation correctness for Qwen3.5 and long-context FA2 performance regression are different failure modes." }, { "left": "issue:44734", "right": "issue:46612", "accept": false, - "reason": "KV-cache continuation crash in serve is not the same as beam-search cache reorder being skipped for certain models." + "reason": "KV-cache continuation indexing in transformers serve and cache-reorder being skipped for legacy models are separate generation-cache bugs." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "Chunked compiled generation inconsistency is unrelated to Qwen3.5 flash-attention generation failure." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and compiled-forward chunked generation are different code paths and failures." + "reason": "Qwen3 auto device-mapping cudaErrorAssert and chunked generation inconsistency are different problems." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Beit legacy-field loading and Bamba torch.compile/SDPA failure are unrelated model-specific bugs." + "reason": "Beit legacy-field loading and Bamba torch.compile+SDPA failure are unrelated." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Legacy-field model loading and BitNet packed-weight unpacking during accelerate loading are different issues." + "reason": "Beit preset loading and BitNet packed-weight unpacking during accelerate loading are different code paths." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype propagation and MPS attention correctness with mismatched head dims are unrelated." + "reason": "Blip2 dtype propagation and an MPS attention correctness issue are not the same bug." }, { "left": "issue:43611", "right": "issue:46846", "accept": false, - "reason": "base_model_prefix loading breakage is not the same as missing linear_attn entries in Qwen3.5 TP planning." + "reason": "base_model_prefix loading breakage and Qwen3.5 tensor-parallel plan omissions are separate v5 regressions." }, { - "left": "issue:43643", - "right": "issue:44561", + "left": "issue:43782", + "right": "issue:46693", "accept": false, - "reason": "Missing fields from trust_remote_code AutoConfig and removal of is_torch_fx_available are separate v5 regressions." + "reason": "weight_only=True loading failure and FA2 long-context performance regression are unrelated." }, { - "left": "issue:43782", - "right": "issue:46693", + "left": "issue:43643", + "right": "issue:44561", "accept": false, - "reason": "Qwen3VL weight_only load error is unrelated to the FA2 long-context performance regression." + "reason": "Missing fields from AutoConfig with trust_remote_code and removal of is_torch_fx_available are both v5 trust_remote_code regressions, but they are different root causes and not one concrete bug." } ] }, @@ -8156,8 +8149,8 @@ "soft_pairs": [ "issue:33666|issue:36331", "issue:43278|issue:43688", - "issue:43720|issue:43927", "issue:43262|issue:43295", + "issue:43720|issue:43927", "issue:43504|issue:44610", "issue:43606|issue:43927", "issue:43493|issue:43825", @@ -8180,263 +8173,233 @@ "issue:43646|issue:44466", "issue:46209|issue:46710", "issue:33357|issue:33666", - "issue:44568|issue:45812", "issue:43278|issue:43701", + "issue:44568|issue:45812", "issue:44488|issue:44843", "issue:44849|issue:46209", "issue:44792|issue:45405", "issue:43927|issue:44361" ], - "prepared_review_unit_hash": "cb6e76d20a072719c701ea4a6b9c39120b337ad04d7ee04135a2e23508f13136", + "prepared_review_unit_hash": "ad3b3b9a0012997e1016af03b2e91ed7078dcee5ea00705065b74f502bd3c291", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37422, - "estimated_input_tokens": 9356, - "estimated_eval_tokens": 18968 + "serialized_chars": 37420, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 }, "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, with only a loose thematic overlap around tokenizer/model-regression reports. I would treat it as noisy rather than a true duplicate cluster.", - "confidence": 0.67, - "canonical_issue_reason": "issue:45812 is the broadest and most central issue in the set: it describes a silent v4\u2192v5 tokenizer regression affecting all Granite models, with clear user impact and multiple references.", + "summary": "This cluster is mostly false-positive similarity: the issues span unrelated tokenizer, config, loading, training, and hardware-specific bugs. No soft pair looks like the same concrete underlying change. Issue 45812 is the best representative because it is the broadest tokenizer regression in the set.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45812 is the most central and broadly scoped bug here: a silent v4\u2192v5 tokenizer regression affecting all Granite models. It best represents the tokenizer-compatibility subset of this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest representative bug report here because it is concrete, regression-focused, and plausibly connected to several other tokenizer-related complaints in the group.", + "best_issue_reason": "issue:45812 is the strongest standalone tracker because it describes a clear, user-facing, high-blast-radius tokenizer bug and is broader than the narrower model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Different problems: multi-GPU training support vs a Trainer custom loss signature mismatch." + "reason": "Unrelated problems: Qwen2-VL multi-GPU training versus a CustomTrainer compute_loss signature mismatch." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Unrelated bugs: dtype drift in eval vs auxiliary-loss normalization." + "reason": "Different bugs: embedding dtype drift in evaluate versus auxiliary-loss normalization in MoE models." }, { - "left": "issue:43720", - "right": "issue:43927", + "left": "issue:43262", + "right": "issue:43295", "accept": false, - "reason": "Different code paths: packed-weight loading vs config losing custom token IDs on save/load." + "reason": "Audio processor sampling-rate default issue versus a multimodal processor.tokenizer regression; different code paths." }, { - "left": "issue:43262", - "right": "issue:43295", + "left": "issue:43720", + "right": "issue:43927", "accept": false, - "reason": "Both involve processors, but one is audio chat-template sampling rate and the other is tokenizer/image handling regression." + "reason": "BitNet packed-weights loading failure and DiaConfig custom token ID persistence are unrelated." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Unrelated model-loading issues: legacy BEiT field vs OmDet-Turbo processor image size mismatch." + "reason": "BEiT pretrained preset loading bug versus OmDet-Turbo image-size mismatch; different models and failure modes." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "Different failure modes: CPU offload device mismatch vs DiaConfig token-ID persistence." + "reason": "CPU offload device mismatch is unrelated to DiaConfig token-ID loss on save/load." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation discrepancy is unrelated to pipeline translation-task messaging." + "reason": "SigLIP2 model-vs-JAX discrepancy is not the same as a pipeline error message about translation support." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both hit apply_chat_template, but the triggering inputs and failure modes differ materially." + "reason": "Both are apply_chat_template-related, but one is batched padding=False crash and the other is assistant tool-call/no-content handling." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Packed-weight loading bug is unrelated to MLukeTokenizer task AttributeError." + "reason": "Packed-weight loading in BitNet and MLukeTokenizer task AttributeError are different bugs." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model-loading regression vs a dependency-version bump issue." + "reason": "Phi-4 multimodal loading failure is unrelated to a bumped MIN_PEFT_VERSION dependency issue." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness is unrelated to importing Trainer being blocked by datasets." + "reason": "MPS attention correctness and Trainer import failure due to datasets are unrelated." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Different model families and bugs: num_labels propagation vs Whisper empty-transcription generation." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription handling are different model-specific bugs." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Similar pattern, but different missing fields on different configs; not the same concrete bug." + "reason": "Both involve missing pad token fields, but they affect different configs and fields in different models." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation issue is unrelated to FP8 inference on B200." + "reason": "Qwen3.5 flash-attention generation bug and FP8 inference on B200 are unrelated hardware/back-end issues." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "Both are from pretrained loading, but the model/tokenizer-specific causes are different." + "reason": "gguf_file loading regression is not the same as failing to load a specific pretrained model." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Device-map forward mismatch is unrelated to sinq quantized save/load corruption." + "reason": "Gemma4 device-map mismatch and sinq save/load quantization break are different problem classes." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug and Mac bus error are clearly unrelated." + "reason": "Tokenizer spacing after added token is unrelated to a MacOS bus error loading a CLIP model." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 inference mismatch vs return_tensors being ignored are different issues." + "reason": "Granite token IDs regression and return_tensors being ignored when text_kwargs is passed are different API bugs." }, { "left": "issue:42796", "right": "issue:44977", "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to Qwen3.5 flash-attention generation." + "reason": "Ministral BPE-marker handling and Qwen3.5 flash-attention generation are unrelated tokenizer vs attention issues." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "Model implementation discrepancy vs training instability under ZeRO-3 are different problems." + "reason": "SigLIP2 implementation discrepancy is not the same as gradient explosion under HfDeepSpeedConfig + ZeRO-3." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Doc-example errors are not the same bug as processor/tokenizer regression." + "reason": "SigLIP2 documentation/example errors and processor.tokenizer access regression are different bugs." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Eval dtype drift and exploding gradients under DeepSpeed are unrelated." + "reason": "Embedding dtype change in evaluation is unrelated to abnormal gradient growth during ZeRO-3 training." }, { "left": "issue:44568", "right": "issue:46489", "accept": false, - "reason": "Both tokenizer regressions, but on different models and different symptoms; not one concrete bug." + "reason": "Different tokenizer regressions: missing BOS/EOS for mdeberta-v3-base versus incorrect DeepSeek-Coder v1 output." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model init breakage is unrelated to inconsistent lm_head.weight serialization." + "reason": "Custom model initialization breakage and lm_head.weight serialization inconsistency are separate v5 issues." }, { "left": "issue:46209", "right": "issue:46710", "accept": false, - "reason": "FP8 inference on B200 is unrelated to the DeepSeek-R1 accuracy regression." + "reason": "FP8 inference failure on B200 and DeepSeek-R1-Distill-Llama accuracy regression are unrelated." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "Mac bus error and multi-GPU training are unrelated." + "reason": "MacOS bus error in CLIP loading is unrelated to Qwen2-VL multi-GPU training." }, { - "left": "issue:44568", - "right": "issue:45812", + "left": "issue:43278", + "right": "issue:43701", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different tokenization behaviors." + "reason": "Embedding dtype drift and resume_from_checkpoint key mismatch do not share an underlying bug." }, { - "left": "issue:43278", - "right": "issue:43701", + "left": "issue:44568", + "right": "issue:45812", "accept": false, - "reason": "Embedding dtype drift is unrelated to resume_from_checkpoint key mismatch." + "reason": "Both are tokenizer bugs, but they affect different models and mechanisms: special-token addition versus Granite token ID corruption." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Different loading failures: one is a model-specific load problem, the other is offline-mode breakage in _patch_mistral_regex." + "reason": "Model loading failure for sleng-bert is unrelated to an unconditional model_info() call breaking offline mode." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 hidden-states bug is unrelated to FP8 inference failure." + "reason": "Qwen3.5 output_hidden_states bug and FP8 inference on B200 are different issues." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Janus test failure is unrelated to the PEFT version bump." + "reason": "A janus test failure is not the same as the MIN_PEFT_VERSION dependency bump." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Config token-ID persistence and MLukeTokenizer task AttributeError are different bugs." - }, - { - "left": "issue:44964", - "right": "issue:45405", - "accept": false, - "reason": "Model loading failure and minimum PEFT version mismatch are unrelated." - }, - { - "left": "issue:43504", - "right": "issue:43646", - "accept": false, - "reason": "Legacy-field load failure and custom model initialization regression are different code paths." - }, - { - "left": "issue:43262", - "right": "issue:43278", - "accept": false, - "reason": "Audio chat-template sampling-rate default is unrelated to eval dtype changes." - }, - { - "left": "issue:43825", - "right": "issue:44977", - "accept": false, - "reason": "Pipeline error-message text is unrelated to Qwen3.5 flash-attention generation." - }, - { - "left": "issue:43606", - "right": "issue:43688", - "accept": false, - "reason": "CPU offload device mismatch and auxiliary-loss normalization are unrelated." + "reason": "DiaConfig custom token IDs are a serialization/config issue, not the same as MLukeTokenizer task AttributeError." } ] }, @@ -8518,8 +8481,8 @@ "issue:43257|issue:43299", "issue:45245|issue:46710", "issue:44493|issue:46285", - "issue:46399|issue:46846", "issue:43334|issue:45759", + "issue:46399|issue:46846", "issue:43334|issue:43572", "issue:43606|issue:44112", "issue:44589|issue:45520", @@ -8531,8 +8494,8 @@ "issue:43844|issue:44964", "issue:43873|issue:46050", "issue:43994|issue:44493", - "issue:45200|issue:45229", "issue:45229|issue:45823", + "issue:45200|issue:45229", "issue:44869|issue:46752", "issue:45588|issue:46050", "issue:42490|issue:43696", @@ -8540,208 +8503,208 @@ "issue:29942|issue:30333", "issue:44977|issue:46701" ], - "prepared_review_unit_hash": "9c74dab803e12cc202acda7a0b16f8f1af9dcdbc34b2c26c76fd0760759ec6ef", + "prepared_review_unit_hash": "6d55b4561c121930326fbf6ee6597c35e9e658a0e1f3cbb53d23eab7c01c847d", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36490, - "estimated_input_tokens": 9123, - "estimated_eval_tokens": 18502 + "serialized_chars": 36481, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The suggested soft pairs are mostly broad topical similarities, not the same underlying bug or change. I would not merge any of them as duplicates; this cluster looks like a noisy mix of unrelated Transformers issues.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue is a true canonical duplicate for this cluster because the items span many unrelated failure modes; if a representative is needed, issue 43257 is the clearest concrete bug report, but it does not subsume the others.", + "summary": "This cluster is mostly a loose thematic grouping of unrelated Transformers issues: model-loading regressions, config/version mismatches, device/offload problems, and a few CI/runtime failures. The soft pairs are generally similar by subsystem or model family, but not the same underlying bug/change.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue clearly subsumes the cluster; 43299 is the closest representative regression-style report, but the items are not true duplicates of one bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43257 is the strongest standalone issue in the set: it is concrete, reproducible, and has enough discussion to be a useful representative\u2014but only as a representative, not as a duplicate target for the whole cluster.", + "best_issue_reason": "issue:43299 is the most representative high-signal report in the set: a clear Transformers dev regression affecting model loading, with a concrete failure mode and broad enough framing to anchor the cluster better than the more niche reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "One is a Qwen2 TP OOM report; the other is a packaging/versioning problem about MIN_PEFT_VERSION. Different bugs and code paths." + "reason": "Different problems: TP loading OOM vs an unreleased PEFT version bump." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Hidden-states output behavior in Qwen3.5 is unrelated to saving incorrect visual encoder keys during save_pretrained." + "reason": "Same model family, but different bugs and code paths: hidden-states handling vs save_pretrained key mapping." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "ImportError for a missing Qwen2.5-VL class and a Float8 storage lookup error are unrelated failures." + "reason": "Unrelated failures: missing Qwen2.5-VL import vs a Float8 storage lookup/type error." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "A legacy-field load issue in BEiT segmentation and a CPU offload device-mismatch crash in bark-small are different subsystems and bugs." + "reason": "Different models and failure modes: BEiT legacy-field loading vs bark-small CPU offload device mismatch." }, { "left": "issue:46821", "right": "issue:46832", "accept": false, - "reason": "TimesFM window_size validation and RT-DETR feature-level handling are unrelated model-specific errors." + "reason": "Different architectures and bugs: TimesFM window_size vs RT-DETR feature-level mismatch." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen3/MoE, but one is about loading with accelerate+deepspeed and the other is a transformers-dev loading regression; not the same concrete bug." + "reason": "Both involve Qwen3 MoE loading, but the triggers differ materially: accelerate+deepspeed conversion vs dev-model loading regression." }, { "left": "issue:45245", "right": "issue:46710", "accept": false, - "reason": "Category-count overflow and an accuracy regression after upgrading transformers are not the same issue." + "reason": "Completely different issues: category-cardinality limit vs accuracy regression after upgrade." }, { "left": "issue:44493", "right": "issue:46285", "accept": false, - "reason": "A position-id key warning and CLIP text-model structure inconsistency are different problems with different fix paths." + "reason": "Different structural regressions: unexpected position-id key vs CLIP text-model class structure change." }, { - "left": "issue:46399", - "right": "issue:46846", + "left": "issue:43334", + "right": "issue:45759", "accept": false, - "reason": "Missing decorator for apply_rotary_pos_emb registration is unrelated to missing linear_attn entries in the TP plan." + "reason": "Both are Qwen multimodal loading issues, but one is a missing config attribute and the other is a factory/unwrapping bug." }, { - "left": "issue:43334", - "right": "issue:45759", + "left": "issue:46399", + "right": "issue:46846", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL text config and not unwrapping text_config in AutoModelForCausalLM are related only at a high level; they are different concrete failures." + "reason": "Related model family, but different concrete fixes: kernel registration decorator vs TP plan coverage for linear_attn." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Qwen3-VL config loading and StableLmConfig missing pad_token_idx are separate model-specific compatibility bugs." + "reason": "Different models and different missing config fields." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "A real device-mismatch crash in model loading is not the same as a stale CI test for a device override." + "reason": "Bark device-mismatch bug vs a GraniteSpeech CI test issue; not the same underlying defect." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage lookup failure and a Python 3.13 flash_attn import KeyError are unrelated environment/runtime issues." + "reason": "Different runtime failures: Float8 storage missing vs flash_attn import-utils KeyError on Python 3.13." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "StableLmConfig field compatibility and init_empty_weights receiving an unexpected argument are different API regressions." + "reason": "Both are version-related breakages, but one is a missing config field and the other is an init_empty_weights signature mismatch." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "BEiT legacy config loading and BigBirdTokenizer special-token registration are unrelated bugs." + "reason": "Different components and symptoms: BEiT preset loading vs BigBirdTokenizer special-token registration." }, { "left": "issue:45072", "right": "issue:46832", "accept": false, - "reason": "bfloat16 dtype mismatches in two wrappers and RT-DETR backbone/feature-level errors do not share the same fix." + "reason": "Different bug classes: dtype mismatch in inference vs RT-DETR backbone feature-count mismatch." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Chunked generation with compiled forward is not the same as cache max-length being ignored outside quantized cache." + "reason": "Both touch generation/cache behavior, but one is compiled-forward inconsistency and the other is cache-length config handling." }, { "left": "issue:43937", "right": "issue:46701", "accept": false, - "reason": "Invalid GenerationConfig for GLM-5 and an image-text-to-text pipeline failure are different code paths." + "reason": "Unrelated failures: invalid GenerationConfig vs a broken image-text-to-text pipeline." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient growth in training is unrelated to loading Phi-4 multimodal with latest transformers." + "reason": "Different models and root causes: ZeRO-3 gradient anomaly vs latest-transformers model loading failure." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization offloading not working and sinq save/load breakage are both quantization-related, but they are not the same failure mode." + "reason": "Both involve quantization/offloading, but one is 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 not the same bug." + "reason": "Different bugs: Siglip2 pipeline/AutoModel behavior vs unexpected position-id key handling." }, { - "left": "issue:45200", - "right": "issue:45229", + "left": "issue:45229", + "right": "issue:45823", "accept": false, - "reason": "Defaulting mm_token_type_ids to zeros and a multi-GPU CUDA OOM are unrelated issues." + "reason": "Same model family, but one is OOM from scale and the other is a device mismatch during forward." }, { - "left": "issue:45229", - "right": "issue:45823", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "Both mention Gemma4, but one is an OOM during inference and the other is a device-map forward mismatch; different fixes." + "reason": "Text-only fine-tuning defaults vs multi-GPU inference OOM are unrelated." }, { "left": "issue:44869", "right": "issue:46752", "accept": false, - "reason": "Whisper timestamp decoding crash and empty-list chat-template IndexError are unrelated input-edge-case bugs." + "reason": "Different decoding/input-edge-case bugs in unrelated components." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "Flash-attention sinkless-model crash and quantized save/load breakage are different features and failure paths." + "reason": "Flash-attention sink-less-model crash vs quantized model save/load breakage are separate bugs." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "A local-path import warning and a GPT-oss CUDA OOM are unrelated." + "reason": "Import warning for trailing-slash paths is unrelated to a GPT-oss CUDA OOM." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Tokenizer slow/fast inconsistency and a trainer _gpu_broadcast_one NoneType error are different issues." + "reason": "Tokenizer fast/slow inconsistency is unrelated to the trainer _gpu_broadcast_one NoneType error." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 test failures and MLflow reporting on failed jobs are unrelated." + "reason": "Flash Attention 2 CI failures and MLflow reporting on failed jobs are different issues." }, { "left": "issue:44977", "right": "issue:46701", "accept": false, - "reason": "Qwen3.5 flash-attention generation failure and a DiffusionGemma pipeline failure are different model-specific bugs." + "reason": "Different models and code paths: Qwen3.5 flash-attention generation vs DiffusionGemma image-text pipeline failure." } ] }, @@ -8826,95 +8789,95 @@ "issue:43404|issue:46710", "issue:45072|issue:45198", "issue:43295|issue:43450", - "issue:45588|issue:45759", "issue:37428|issue:43957", + "issue:45588|issue:45759", "issue:43540|issue:45200", "issue:43295|issue:43873", "issue:44186|issue:46192", "issue:45823|issue:46192", "issue:43295|issue:44610", "issue:43572|issue:46821", - "issue:45200|issue:45823", "issue:43550|issue:43844", + "issue:45200|issue:45823", "issue:29870|issue:29942", "issue:43450|issue:43575", "issue:45520|issue:46178", "issue:43572|issue:44987", + "issue:37428|issue:43824", "issue:42490|issue:43493", "issue:44554|issue:45356", - "issue:37428|issue:43824", "issue:39692|issue:46192" ], - "prepared_review_unit_hash": "151dbff751ea8e8dc6171cfba483d7b6f9be604466010b391049b6f94dd666a0", + "prepared_review_unit_hash": "27d192f34deb934a7e94bcd819bff452281477b4ad1bfba9388b5e42901fbdd9", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36325, - "estimated_input_tokens": 9082, - "estimated_eval_tokens": 18420 + "serialized_chars": 36330, + "estimated_input_tokens": 9083, + "estimated_eval_tokens": 18422 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous; the soft links mostly connect superficially similar tokenizer/processor/model-loading issues rather than the same underlying bug. No PRs are present.", - "confidence": 0.31, - "canonical_issue_reason": "Issue 43295 is the broadest cross-cutting regression in the set, centered on processor/tokenizer behavior and multimodal custom code breakage, which best matches the cluster\u2019s loose theme.", + "summary": "This cluster is mostly heterogeneous. A few items are thematically adjacent around multimodal processor/tokenizer behavior, but none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43295 is the broadest representative of the recurring multimodal processor/tokenizer regression theme and sits closest to the few adjacent soft matches.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43295 is the best representative issue because it captures a high-level multimodal processor/tokenizer regression with wider applicability than the more model-specific reports.", + "best_issue_reason": "issue:43295 best represents the main subtheme in this cluster; it is broad, user-facing, and closer to multiple nearby processor-related reports than the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Different failures in unrelated areas: compiler/template-node parsing vs flash_attn import/version handling." + "reason": "Different failures: template-node compilation vs import-time backend lookup. Same broad area only by coincidence." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing bug and TrainerState token counter update are unrelated." + "reason": "Tokenizer spacing bug and TrainerState counter update are unrelated code paths." }, { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Both mention SAM3, but one is a dummy Conv2d implementation detail and the other is a memory overflow; not the same bug." + "reason": "Both mention Sam3, but one is a dummy layer concern and the other is a memory overflow report; not the same bug." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export issue and SigLIP2 zero-probabilities issue are different code paths and symptoms." + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability behavior." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness and a flex_attention import version check are separate problems." + "reason": "Both touch attention/flex attention, but one is an MPS correctness issue and the other is a version-check import bug." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both affect serve, but one is continuous-batching multimodal input handling and the other is KV-cache continuation tensor indexing." + "reason": "Both are serve/runtime crashes, but one is multimodal input handling and the other is KV-cache continuation indexing." }, { "left": "issue:41720", "right": "issue:46710", "accept": false, - "reason": "CUDA device-map assert on Qwen3 is unrelated to an accuracy regression in DeepSeek after upgrade." + "reason": "A device-mapping CUDA assert on Qwen3 is not the same as an accuracy regression on DeepSeek after upgrade." }, { "left": "issue:43493", @@ -8926,43 +8889,43 @@ "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and rope-validation TypeError are distinct regressions." + "reason": "Gemma4Processor _tokenizer crash and rope-param validation type error are different subsystems." }, { "left": "issue:43404", "right": "issue:46710", "accept": false, - "reason": "Weight tying in Mistral3 and DeepSeek accuracy regression are not the same issue." + "reason": "Untied lm_head weights in Mistral3 is not the same problem as a model accuracy regression." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Dtype mismatches in inference and Wav2Vec2 save_pretrained/tokenization failures are unrelated." + "reason": "dtype mismatch in inference and Wav2Vec2 save/tokenization failure are unrelated." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer regression and batched video shape bug are different multimodal bugs." + "reason": "Both are processor-related, but one is processor.tokenizer/image handling and the other is batched video output shape; different bugs." }, { - "left": "issue:45588", - "right": "issue:45759", + "left": "issue:37428", + "right": "issue:43957", "accept": false, - "reason": "Flash-attention sinkless-model crash and Qwen multimodal config unwrapping are separate code-path issues." + "reason": "ImportError for a flash-attention symbol and meta-device loading failures are different issues." }, { - "left": "issue:37428", - "right": "issue:43957", + "left": "issue:45588", + "right": "issue:45759", "accept": false, - "reason": "ImportError for a flash-attention helper and meta-device model loading breakage are different problems." + "reason": "Both involve multimodal/serve or composite configs, but the crash conditions and code paths differ." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video input processing error in Qwen3OmniMoe and Gemma4 mm_token_type_ids defaulting are unrelated." + "reason": "Video input processing in Qwen3OmniMoe is unrelated to Gemma4 mm_token_type_ids defaults." }, { "left": "issue:43295", @@ -8974,61 +8937,67 @@ "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash and return_tensors being ignored are different tokenizer issues." + "reason": "LayoutLMv2Tokenizer crashes on NER/padding are unrelated to return_tensors being ignored with text_kwargs." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Gemma4 device-map mismatch and return_tensors handling are unrelated." + "reason": "Gemma4 device_map forward mismatch and return_tensors handling are different failure modes." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Both involve processors, but one is a tokenizer/processor regression while the other is an OmDet-Turbo image-size mismatch." + "reason": "Both are preprocessing/model-shape mismatches, but they affect different models and input pipelines." }, { "left": "issue:43572", "right": "issue:46821", "accept": false, - "reason": "StableLmConfig pad_token_idx missing and TimesFM window_size AttributeError are unrelated config issues." + "reason": "StableLmConfig pad_token_idx and TimesFM window_size AttributeError are unrelated config bugs." }, { - "left": "issue:45200", - "right": "issue:45823", + "left": "issue:43550", + "right": "issue:43844", "accept": false, - "reason": "Gemma4 token-type-id defaulting and Gemma4 PLE device mismatch are separate bugs." + "reason": "torch.compile/SDPA failure and DeepSpeed gradient overflow are different training/runtime issues." }, { - "left": "issue:43550", - "right": "issue:43844", + "left": "issue:45200", + "right": "issue:45823", "accept": false, - "reason": "torch.compile/SDPA failure and ZeRO-3 gradient increase are unrelated training/runtime issues." + "reason": "Same model family (Gemma4) but distinct bugs: token-type defaults vs device placement." }, { "left": "issue:29870", "right": "issue:29942", "accept": false, - "reason": "BART decoder_attention_mask assignment bug and Flash Attention 2 test failures are different issues." + "reason": "Bart decoder_attention_mask assignment and Flash Attention 2 test failures are unrelated." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Batched video processor shape bug and model-loading OOM are not the same underlying problem." + "reason": "Batched video processor shape bug and tensor-parallel OOM on Qwen2-57B are unrelated." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "flash_attn import_utils KeyError and tensorflow_text backend mapping error are different import/backend issues." + "reason": "Both are import/backend mapping issues, but they involve different backend registries and different symptoms." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLmConfig update regression and failure loading physical-intelligence/fast are unrelated." + "reason": "StableLmConfig pad_token_idx regression and loading physical-intelligence/fast on 5.1.0 are not the same bug." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different missing imports: flash-attention helper vs Qwen2_5_VLForConditionalGeneration." }, { "left": "issue:42490", @@ -9040,19 +9009,13 @@ "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness issue and Kimi tokenizer codec/regex regression are different problems." - }, - { - "left": "issue:37428", - "right": "issue:43824", - "accept": false, - "reason": "Different import failures: flash attention helper vs Qwen2.5 VL model symbol." + "reason": "MPS attention correctness and tokenizer codec regression are different problems." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 doc-example errors and ignored return_tensors behavior are unrelated." + "reason": "SigLIP2 documentation/model-processor mismatch and return_tensors handling are unrelated." } ] }, @@ -9130,42 +9093,42 @@ "issue:44521|issue:45290", "issue:44466|issue:45127", "issue:45381|issue:46209", - "issue:43653|issue:44112", "issue:43278|issue:43856", "issue:44737|issue:46894", "issue:28218|issue:35141", - "issue:43824|issue:43975", + "issue:43653|issue:44112", "issue:42915|issue:43824", + "issue:43824|issue:43975", "issue:43950|issue:46285", "issue:43452|issue:46512", "issue:43262|issue:43450", - "issue:45200|issue:46899", - "issue:45200|issue:46932", "issue:43334|issue:43525", + "issue:45200|issue:46899", "issue:44442|issue:45198", + "issue:45200|issue:46932", "issue:44792|issue:45563", - "issue:44479|issue:46192", "issue:44898|issue:45412", + "issue:44479|issue:46192", "issue:43819|issue:44112", "issue:45200|issue:45325", - "issue:43827|issue:43901", "issue:43335|issue:43550", - "issue:41720|issue:46209" + "issue:41720|issue:46209", + "issue:43827|issue:43901" ], - "prepared_review_unit_hash": "e6a90756bee7dd4f427404a447127a2708b318cf5065b9f4c7134c3d421ffaae", + "prepared_review_unit_hash": "d3e54d621410d443c06f487aeb437afd505349409aaf467067b41b4d43d2a1e1", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 35339, + "serialized_chars": 35340, "estimated_input_tokens": 8835, "estimated_eval_tokens": 17926 }, @@ -9174,162 +9137,168 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is noisy and mostly mixes unrelated issues across tokenization, multimodal processors, model config, and hardware-specific inference failures. None of the soft pairs look like safe duplicate merges.", - "confidence": 0.76, - "canonical_issue_reason": "issue:44521 is the most central and referenced item in the set, and it sits in the broad multimodal `apply_chat_template`/processor area that recurs in nearby reports. It is the best representative of the cluster, though the surrounding items are not true duplicates of it.", + "summary": "This cluster is mostly heterogeneous: the soft links are broad topical similarities across different regressions, not true duplicates. I would not merge any of the proposed soft pairs.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 44521 is the strongest representative of the cluster\u2019s most active multimodal/template-related slice, with high discussion and inbound references; if one issue must serve as canonical, it is the best central example.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the strongest single issue to anchor this noisy cluster because it is active, user-facing, and most likely to attract related processor/chat-template regressions. It is still only a loose representative, not a confirmed duplicate hub.", + "best_issue_reason": "44521 is the most broadly representative and active issue in the set, but only as a loose canonical for the multimodal `apply_chat_template` area; the overall cluster is not a duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Different generation bugs: one is `_update_model_kwargs_for_generation` after `sync_gpus`, the other is Qwen3.5 hidden-states output behavior." + "reason": "Different bugs: generation kwargs handling after `sync_gpus` vs `output_hidden_states` behavior in Qwen3.5." }, { "left": "issue:43295", "right": "issue:43493", "accept": false, - "reason": "One is a processor/tokenizer API regression, the other is a SigLIP2 implementation discrepancy; same broad area but not the same defect." + "reason": "Different model paths: processor/tokenizer image passthrough regression vs SigLIP2 implementation discrepancy." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both involve `apply_chat_template`, but one is all-zero assistant masks and the other is a crash on tool-call assistant messages; different failure modes and likely different fixes." + "reason": "Both involve `apply_chat_template`, but one is zeroed assistant masks in multimodal inputs and the other is a crash on tool-call messages; not the same concrete bug." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Serialization of `lm_head.weight` on different devices is not the same bug as LoRA merge collapse with extended vocab and tied weights." + "reason": "Both mention tied weights, but one is serialization of `lm_head.weight` and the other is LoRA merge collapse with extended vocab; different failure modes." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Qwen2.5-VL vision position IDs for video input is unrelated to a generic FP8 inference failure on B200." - }, - { - "left": "issue:43653", - "right": "issue:44112", - "accept": false, - "reason": "Tokenizer special-token registration bug and a GraniteSpeech CI device-override test failure are unrelated." + "reason": "Different issues: Qwen2.5-VL video position IDs vs FP8 inference failure on B200." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Training/eval embedding dtype drift is unrelated to Qwen3 MoE memory usage." + "reason": "One is dtype drift between training/eval for embeddings; the other is MoE training memory inefficiency." }, { "left": "issue:44737", "right": "issue:46894", "accept": false, - "reason": "XLNet CPU `arange` placement bug and CohereASR double-shift training loss bug are different code paths." + "reason": "Unrelated code paths: XLNet positional encoding device placement vs CohereASR double-shift training loss bug." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Both touch tokenization/embeddings, but one is extra spacing after added tokens and the other is output embedding reinit after resizing." + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resize." }, { - "left": "issue:43824", - "right": "issue:43975", + "left": "issue:43653", + "right": "issue:44112", "accept": false, - "reason": "An import error for a model class and a detokenization regression are not the same underlying issue." + "reason": "BigBirdTokenizer special-token registration bug is unrelated to the GraniteSpeech CI device-override test failure." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3Moe FP8 failure and Qwen2.5-VL import failure are unrelated model-loading problems." + "reason": "Different failures: Qwen3Moe FP8 config issue vs missing Qwen2_5_VL import." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Importing Qwen2_5_VL is unrelated to DeepSeek detokenization in v5." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Non-persistent buffer corruption on load and CLIP text-model structure inconsistency are distinct regressions." + "reason": "Both are v5 regressions, but one corrupts non-persistent buffers on load and the other is a CLIP text-model structure inconsistency." }, { "left": "issue:43452", "right": "issue:46512", "accept": false, - "reason": "GGUF loading breakage is unrelated to `from_config(dtype=...)` leaving `model.config.dtype` stale." + "reason": "GGUF loading/from_pretrained breakage is unrelated to `from_config(dtype=...)` leaving `config.dtype` stale." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and batched video processor shape handling are different media-processing bugs." + "reason": "Audio `apply_chat_template` sampling-rate defaulting and video batched-shape output are different processor bugs." }, { - "left": "issue:45200", - "right": "issue:46899", + "left": "issue:43334", + "right": "issue:43525", "accept": false, - "reason": "`mm_token_type_ids` defaulting to zeros is not the same as unsafe dtype casts in quantized multimodal embedders." + "reason": "Similar symptom (`pad_token_id` missing), but different model configs and likely separate fixes; not enough evidence of one shared bug." }, { "left": "issue:45200", - "right": "issue:46932", + "right": "issue:46899", "accept": false, - "reason": "Text-only `mm_token_type_ids` behavior and quantized vision/audio embedder dtype casting are separate issues." + "reason": "Text-only `mm_token_type_ids` defaulting and quantized multimodal embedder dtype casting are different Gemma4 problems." }, { - "left": "issue:43334", - "right": "issue:43525", + "left": "issue:44442", + "right": "issue:45198", "accept": false, - "reason": "Both mention missing `pad_token_id`, but on different config classes/models; this looks model-specific rather than one shared bug." + "reason": "Tokenizer auto-loading failure and Wav2Vec2 save/tokenization failure are different model-specific issues." }, { - "left": "issue:44442", - "right": "issue:45198", + "left": "issue:45200", + "right": "issue:46932", "accept": false, - "reason": "FastSpeech2 tokenizer loading and Wav2Vec2 save/tokenization failure are different tokenizer/model problems." + "reason": "`mm_token_type_ids` defaults vs unsafe dtype casting in quantized vision/audio embedders are distinct bugs." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "A Janus image-generation test failure is unrelated to a stale `num_return_sequences` warning in paged generation." + "reason": "Janus image generation test failure is unrelated to a stale paged-generate warning." }, { - "left": "issue:44479", - "right": "issue:46192", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "A video-input regression and a `return_tensors` override bug are unrelated." + "reason": "Perceiver non-default-resolution classification failure is unrelated to RT-DETR memory not being released." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:44479", + "right": "issue:46192", "accept": false, - "reason": "Perceiver interpolation failure and RT-DETR memory retention are different defects." + "reason": "Video-input regression for Qwen2.5/Qwen3 variants is unrelated to `return_tensors` being ignored when `text_kwargs` is passed." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC latent reconstruction mismatch and GraniteSpeech CI device override failure do not share the same code path." + "reason": "DAC latent conversion mismatch and a GraniteSpeech CI test flake are unrelated." }, { - "left": "issue:43827", - "right": "issue:43901", + "left": "issue:45200", + "right": "issue:45325", "accept": false, - "reason": "Both are docs-related, but they reference different deprecated pipeline behaviors and different documentation fixes." + "reason": "Gemma4 token-type defaults and Qwen2.5-VL rope-index scaling are different model bugs." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and Bamba torch.compile SDPA failure are unrelated model bugs." + "reason": "SwitchTransformers sparse-layer creation and Bamba SDPA torch.compile failure are unrelated." }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Hardware-specific Qwen3 A800 device-mapping assert and FP8 inference failure on B200 are different issues." + "reason": "Different models and failure modes: Qwen3 auto device mapping assert vs FP8 inference on B200." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but one is about pipeline removal and the other about `return_all_scores` wording; not the same issue." } ] }, @@ -9374,6 +9343,7 @@ "issue:43825", "issue:44112", "issue:44206", + "issue:44351", "issue:44448", "issue:44466", "issue:44514", @@ -9389,10 +9359,8 @@ "issue:45412", "issue:45520", "issue:45584", - "issue:45588", "issue:45593", "issue:45759", - "issue:45859", "issue:46170", "issue:46234", "issue:46285", @@ -9404,13 +9372,13 @@ ], "soft_pairs": [ "issue:44561|issue:46234", - "issue:43819|issue:45083", "issue:44610|issue:44857", - "issue:43388|issue:43526", + "issue:43819|issue:45083", "issue:44857|issue:45593", + "issue:43388|issue:43526", "issue:45759|issue:46285", - "issue:44995|issue:46858", "issue:46512|issue:46821", + "issue:44995|issue:46858", "issue:46424|issue:46858", "issue:43606|issue:45083", "issue:44206|issue:45584", @@ -9429,219 +9397,219 @@ "issue:43335|issue:45083", "issue:43335|issue:45593", "issue:43335|issue:44112", - "issue:29127|issue:32090", "issue:44514|issue:44521", + "issue:29127|issue:32090", "issue:45290|issue:46326", - "issue:45588|issue:45859" + "issue:37428|issue:44351" ], - "prepared_review_unit_hash": "63379649a6ba00f4b16e17bd731d933bb43e3b5a2617738727d9358e3680182e", + "prepared_review_unit_hash": "249e0c9235468b103b0893e19b790de568fdf6b549bcdb5be26ee629c43e416b", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 37139, - "estimated_input_tokens": 9285, - "estimated_eval_tokens": 18826 + "serialized_chars": 36514, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated Transformers bug reports that only share broad surface themes (imports, cache, chat templates, vision, CI). I don\u2019t see a true duplicate cluster here, so all soft edges should be rejected.", - "confidence": 0.92, - "canonical_issue_reason": "No single canonical issue stands out: the items span many different subsystems and failure modes, so treating any one as the duplicate anchor would be misleading.", + "summary": "These issues are only loosely similar by embedding; they span many unrelated bugs across import utils, caching, processors, trainers, and model-specific regressions. None of the soft-pair candidates look like the same underlying bug/change, so all proposed duplicate links should be rejected.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43335 is the densest hub in the similarity graph, so it is the least arbitrary anchor for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative because the pairwise similarities are thematic rather than the same underlying bug.", + "best_issue_reason": "issue:43335 is the most connected candidate, but the surrounding items are too heterogeneous for this to represent a true duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Different missing utilities in import_utils; they affect separate symbols and are not the same bug." - }, - { - "left": "issue:43819", - "right": "issue:45083", - "accept": false, - "reason": "One is a DAC latent/STE mismatch, the other is a qwen3_omni_moe feature-length helper issue." + "reason": "Different missing-symbol import regressions in import_utils; not the same code path." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor image-size mismatch vs AMP/CUDA loss crash; different code paths and symptoms." + "reason": "Processor image-size mismatch vs AMP/CUDA loss crash; unrelated failures." }, { - "left": "issue:43388", - "right": "issue:43526", + "left": "issue:43819", + "right": "issue:45083", "accept": false, - "reason": "Both involve labels, but one drops elements in gather_for_metrics while the other truncates labels in BeitImageProcessorFast." + "reason": "DAC forward/from_latents mismatch is unrelated to qwen3_omni_moe feature-length behavior." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "One is a numeric stability crash, the other is missing auxiliary losses when denoising is disabled." + "reason": "One is a float16 CUDA crash, the other is missing auxiliary losses; different bugs." }, { - "left": "issue:45759", - "right": "issue:46285", + "left": "issue:43388", + "right": "issue:43526", "accept": false, - "reason": "Different composite-model/config structure regressions; not the same underlying loading issue." + "reason": "Tuple-label truncation in metrics vs BEiT fast processor label reduction; different logic." }, { - "left": "issue:44995", - "right": "issue:46858", + "left": "issue:45759", + "right": "issue:46285", "accept": false, - "reason": "Stale indexer cache on second forward pass vs device-tensor seq-length breaking static-cache slice arithmetic." + "reason": "Composite Qwen text_config unwrapping is not the same as CLIP text-model structure drift." }, { "left": "issue:46512", "right": "issue:46821", "accept": false, - "reason": "Stale model.config.dtype after from_config(dtype=...) is unrelated to a window_size AttributeError." + "reason": "Stale dtype metadata vs TimesFM window_size AttributeError; different subsystems." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Stale indexer cache on second forward pass vs static-cache seq-length tensor arithmetic." }, { "left": "issue:46424", "right": "issue:46858", "accept": false, - "reason": "max_cache_length handling for quantized cache is a different bug from get_seq_length returning a device tensor." + "reason": "Quantized-cache max length handling and static-cache seq-length device tensor are different bugs." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch for bark-small is unrelated to a qwen3_omni_moe length helper bug." + "reason": "CPU offload device mismatch is unrelated to the qwen3_omni_moe helper function." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Unsupported center arg in feature extraction is a different failure from Whisper empty-transcription generation." + "reason": "Unsupported feature-extractor arg vs empty-transcription generation failure; different paths." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Zero probabilities in SigLIP2 and RT-DETR memory leak are unrelated model-specific issues." + "reason": "Zero probabilities in SigLIP2 and RT-DETR memory release are unrelated." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformersConfig is unrelated to BigBirdTokenizer special-token registration." + "reason": "Sparse-layer creation bug vs tokenizer special-token registration bug; not the same issue." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Missing _flash_supports_window_size import and Torch 2.10 incompatibility are different breakages." + "reason": "Missing flash-attention utility import is not the same as Torch 2.10 incompatibility." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures are unrelated to the NoneType passed to _gpu_broadcast_one." + "reason": "Flash Attention 2 test failures and NoneType GPU broadcast are different failures." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "PyTorch-missing ImportError and a MacOS bus error in community CLIP are separate problems." + "reason": "Missing PyTorch import and Mac bus error in CLIP model are unrelated." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is not the same as a stale GraniteSpeech CI test." + "reason": "torch.compile/SDPA model failure vs stale CI device-override test; different bugs." }, { "left": "issue:43611", "right": "issue:44466", "accept": false, - "reason": "base_model_prefix loading breakage and lm_head.weight serialization inconsistency are different v5 regressions." + "reason": "base_model_prefix loading regression does not match lm_head.weight serialization inconsistency." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Creating a sparse layer at zero sparse layers is unrelated to BEiT pretrained-preset legacy-field loading." + "reason": "SwitchTransformers sparse-layer config bug and BEiT legacy-field load failure are unrelated." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output differences are a different issue from recurrent states resetting in modular_qwen3_5." + "reason": "Pegasus output regression vs Qwen recurrent-state reset bug; different models and code paths." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "flash_attn lookup on Python 3.13 and Windows ROCm torch.distributed dependency failures are unrelated." + "reason": "Python 3.13 flash_attn import error and ROCm/Windows distributed dependency issue are distinct." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 JIT parsing/IndentationError is not the same as the qwen3_omni_moe helper behavior." + "reason": "Python 3.13 JIT parse/import problem is unrelated to qwen3_omni_moe length helper behavior." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "dtype mismatch between train/eval and a pipeline translation-support message are unrelated." + "reason": "dtype drift between train/eval is unrelated to chat-template translation messaging." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe output-length helper are different bugs." + "reason": "Sparse-layer config creation and helper output-length behavior are different problems." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "A misbuilt sparse layer is unrelated to D-FINE not using auxiliary losses when denoising is off." + "reason": "SwitchTransformers sparse-layer bug is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "SwitchTransformers config creation is not the same as a GraniteSpeech CI device-override failure." + "reason": "Different models and failure modes; not a shared bug." }, { - "left": "issue:29127", - "right": "issue:32090", + "left": "issue:44514", + "right": "issue:44521", "accept": false, - "reason": "LayoutLMv3 box-validation messaging and Trainer _gpu_broadcast_one NoneType errors are unrelated." + "reason": "Batch padding crash and all-zero assistant masks are related area but not the same concrete bug." }, { - "left": "issue:44514", - "right": "issue:44521", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "Both touch apply_chat_template, but one crashes on batched padding=False input while the other yields zero assistant masks for multimodal input." + "reason": "LayoutLMv3 error-message clarity issue is unrelated to Trainer `_gpu_broadcast_one` TypeError." }, { "left": "issue:45290", "right": "issue:46326", "accept": false, - "reason": "Tool-call/no-content crash and mistral_common continue_final_message behavior are different chat-template edge cases." + "reason": "Tool-call/no-content chat-template crash vs mistral_common continue_final_message bug; different triggers." }, { - "left": "issue:45588", - "right": "issue:45859", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "flash_attention.py s_aux=None crash and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are separate problems." + "reason": "Missing internal flash-attention import and missing HybridCache export are separate import regressions." } ] }, @@ -9675,7 +9643,6 @@ "issue:33357", "issue:34689", "issue:35141", - "issue:37428", "issue:38175", "issue:39692", "issue:42490", @@ -9691,7 +9658,6 @@ "issue:43854", "issue:44112", "issue:44206", - "issue:44351", "issue:44360", "issue:44448", "issue:44493", @@ -9706,7 +9672,9 @@ "issue:45310", "issue:45459", "issue:45478", + "issue:45588", "issue:45823", + "issue:45859", "issue:45920", "issue:46192", "issue:46234", @@ -9715,9 +9683,9 @@ "issue:46697" ], "soft_pairs": [ - "issue:37428|issue:44351", - "issue:43720|issue:44112", + "issue:45588|issue:45859", "issue:42490|issue:43854", + "issue:43720|issue:44112", "issue:34689|issue:35141", "issue:43819|issue:44493", "issue:39692|issue:43526", @@ -9733,8 +9701,8 @@ "issue:44617|issue:46697", "issue:43701|issue:45083", "issue:44206|issue:45823", - "issue:32090|issue:33357", "issue:44857|issue:45072", + "issue:32090|issue:33357", "issue:42490|issue:43575", "issue:43550|issue:45083", "issue:29942|issue:35141", @@ -9745,214 +9713,214 @@ "issue:29870|issue:32090", "issue:45920|issue:46192" ], - "prepared_review_unit_hash": "432ea69242708d598fe532e6d0b3e1f320ee0a6ef6ba97f93f00bf26e7c52a0d", + "prepared_review_unit_hash": "584599ab22f9b7e994de378dc4e8dc5f2fd2e4fa3231de2038ea2b6d31082e43", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36459, - "estimated_input_tokens": 9115, - "estimated_eval_tokens": 18486 + "serialized_chars": 36471, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained regressions; the rest are similar only at a broad subsystem/topic level and should not be merged.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45310 is the earlier, clearer report for the one confirmed duplicate pair and best captures the concrete Qwen3.5 MoE from_pretrained regression.", + "summary": "Almost all soft pairs are false positives: they point to different bugs, models, or subsystems. The only clear duplicate is the repeated Qwen3.5 MoE from_pretrained regression in issue:45310 and issue:45478.", + "confidence": 0.97, + "canonical_issue_reason": "issue:45310 is the better canonical issue: it is the earlier report, has a slightly fuller description, and matches issue:45478 exactly in title and bug report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the best representative issue for the cluster because it is the canonical target of the only true duplicate and is the most specific concrete bug report among the matched pair.", + "best_issue_reason": "issue:45310 is the best issue to represent the duplicate pair because it is the original report for the same Qwen3.5 MoE from_pretrained failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:45588", + "right": "issue:45859", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and different code paths (_flash_supports_window_size vs HybridCache)." + "reason": "Both involve attention-related model code, but one is a flash-attention sinkless-model crash and the other is a CPU offload rotary-embedding incompatibility; different failure paths." }, { - "left": "issue:43720", - "right": "issue:44112", + "left": "issue:42490", + "right": "issue:43854", "accept": false, - "reason": "Both are CI/runtime failures, but the root causes are unrelated: BitNet packed-weight loading vs GraniteSpeech stale device override." + "reason": "Local-path trailing slash warning is unrelated to GLM-4.7-Flash loading correctness in tests; different surfaces and symptoms." }, { - "left": "issue:42490", - "right": "issue:43854", + "left": "issue:43720", + "right": "issue:44112", "accept": false, - "reason": "A trailing-slash local-path warning is unrelated to a model-loading failure for zai-org/GLM-4.7-Flash." + "reason": "BitNet packed-weight loading failure is unrelated to a GraniteSpeech CI device-override test regression." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Both involve model loading/init, but one is a specific Llama 3.2 vision loading breakage and the other is embedding reinitialization when resizing token embeddings." + "reason": "Llama 3.2 Vision model loading breakage is unrelated to embedding reinitialization after resize_token_embeddings with tie_word_embedding=False." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "Different bugs: DAC.from_latents/forward mismatch vs unexpected position-id keys in many models." + "reason": "DAC latent/forward mismatch is a model math bug; unexpected position-id keys is a separate loading/input-key regression." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Docs example failures for SigLIP2 are unrelated to BEiT Fast processor label reduction behavior." + "reason": "SigLIP2 docs example errors and BeitImageProcessorFast reduce_labels behavior are different components and different bug classes." }, { "left": "issue:46431", "right": "issue:46566", "accept": false, - "reason": "Chat-template absence in a processor and DiffusionGemma staying on meta with device_map=auto are different issues." + "reason": "Missing chat template error is unrelated to DiffusionGemma staying on meta with device_map=\"auto\"." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch for bark-small is not the same as a float16 AMP CUDA crash in LwDetrImageLoss." + "reason": "CPU offload device mismatch in bark-small is unrelated to LwDetrImageLoss crashing under float16 AMP on CUDA." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Both are v5 regressions, but they affect different model/tokenizer code paths and are not the same underlying bug." + "reason": "Pegasus output drift across v4/v5 is a model behavior regression, while DeepSeek tokenizer incorrect IDs is a tokenizer regression." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "These are effectively the same bug report: identical Qwen3.5 MoE from_pretrained failure on transformers>=5.4.0." + "reason": "Same title, same Qwen3.5 MoE from_pretrained error, and both describe the same underlying regression." }, { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "The checkpoint-resume epoch_iterator error and a training-start failure for FineWeb are too different to be duplicates." + "reason": "resume_from_checkpoint epoch_iterator failure and FineWeb SLM training not starting are different training issues with no concrete shared code path shown." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing after an added token is unrelated to a Trainer _gpu_broadcast_one NoneType type error." + "reason": "Tokenizer added-space bug and trainer _gpu_broadcast_one NoneType error are unrelated." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA/extended-vocab collapse with tied embeddings is unrelated to the 2^24 category limit error." + "reason": "LoRA merge with extended vocabulary causing collapse is unrelated to the category-count limit error." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer protobuf error masking and a missing import-utils availability helper are unrelated import/exception bugs." + "reason": "Tokenizer protobuf error handling and missing is_falcon_mamba_ssm_available helper are unrelated import/runtime issues." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 doc example failures are unrelated to the position-id key regression affecting many models." + "reason": "Documentation/example issues do not match a position-id-key regression in model loading." }, { "left": "issue:44617", "right": "issue:46697", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Swin2SR image over-padding." + "reason": "Sam3Video OOM and Swin2SR over-padding are unrelated model/image-processing bugs." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "A resume_from_checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + "reason": "resume_from_checkpoint key mismatch is a training/checkpoint issue, not the qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "LasrFeatureExtractor's unsupported center arg and Gemma4 PLE device mismatch are different bugs." + "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to Gemma4 PLE device-map forward mismatch." }, { - "left": "issue:32090", - "right": "issue:33357", + "left": "issue:44857", + "right": "issue:45072", "accept": false, - "reason": "A Trainer broadcast type error is unrelated to the MacOS bus error on a CLIP community model." + "reason": "Different failures: AMP CUDA crash in LwDetrImageLoss versus bfloat16 inference dtype mismatch for SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:44857", - "right": "issue:45072", + "left": "issue:32090", + "right": "issue:33357", "accept": false, - "reason": "Both mention dtype issues, but they affect different models/code paths and are not the same concrete failure." + "reason": "NoneType broadcast error in Trainer is unrelated to a MacOS bus error loading a CLIP community model." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "A trailing-slash import warning is unrelated to an OOM when loading Qwen2-57B-A14B-Instruct with tensor parallelism." + "reason": "Trailing-slash import warning is unrelated to tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to the qwen3_omni_moe feature-length helper issue." + "reason": "torch.compile+SDPA failure in Bamba-9B-v2 is unrelated to the qwen3_omni_moe helper function behavior." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures are not the same bug as output-embedding reinitialization after resizing token embeddings." + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization after resizing token embeddings." }, { "left": "issue:27957", "right": "issue:34689", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 vision model-loading regressions are unrelated." + "reason": "XLMRoberta with Flash Attention 2 and Llama 3.2 Vision model loading are distinct model-specific problems." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 is not the same issue as unexpected position-id keys across models." + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to unexpected position-id keys." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "A DSA indexer ReLU comment and removal of is_torch_fx_available breaking trust_remote_code models are unrelated." + "reason": "A missing ReLU discussion in the DSA indexer is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "BF16-to-FP32 embedding dtype drift and an undefined video-token helper are different code-path bugs." + "reason": "Embedding dtype drift in evaluate is unrelated to undefined video-branch helpers in multimodal token counting." }, { "left": "issue:29870", "right": "issue:32090", "accept": false, - "reason": "Bart generate() decoder_attention_mask initialization and the _gpu_broadcast_one NoneType failure are unrelated." + "reason": "BART decoder_attention_mask assignment bug is unrelated to the _gpu_broadcast_one NoneType type error." }, { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "A tokenizer-v5 regression producing wrong token IDs is unrelated to return_tensors being ignored when text_kwargs is passed explicitly." + "reason": "Wrong AutoTokenizer token IDs across multiple models is unrelated to return_tensors being ignored when text_kwargs is passed." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative: the only accepted soft pair is an exact-title duplicate with strong textual overlap (issue:45310 / issue:45478), and the rejected pairs are justified by clearly different bugs or code paths. No overstatement detected." + "feedback": "The analysis is grounded in the packet and stays conservative overall. The only clearly supported duplicate is issue:45310 / issue:45478 with identical titles and closely matching descriptions. The soft-pair rejections are mostly justified by distinct titles and bug classes, with no obvious overstatement." }, "evaluator_used": true, "retried": false, @@ -10026,11 +9994,11 @@ "issue:44448|issue:44977", "issue:45290|issue:45381", "issue:10000|issue:15354", - "issue:44977|issue:45229", "issue:44855|issue:46399", + "issue:44977|issue:45229", "issue:43278|issue:43381", - "issue:40444|issue:43526", "issue:29870|issue:33290", + "issue:40444|issue:43526", "issue:43927|issue:45698", "issue:45290|issue:46655", "issue:44857|issue:45588", @@ -10040,241 +10008,235 @@ "issue:44561|issue:44779", "issue:43873|issue:45229", "issue:42490|issue:43873", - "issue:25251|issue:32090", "issue:44448|issue:44466", + "issue:25251|issue:32090", "issue:43976|issue:45440", "issue:44568|issue:45491", "issue:29127|issue:34567", "issue:45588|issue:45823", "issue:44871|issue:44977", - "issue:33357|issue:39692", "issue:43381|issue:43701", - "issue:41720|issue:45229", + "issue:33357|issue:39692", "issue:43278|issue:43575", + "issue:41720|issue:45229", "issue:44977|issue:45563", "issue:45198|issue:45803", "issue:44186|issue:45198" ], - "prepared_review_unit_hash": "444f5f9289ead52a9d682695f84c93c11912dc901706a56d9c71bcb0b9dce25a", + "prepared_review_unit_hash": "0cdabcb724e49e6308e80727e73d8524298a1c7e651fced0acddb6395a686d88", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37330, - "estimated_input_tokens": 9333, - "estimated_eval_tokens": 18922 + "serialized_chars": 37334, + "estimated_input_tokens": 9334, + "estimated_eval_tokens": 18924 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is thematically related to Transformers bugs, but the soft-linked pairs are not true duplicates: they involve different models, APIs, and failure modes. I would not merge any of the suggested edges.", - "confidence": 0.93, - "canonical_issue_reason": "No pair looks like the same underlying bug. If one issue has to serve as the representative, issue 43873 is the best umbrella candidate because it is open, high-activity, and broadly about a core runtime path (offloading/quantization), but it does not subsume the others.", + "summary": "This cluster is a grab-bag of unrelated Transformers bugs across generation, tokenizers, multimodal processors, Trainer, cache/offloading, and serialization. I don\u2019t see any soft pair that is clearly the same underlying defect; the strongest representative issue is the active custom-module save/load regression.", + "confidence": 0.62, + "canonical_issue_reason": "issue:45698 is the most representative and actionable bug here: it is open, has multiple inbound references, and describes a concrete save_pretrained/from_pretrained regression with broad impact.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the strongest single representative for the cluster as a whole due to its broader scope and higher discussion/inbound reference activity; however, it is still not a duplicate anchor for the other issues.", + "best_issue_reason": "issue:45698 has the clearest reproducible failure, active discussion, and broader downstream impact than the other candidates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Different models and bugs: Pegasus v4/v5 output divergence vs Qwen3.5 flash-attention generation failure." + "reason": "Both are v5 output regressions, but they hit different models and different code paths (Pegasus generation vs Qwen3.5 flash-attention)." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Different code paths: chat template/tool-call crash vs Qwen2.5-VL video vision_position_ids issue." + "reason": "Different multimodal failures: chat template tool-call handling vs Qwen2.5-VL video vision_position_ids." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Unrelated failures: DistilBertModel German issue vs GeneratorExp/torch.jit.script export problem." + "reason": "Different problem classes: a DistilBert model error vs torch.jit.script GeneratorExp export failure." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:44855", + "right": "issue:46399", "accept": false, - "reason": "Different root causes: flash-attention generation bug vs multi-GPU CUDA OOM on Gemma4 31B." + "reason": "Both involve model import/runtime errors, but the root causes are unrelated: Python 3.13 TorchScript parsing vs a missing decorator in Qwen3.5 kernel registration." }, { - "left": "issue:44855", - "right": "issue:46399", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "Different model/runtime problems: Python 3.13 import/JIT parsing error vs missing decorator registration in Qwen3.5 kernels." + "reason": "One is a generation correctness issue, the other is a multi-GPU CUDA OOM; same model family is not enough to make them duplicates." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Both involve eval mode, but one is dtype drift and the other is gradient checkpointing being disallowed; not the same bug." + "reason": "Both mention eval-mode behavior, but one is dtype drift and the other is gradient checkpointing being unsupported in eval." }, { - "left": "issue:40444", - "right": "issue:43526", + "left": "issue:29870", + "right": "issue:33290", "accept": false, - "reason": "Different subsystems: multimodal finetuning with multiple images vs BeitImageProcessorFast label reduction bug." + "reason": "UnboundLocalError during BART generation and DeepSpeed Adafactor OOM are unrelated failure modes." }, { - "left": "issue:29870", - "right": "issue:33290", + "left": "issue:40444", + "right": "issue:43526", "accept": false, - "reason": "Different failures: BART generate UnboundLocalError vs Adafactor/deepspeed OOM." + "reason": "Different subsystems and symptoms: IterableDataset with multi-image prompts vs BeitImageProcessorFast label reduction bug." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both touch save/load, but one loses custom token IDs in config while the other loads the wrong custom module; not the same underlying bug." + "reason": "Both involve save/load behavior, but one corrupts config token IDs while the other loads the wrong custom module after save_pretrained." }, { "left": "issue:45290", "right": "issue:46655", "accept": false, - "reason": "Different components: chat template tool-call crash vs SmolVLM video padding calculation bug." + "reason": "Both are preprocessing crashes, but the underlying bugs are different: chat-template tool calls vs video padding calculation." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different crash sites and causes: float16 AMP loss issue vs flash_attention.py handling of s_aux=None." + "reason": "Different crash sites and causes: AMP/CUDA loss computation vs flash_attention handling of s_aux=None." }, { "left": "issue:37428", "right": "issue:42175", "accept": false, - "reason": "Completely different issues: missing flash_attention import symbol vs TensorFlow not included in torch-only install." + "reason": "Import failure for a missing flash-attention helper is unrelated to TensorFlow not being installed in a torch-only backend setup." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Different trainer/deepspeed failures: NoneType broadcast error vs optimizer memory OOM." + "reason": "Trainer distributed broadcast of NoneType and Adafactor OOM in DeepSpeed are separate issues." }, { "left": "issue:46421", "right": "issue:46858", "accept": false, - "reason": "Both relate to cache/prefill, but one is spurious recompilation from lazy init and the other is a device-tensor arithmetic bug." + "reason": "Both are cache-related, but one is recompilation from lazy init and the other is device-tensor return value breaking slice arithmetic." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Different regressions: removal of is_torch_fx_available affecting trust_remote_code vs Deepseek tokenizer output changes in v5." + "reason": "A removed utility breaking trust_remote_code models is not the same as DeepSeek tokenizer behavior changes in v5." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Both are memory-related, but one is offloading/quantization behavior and the other is multi-GPU CUDA OOM; not the same bug." + "reason": "Quantization/offloading not working and Gemma4 multi-GPU OOM are different resource-management bugs." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Different areas: local-path import warning vs offloading with quantization." + "reason": "A path-trailing-slash import warning is unrelated to quantization offloading behavior." }, { - "left": "issue:25251", - "right": "issue:32090", + "left": "issue:44448", + "right": "issue:44466", "accept": false, - "reason": "Different subsystems and symptoms: pipeline top_k nesting vs Trainer distributed broadcast type error." + "reason": "Both are v5 regressions, but one is output mismatch and the other is serialization of tied weights; different code paths." }, { - "left": "issue:44448", - "right": "issue:44466", + "left": "issue:25251", + "right": "issue:32090", "accept": false, - "reason": "Both are v5 regressions, but one is model output divergence and the other is serialization of tied weights; separate bugs." + "reason": "Pipeline top_k nesting behavior and Trainer broadcast errors are unrelated." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Different issues: Python version compatibility vs native DeepSeekV3MoE divergence from remote code." + "reason": "Python version compatibility for Transformers 5.1.0 is not the same as DeepSeekV3 native-vs-remote implementation divergence." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token insertion bug vs Gemma3 NaN embeddings from sliding-window attention." + "reason": "Tokenizer BOS/EOS insertion bug and Gemma3 NaN embeddings are distinct tokenizer/model execution problems." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "Different concerns: error-message clarity in LayoutLMv3 vs TrainerState token-count tracking." + "reason": "LayoutLMv3 error messaging and TrainerState token counter not updating are different concerns." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Different failure modes in different model paths: flash-attention auxiliary-state crash vs Gemma4 PLE device mismatch." + "reason": "Both touch Gemma4/flash attention/device placement, but one is a flash-attention AttributeError and the other is a forward device mismatch." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Different model/config bug vs flash-attention generation bug; no shared root cause." - }, - { - "left": "issue:33357", - "right": "issue:39692", - "accept": false, - "reason": "One is a MacOS bus error in CLIP loading; the other is a SigLIP2 documentation example problem with quantization/model mismatch." + "reason": "Gemma eos_token_id config mismatch and Qwen3.5 flash-attention generation issues do not share the same root cause." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Different trainer lifecycle issues: eval-mode checkpointing restriction vs resume-from-checkpoint key mismatch." + "reason": "Eval-mode gradient checkpointing limitation and resume_from_checkpoint key mismatch are unrelated Trainer issues." }, { - "left": "issue:41720", - "right": "issue:45229", + "left": "issue:33357", + "right": "issue:39692", "accept": false, - "reason": "Different device-placement failures: cudaErrorAssert on auto mapping vs CUDA OOM in multi-GPU inference." + "reason": "MacOS bus error with a community CLIP model and SigLIP2 documentation example errors are not the same defect." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Different problems: eval dtype change vs tensor-parallel loading OOM." + "reason": "BF16-to-FP32 dtype drift in eval is unrelated to Qwen2-57B tensor-parallel OOM." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Auto device mapping cudaErrorAssert and Gemma4 multi-GPU OOM are different runtime failures." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Different generation bugs: flash-attention failure vs stale warning about num_return_sequences." + "reason": "Flash-attention generation failure and stale paged-generate warning are different behaviors." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Different components and defects: Wav2Vec2 save/tokenization failure vs bare-except swallowing errors in FuyuBatchFeature." + "reason": "Wav2Vec2 save/tokenization failure and Fuyu bare-except swallowing errors are unrelated." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "Different model/tokenizer bugs: LayoutLMv2 tokenizer crashes vs Wav2Vec2 save_pretrained/tokenization failure." - }, - { - "left": "issue:43381", - "right": "issue:43701", - "accept": false, - "reason": "Same as above: unrelated eval checkpointing and checkpoint-resume key mismatch bugs." + "reason": "LayoutLMv2 tokenizer crashes and Wav2Vec2 save/tokenization failure are distinct tokenizer bugs." } ] }, @@ -10355,16 +10317,16 @@ "issue:45372|issue:45588", "issue:44206|issue:44479", "issue:43756|issue:45440", - "issue:42915|issue:42994", "issue:29942|issue:30064", + "issue:42915|issue:42994", "issue:43653|issue:44442", - "issue:45491|issue:46693", "issue:45459|issue:45712", - "issue:43412|issue:45588", + "issue:45491|issue:46693", "issue:44265|issue:45198", + "issue:43412|issue:45588", "issue:43065|issue:43994", - "issue:45372|issue:45406", "issue:29127|issue:33357", + "issue:45372|issue:45406", "issue:43065|issue:45412", "issue:45245|issue:45563", "issue:43575|issue:43701", @@ -10381,226 +10343,226 @@ "issue:44442|issue:45072", "issue:44464|issue:45584" ], - "prepared_review_unit_hash": "6cafbdd3a5e5c8736182cd907b2164b6d6dab46199efcec435bca1213b2befea", + "prepared_review_unit_hash": "146ba47a4fd9ad9d9e18451336c012dbdbc5831fc082ade3f935c4dbc7c81b1d", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37371, - "estimated_input_tokens": 9343, - "estimated_eval_tokens": 18942 + "serialized_chars": 37374, + "estimated_input_tokens": 9344, + "estimated_eval_tokens": 18944 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is not a duplicate cluster: the issues span many unrelated Transformers areas (tokenizers, generation, vision, export, quantization, memory, and model-loading regressions). All soft edges should be rejected.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the items do not share one underlying bug, so there is no single representative duplicate to choose.", + "summary": "These are mostly unrelated Transformers bug reports spanning tokenizers, model loading, generation, quantization, export, and memory issues. None of the soft pairs looks like the same concrete bug, so this cluster should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue cleanly represents the cluster because the items are heterogeneous. If forced, issue 44462 is the broadest user-facing loading bug, but it is not a true umbrella duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue: the cluster is heterogeneous, so any single issue would be arbitrary rather than a true canonical duplicate.", + "best_issue_reason": "Issue 44462 is the most generally phrased and broadly reusable bug report in the set, but the cluster is too diverse for it to be a strong canonical duplicate anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer-related failures, but they affect different models and different code paths (MLuke task usage vs Wav2Vec2 save/tokenization)." + "reason": "Different tokenizer/model families and different failures: MLukeTokenizer AttributeError vs Wav2Vec2 save/tokenization failure." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "Unrelated bugs: loading with base_model_prefix vs rope-validation TypeError on JSON-loaded lists." + "reason": "Model loading with base_model_prefix is unrelated to a TypeError in rope validation list handling." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "Tokenizer special-token registration bug vs recurrent-state reset in Qwen3.5 MoE; different subsystems and symptoms." + "reason": "BigBirdTokenizer special-token registration bug is unrelated to Qwen recurrent-state reset logic." }, { "left": "issue:29870", "right": "issue:34689", "accept": false, - "reason": "generate()-time UnboundLocalError in BART vs a model-loading regression for Llama 3.2 Vision; not the same bug." + "reason": "BART generate() decoder mask assignment error is unrelated to Llama 3.2 vision model loading regression." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 processor import failure vs flash-attention integration crash on s_aux=None; different failure points." + "reason": "Gemma 4 processor import/dependency failure is a different bug from flash_attention.py crashing on s_aux=None." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Feature extractor argument regression vs video-input regression in Qwen VL models; different concrete issues." + "reason": "Unsupported center argument in a feature extractor is unrelated to a Qwen video-input regression." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Both concern model-implementation fidelity, but they are different models and different discrepancies." - }, - { - "left": "issue:42915", - "right": "issue:42994", - "accept": false, - "reason": "Quantization-related, but one is FP8 config/model failure and the other is quantized model saving; not the same bug." + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are different model implementation bugs." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation map processing are unrelated." + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems and symptoms." }, { - "left": "issue:43653", - "right": "issue:44442", + "left": "issue:42915", + "right": "issue:42994", "accept": false, - "reason": "Both touch tokenizers, but one is a missing special token registration and the other is AutoTokenizer loading a specific tokenizer class." + "reason": "Qwen3Moe FP8 loading failure and quantized model saving failure are both quantization-related but not the same bug." }, { - "left": "issue:45491", - "right": "issue:46693", + "left": "issue:43653", + "right": "issue:44442", "accept": false, - "reason": "Both mention long-context/FA2, but one is NaN embeddings in Gemma3 and the other is a performance regression; different bugs." + "reason": "Both are tokenizer-related, but BigBird mask-token registration and FastSpeech2ConformerTokenizer loading are distinct issues." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Tokenizer error masking vs leftover dummy classes; no shared underlying issue." + "reason": "Hiding tokenizer decode errors when protobuf is missing is unrelated to leftover dummy classes in dummy_pt_objects.py." }, { - "left": "issue:43412", - "right": "issue:45588", + "left": "issue:45491", + "right": "issue:46693", "accept": false, - "reason": "ExecuTorch export failure for Gemma3n vs flash-attention crash in sink-less models; unrelated code paths." + "reason": "Gemma3 NaNs in mixed-length batching are a correctness bug, while the FA2 issue is a performance regression." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export failure with torch_compilable_check vs Wav2Vec2 save/tokenization failure; different problems." + "reason": "torch.export failures from torch_compilable_check are unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { - "left": "issue:43065", - "right": "issue:43994", + "left": "issue:43412", + "right": "issue:45588", "accept": false, - "reason": "Different model-specific issues: dummy Conv2d in Sam3PixelDecoder vs bad AutoModel/pipeline outputs for SigLIP2." + "reason": "Executorch export guard/erfinv support and flash_attention sinkless crashes are different code paths." }, { - "left": "issue:45372", - "right": "issue:45406", + "left": "issue:43065", + "right": "issue:43994", "accept": false, - "reason": "Both involve Gemma 4, but one is a dependency import error during loading and the other is a missing _tokenizer attribute in serve." + "reason": "A dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to SigLIP2 producing nonsense outputs through AutoModel/pipeline." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error messaging vs CLIP bus error on macOS; unrelated." + "reason": "LayoutLMV3 error-message clarity and a MacOS bus error in CLIP loading are not the same underlying bug." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma 4 processor import failure and serve crashing due to missing _tokenizer are different defects in different paths." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Sam3PixelDecoder dummy layer issue vs RT-DETR memory leak; different models and different symptoms." + "reason": "Dummy Conv2d placeholders and RT-DETR memory leaks are unrelated issues." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count runtime limit vs stale generate() warning; unrelated." + "reason": "Category-count runtime limit and stale warning behavior in paged generate() are unrelated generation bugs." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM on Qwen2-57B vs resume_from_checkpoint key mismatch; not the same bug." + "reason": "Tensor-parallel OOM on load and resume_from_checkpoint key mismatch are distinct failures." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "DeepSeekV3MoE implementation divergence vs D-FINE auxiliary-loss behavior; unrelated." + "reason": "DeepseekV3MoE divergence from remote implementation and D-FINE auxiliary-loss omission are unrelated model bugs." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Local-path import warning vs gradient-checkpointing eval-mode restriction; unrelated." + "reason": "Import warning for local paths ending in slash is unrelated to gradient checkpointing being disallowed in eval mode." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 box-clarity issue vs embedding reinitialization in post_init; different code paths." + "reason": "LayoutLMV3 box-information messaging and post_init reinitializing output embeddings are different bugs." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma 4 dependency import failure vs protobuf-related tokenizer error masking; different root causes." + "reason": "Gemma 4 loading import failure and protobuf error masking in tokenizer code are unrelated." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime special-token post-processor update bug vs Gemma3 mixed-length NaNs; unrelated." + "reason": "Runtime special-token/post-processor update behavior is unrelated to Gemma3 NaNs with mixed-length batching." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 inference mismatch vs RT-DETR memory leak; not the same underlying bug." + "reason": "SigLIP2 pipeline/AutoModel output issues are unrelated to RT-DETR memory retention after deletion." }, { "left": "issue:45406", "right": "issue:46701", "accept": false, - "reason": "Both are Gemma-related, but serve processor state and image-text-to-text pipeline failures are different issues." + "reason": "Gemma4Processor serve crash and DiffusionGemmaGenerationOutput pipeline failure are different issues." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Tokenizer decoder-type mismatch in v5 vs AutoTokenizer ignoring tokenizer.json; related area, but not the same defect." + "reason": "Wrong tokenizer decoder type in v5 and AutoTokenizer ignoring tokenizer.json are related broadly, but not the same concrete bug." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "Both are memory complaints, but one is a leak on deletion and the other is an overflow during sam3-litetext execution." + "reason": "RT-DETR garbage-collection memory leak and sam3-litetext memory overflow are different memory problems." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "OOM in tensor-parallel model loading vs higher reserved memory under int4 quantization; different causes and fixes." + "reason": "Tensor-parallel OOM on model load and int4 quantization reserved-memory regression are distinct bugs." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "AutoTokenizer loading failure vs dtype mismatch in bfloat16 inference; unrelated." + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to dtype mismatches in bfloat16 inference." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency vs Whisper empty-transcription failure after token alignment; different bugs." + "reason": "Chunked generation inconsistency with compiled forward and Whisper empty-transcription failure after align_special_tokens are different generation bugs." } ] }, @@ -10684,8 +10646,8 @@ "issue:43526|issue:44186", "issue:33290|issue:34689", "issue:33290|issue:35141", - "issue:46121|issue:46399", "issue:43334|issue:43531", + "issue:46121|issue:46399", "issue:45250|issue:46640", "issue:43381|issue:43873", "issue:45588|issue:46821", @@ -10700,23 +10662,23 @@ "issue:45003|issue:45440", "issue:43720|issue:44442", "issue:42175|issue:46234", - "issue:43493|issue:43827", - "issue:27957|issue:29870" + "issue:27957|issue:29870", + "issue:43493|issue:43827" ], - "prepared_review_unit_hash": "496d3b753e483b03458c5bdd6c8464532fa042a83dbee2e6919baf5c737fdc1f", + "prepared_review_unit_hash": "03780444fb37c5e268a14e6bf721581b3c14a59d3058461818dd89cd6c6a0458", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35969, + "serialized_chars": 35971, "estimated_input_tokens": 8993, "estimated_eval_tokens": 18242 }, @@ -10725,180 +10687,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is not a true duplicate cluster: the items span many unrelated regressions/features across vision, audio, quantization, loading, and docs. None of the soft pairs shows the same concrete bug or code path, so all should be rejected.", - "confidence": 0.97, - "canonical_issue_reason": "No single issue is a good canonical for this set; the cluster is heterogeneous and the nearest-looking pairs still describe different bugs.", + "summary": "No strong duplicate clusters emerge from the provided issues; most soft pairs are thematically related at a high level only and do not describe the same concrete bug or change.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46640 is the broadest and most current Flash Attention 2 report among the set, so it works best as a representative umbrella issue.", "canonical_pr_reason": null, - "best_issue_reason": "No strong best issue candidate: a few items are thematically close (e.g. Flash Attention 2 or Qwen2.5-VL), but they do not collapse to one underlying fix.", + "best_issue_reason": "issue:46640 has the clearest generic FA2 framing, recent activity, and inbound references, making it the best single issue to anchor this group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "Different failures: BeitImageProcessorFast label reduction vs Kosmos2.5 OCR long-input indexing." + "reason": "Different components and failures: BeitImageProcessorFast label reduction vs Kosmos2.5 OCR long-input indexing." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Different Qwen problems: FineGrainedFP8Config failure vs Qwen3-VL weight-shape mismatch on load." + "reason": "Different models and bugs: Qwen3MoE FP8 config failure vs Qwen3-VL weight-shape loading mismatch." }, { "left": "issue:43526", "right": "issue:46693", "accept": false, - "reason": "Unrelated: image processor label handling vs FA2 long-context performance regression." + "reason": "Unrelated issues: image processor label handling vs FA2 long-context performance regression." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Missing import/export for Qwen2.5-VL class is not the same as a vague dependency issue." + "reason": "ImportError for a missing Qwen2.5-VL class is not the same as a generic dependency issue." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "Different subsystems and bugs: Flash Attention 2 support vs LayoutLMv3 box-validation messaging." + "reason": "Different subsystems: XLM-RoBERTa Flash Attention 2 vs LayoutLMv3 error-message clarity." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "FA2 model support and Adafactor OOM in DeepSpeed are unrelated." + "reason": "Flash Attention 2 model support is unrelated to Adafactor OOM in DeepSpeed." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "SigLIP2 incorrect outputs and SAM3-LiteText memory overflow are different issues." + "reason": "Different model families and symptoms: SigLIP2 bad outputs vs SAM3 LiteText memory overflow." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Image-processor label reduction bug is unrelated to LayoutLMv2 tokenizer padding/truncation crashes." + "reason": "Different tokenizer/processor bugs in different models; not the same code-path." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "Adafactor/DeepSpeed OOM is not the same as Llama 3.2 Vision loading breakage." + "reason": "Optimizer memory issue is unrelated to Llama 3.2 Vision loading breakage." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Optimizer memory blowup and embedding reinitialization are different code-path bugs." + "reason": "OOM in DeepSpeed Adafactor is unrelated to embedding reinitialization after resize." }, { - "left": "issue:46121", - "right": "issue:46399", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Rope-validation TypeError and missing decorator during registration are separate failures." + "reason": "Both are Qwen-related, but one is a missing config attribute and the other is a sliding-window behavior issue." }, { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:46121", + "right": "issue:46399", "accept": false, - "reason": "Qwen3-VL pad_token_id load error is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Both mention rope/rotary code, but they are different failures in different paths and models." }, { "left": "issue:45250", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but the titles are too generic and no shared concrete failure path is evident." + "reason": "Too generic to treat as duplicates; both mention FA2, but the concrete bug/change is not the same from the titles alone." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode and quantization offloading are different bugs." + "reason": "Gradient checkpointing in eval mode and quantization/offloading are separate behaviors." }, { "left": "issue:45588", "right": "issue:46821", "accept": false, - "reason": "FA integration crash on sink-less models is unrelated to TimesFM window_size AttributeError." + "reason": "Different AttributeError sites and different models; no shared underlying bug is evident." }, { "left": "issue:43761", "right": "issue:45859", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen3_5Moe CPU-offload incompatibility do not share a fix." + "reason": "CLIP hidden_states regression is unrelated to CPU-offload incompatibility in Qwen rotary embeddings." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Batched chat-template padding crash and Qwen2.5-VL video vision_position_ids bug are distinct." + "reason": "Both are Qwen2.5-VL, but batched chat-template padding crashes and wrong video vision_position_ids are distinct bugs." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "v5 custom model init breakage is not the same as patchtsmixer's optional post_init policy issue." + "reason": "The second is a specific patchtsmixer post_init constraint, not clearly the same concrete issue as the broad v5 custom-init regression." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug and a category-count runtime limit are unrelated." + "reason": "Qwen3.5 flash-attention generation failure is unrelated to the category-cardinality runtime error." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Recurrent-state reset with cache is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "Different model paths and failures: recurrent-state reset vs Whisper batch_decode token skipping." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading failure and Perceiver non-default-resolution failure are different model paths." + "reason": "Tokenizer loading failure and Perceiver image-resolution failure are unrelated." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and LasrFeatureExtractor center-arg crash are unrelated." + "reason": "Audio chat-template sampling-rate default and unsupported center arg in feature extraction are different bugs." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "Local-path import warning has nothing to do with Qwen3.5 recurrent-state reset." + "reason": "Local-path import warning is unrelated to recurrent-state cache behavior." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access and DeepseekV3MoE divergence are not the same bug." + "reason": "Unsafe sys.modules access and DeepSeekV3 divergence are different code paths." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and FastSpeech2ConformerTokenizer loading are different failures." + "reason": "Packed-weight loading in BitNet is unrelated to FastSpeech2ConformerTokenizer loading." }, { "left": "issue:42175", "right": "issue:46234", "accept": false, - "reason": "Missing TensorFlow backend after torch install is unrelated to a missing import_utils helper." + "reason": "Packaging/backend dependency coverage and a missing import_utils helper are not the same bug." }, { - "left": "issue:43493", - "right": "issue:43827", + "left": "issue:27957", + "right": "issue:29870", "accept": false, - "reason": "SigLIP2 implementation discrepancy and docs still referencing pipeline() are unrelated." + "reason": "Different model/task failures: XLM-RoBERTa FA2 vs BART generate() decoder_attention_mask bug." }, { - "left": "issue:27957", - "right": "issue:29870", + "left": "issue:43493", + "right": "issue:43827", "accept": false, - "reason": "XLM-Roberta Flash Attention 2 support and BART decoder_attention_mask generate error are different bugs." + "reason": "SigLIP2 implementation discrepancy and docs still referencing pipeline() are unrelated." } ] }, @@ -10977,13 +10939,13 @@ "issue:45325|issue:46697", "issue:44062|issue:45907", "issue:43873|issue:44871", - "issue:43749|issue:43931", - "issue:43643|issue:44913", "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:43749|issue:43931", "issue:43316|issue:45759", "issue:44442|issue:45803", - "issue:44190|issue:44496", "issue:44987|issue:46178", + "issue:44190|issue:44496", "issue:43761|issue:44079", "issue:43232|issue:43388", "issue:43931|issue:45216", @@ -11001,202 +10963,202 @@ "issue:16998|issue:22355", "issue:43295|issue:44871" ], - "prepared_review_unit_hash": "25d1f0f42823a7e364516cb1cd9593409daa39b2526b2c77dec2853528b7ccc4", + "prepared_review_unit_hash": "ce19fc3d99b54ed64bbd1a2ff91bf4c0d28bb61b9a99a49e072ae66b254ca79a", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36164, - "estimated_input_tokens": 9041, - "estimated_eval_tokens": 18338 + "serialized_chars": 36167, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "I found no true duplicate pairs in this set. The soft matches are mostly superficial keyword overlaps across different models, configs, or failure modes, so they should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical duplicate issue stands out; the candidate pairs are heterogeneous and do not describe the same underlying bug.", + "summary": "This cluster is heterogeneous: the issues span unrelated bugs in loading, generation, tokenization, config handling, offloading, and training utilities. None of the soft pairs appears to describe the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the set mixes unrelated failures across many subsystems with no shared root cause or duplicate thread.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the broadest and most actively discussed standalone report, but it still does not subsume the other issues.", + "best_issue_reason": "No single issue is a strong global representative; 43873 has the broadest/highest-signal discussion, but it still describes a different offloading/quantization problem than the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "Different problems: optional post_init validation vs a video-input regression for Qwen models." + "reason": "Different problems: a patchtsmixer post_init policy question versus a video-input regression in Qwen-like models." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config/API discrepancies, but they involve different fields and different model families." + "reason": "Both are config API mismatches, but they involve different missing fields and different models, so not the same bug." }, { "left": "issue:45325", "right": "issue:46697", "accept": false, - "reason": "Different code paths: Qwen2.5-VL rope/position-id scaling vs Swin2SR image padding." + "reason": "One is a Qwen2.5-VL rope/position_ids scaling bug; the other is Swin2SR image over-padding." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "Tokenizer AddedToken keyword conflict is unrelated to list_repo_files not forwarding revision." + "reason": "Tokenizer AddedToken argument collision is unrelated to list_repo_files not forwarding revision." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma-3 eos_token_id mismatch are distinct issues." + "reason": "Offloading/quantization behavior is unrelated to a Gemma-3 eos_token_id configuration inconsistency." }, { - "left": "issue:43749", - "right": "issue:43931", + "left": "issue:44855", + "right": "issue:45137", "accept": false, - "reason": "FSDP CPU-RAM-efficient loading breakage is unrelated to a Qwen3-VL weight-shape mismatch." + "reason": "Python 3.13 TorchScript parsing error is unrelated to a DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Missing fields from trust_remote_code loading is unrelated to GPTNeoX rotary_pct not persisting on reload." + "reason": "Both are reload/config issues, but one drops fields with trust_remote_code while the other resets rotary_pct on reload." }, { - "left": "issue:44855", - "right": "issue:45137", + "left": "issue:43749", + "right": "issue:43931", "accept": false, - "reason": "Python 3.13 JIT parsing failure and DeepSpeed ZeRO3 deque underflow are separate failures." + "reason": "FSDP CPU RAM efficient loading failure and Qwen3-VL weight-shape mismatch are different loading bugs." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both mention config objects, but one is a Gemma3TextConfig API discrepancy and the other is Qwen multimodal from_config unwrapping." + "reason": "A Gemma3TextConfig field discrepancy is not the same as failing to unwrap text_config in composite Qwen multimodal configs." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure and a bare-except bug in Fuyu tensor conversion are not the same bug." + "reason": "Tokenizer loading failure and a bare-except swallowing errors are unrelated code paths." }, { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:44987", + "right": "issue:46178", "accept": false, - "reason": "Local dataset loading in an example script is unrelated to unrecognized model/config loading." + "reason": "A model-loading regression is unrelated to a missing BACKENDS_MAPPING entry for tensorflow_text." }, { - "left": "issue:44987", - "right": "issue:46178", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "Loading physical-intelligence/fast failing is unrelated to a backend-mapping error for tensorflow_text." + "reason": "Local dataset loading in a training script is unrelated to an unrecognized model config/model_type error." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "Hidden_states missing from CLIPVisionModel.forward is not the same as generic ModelOutput key assignment behavior." + "reason": "CLIPVisionModel failing to return hidden_states is not the same as generic ModelOutput key assignment behavior." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs sync issue and gather_for_metrics label truncation are different trainer/runtime bugs." + "reason": "Generation kwargs syncing and gather_for_metrics label truncation are different training utilities with different failures." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Qwen3-VL load-time shape mismatch and Qwen3.5 save_pretrained corruption are related only at a very broad level." + "reason": "Weight-shape mismatch on load is unrelated to save_pretrained producing a wrong checkpoint for Qwen3.5." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "One is a missing output field; the other is divergent attention-kernel behavior under torch.compile." + "reason": "Hidden_states output regression is unrelated to attention-kernel divergence under torch.compile." }, { "left": "issue:46829", "right": "issue:46932", "accept": false, - "reason": "VideoPrism output field mix-up and Gemma4 quantized embedder dtype casting are unrelated." + "reason": "A wrong hidden_states field in classification output is unrelated to dtype casting for quantized Gemma4 embedders." }, { "left": "issue:43701", "right": "issue:46506", "accept": false, - "reason": "Checkpoint resume key mismatch and CPU/CUDA result divergence are separate issues." + "reason": "Checkpoint key mismatch has no relation to CPU/CUDA result divergence in PPDocLayoutV3." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "A parallel script failure is unrelated to Qwen2-57B-A14B-Instruct OOM during tp loading." + "reason": "3d_parallel.py failing and Qwen2-57B tensor-parallel OOM are different distributed-memory issues." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained are different save/load bugs." + "reason": "Saving a dequantized model and loading the wrong custom module after save_pretrained are different save/load bugs." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM under tensor-parallel load and key error loading MobileLLM are distinct model-loading failures." + "reason": "An OOM during tensor-parallel load is not the same as a key error when loading MobileLLM-125M." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile SDPA failure for Bamba is unrelated to resume_from_checkpoint key mismatch." + "reason": "torch.compile/SDPA behavior is unrelated to resume_from_checkpoint key mismatches." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and a NoneType broadcast error in Trainer are unrelated." + "reason": "XLMRoberta flash-attention support and a NoneType GPU broadcast error are unrelated." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is not the same as checkpoint key mismatch." + "reason": "CPU offload device mismatch is unrelated to checkpoint resume key mismatches." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k nesting behavior is unrelated to a missing PyTorch dependency." + "reason": "Pipeline top_k nesting behavior is unrelated to missing PyTorch dependency detection." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Processor/tokenizer regression and a pipeline translation error message issue are separate regressions." + "reason": "Processor/tokenizer regression in v4.57.5 is unrelated to a pipeline error-message wording issue in v5." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "DeBERTa-V3 model_max_length discussion is unrelated to a transformers.onnx import error." + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to the missing transformers.onnx import error." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer regression and Gemma-3 eos_token_id inconsistency do not share the same code path." + "reason": "Custom processor/tokenizer regression is unrelated to Gemma-3 eos_token_id inconsistency." } ] }, @@ -11276,8 +11238,8 @@ "issue:44077|issue:44534", "issue:43316|issue:43716", "issue:43316|issue:46821", - "issue:44871|issue:46050", "issue:43646|issue:44913", + "issue:44871|issue:46050", "issue:43452|issue:44496", "issue:45803|issue:46365", "issue:44162|issue:44792", @@ -11289,205 +11251,205 @@ "issue:43650|issue:46202", "issue:43540|issue:44514", "issue:43761|issue:44479", - "issue:45542|issue:46234", "issue:43065|issue:44617", + "issue:45542|issue:46234", "issue:43504|issue:45698", "issue:43572|issue:46178", - "issue:43388|issue:43688", "issue:44792|issue:44977", + "issue:43388|issue:43688", "issue:44792|issue:45600", "issue:45127|issue:45200" ], - "prepared_review_unit_hash": "509e00bfc451f279ff80979ff041e94c4582aa48396d3e046b2ce656b594c60a", + "prepared_review_unit_hash": "604c7ac5d159a1f6220561239c318f2c2103031f75710d40d18a9684db147652", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35748, - "estimated_input_tokens": 8937, - "estimated_eval_tokens": 18130 + "serialized_chars": 35742, + "estimated_input_tokens": 8936, + "estimated_eval_tokens": 18128 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is not a true duplicate cluster: the issues span many unrelated models, configs, and subsystems, with only superficial title similarity. All soft-edge pairs look like separate bugs/changes and should stay split.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue fits: the items cover many unrelated regressions across different models and subsystems.", + "summary": "This cluster is mostly a set of semantically similar but actually unrelated Transformers issues. The soft pairs share broad themes like loading, config, cache, or multimodal regressions, but not the same concrete bug/change path, so I would not merge any of them.", + "confidence": 0.83, + "canonical_issue_reason": "There is no single true canonical issue here because the cluster is heterogeneous. If an anchor is required, issue 45698 is the broadest representative of the loading/save-pretrained regression subtheme, but it should not be treated as a duplicate umbrella for the rest.", "canonical_pr_reason": null, - "best_issue_reason": "None of the issues is a strong global representative; the cluster lacks one underlying bug and should be split.", + "best_issue_reason": "Issue 45698 is the best global anchor: it has the strongest discussion/references and captures a common serialization/loading failure mode that appears more central than the other one-off regressions. Still, it is not a valid duplicate representative for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Both involve tokenization/eos handling, but they are different models and different failure modes (Gemma eos config mismatch vs Whisper empty transcription)." + "reason": "Different models and failure modes: Gemma EOS config inconsistency vs Whisper empty-transcription generation after token alignment." }, { "left": "issue:46234", "right": "issue:46291", "accept": false, - "reason": "Both are import-utils/dependency issues, but one is a missing availability helper and the other is a pipeline import regression from kernels version handling." + "reason": "Both are import/runtime dependency problems, but one is a missing utility symbol and the other is a pipelines import error from kernels versioning; not the same bug." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Both mention chunked generation/cache behavior, but one is output inconsistency under compiled forward and the other is a recompilation issue from lazy cache initialization." + "reason": "Both mention chunked generation/cache, but one is wrong outputs under compiled forward and the other is spurious recompilation from lazy cache init." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Both are v5 behavior regressions, but they describe different bugs: post_init allowance vs junk-filled non-persistent buffers." + "reason": "Different regressions: optional post_init validation for patchtsmixer vs non-persistent buffers being filled with junk in v5." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Different subsystems: config API discrepancy for Gemma3TextConfig vs image-preprocessor/model dtype mismatch for Mistral-3." + "reason": "Unrelated model/config discrepancy versus image preprocessor dtype mismatch; same release family at most, not the same bug." }, { "left": "issue:43316", "right": "issue:46821", "accept": false, - "reason": "These are unrelated config issues in different model families; one is Gemma text config API, the other is a TimesFM window_size attribute error." + "reason": "Gemma3TextConfig API mismatch is unrelated to TimesFM window_size AttributeError." }, { - "left": "issue:44871", - "right": "issue:46050", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "Both involve save/load roundtrips, but one is an eos_token_id config mismatch and the other is a broken quantized-model roundtrip." + "reason": "General v5 custom model init breakage vs GPTNeoX rotary_pct reload bug; different code paths and symptoms." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:44871", + "right": "issue:46050", "accept": false, - "reason": "The second is a specific GPTNeoXConfig reload bug; it does not look like the same underlying custom-model initialization breakage." + "reason": "EOS token config mismatch is unrelated to sinq quantization save/load corruption." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "Both are from_pretrained loading failures, but one is a gguf_file API break and the other is a missing model_type in config.json." + "reason": "Both are from_pretrained loading issues, but one is gguf_file handling and the other is missing model_type in config; different root causes." }, { "left": "issue:45803", "right": "issue:46365", "accept": false, - "reason": "Completely different failures: Fuyu tensor conversion exception handling vs Trainer import interference from datasets." + "reason": "Bare except in tensor conversion is unrelated to Trainer import failure caused by datasets dependency interaction." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "Unrelated model families and code paths: ESM2 breakage vs a janus image-generation test failure." + "reason": "ESM2 model breakage and Janus image-generation test failure are different subsystems and bugs." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Same family name, but different bugs: weight tying in Mistral3ForConditionalGeneration vs FlashAttention failure in Ministral-3." + "reason": "Mistral3 weight tying bug and Ministral-3 FlashAttention failure are distinct model issues." }, { "left": "issue:44242", "right": "issue:46894", "accept": false, - "reason": "Both touch training loss, but one is missing load-balancing loss under a flag and the other is a double-shift bug in ASR loss computation." + "reason": "Load-balancing loss gating is unrelated to CohereASR training-loss double-shift behavior." }, { "left": "issue:45632", "right": "issue:46424", "accept": false, - "reason": "Different cache bugs: path collision for local trust_remote_code models vs max_cache_length being ignored outside quantized cache." + "reason": "Cache path collision for trust_remote_code is unrelated to max_cache_length being honored only for quantized cache." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "No overlap in underlying bug: exception handling in FuyuBatchFeature vs Windows/ROCm import errors from torch.distributed dependency." + "reason": "Different failure classes: exception swallowing in tensor conversion vs Windows/ROCm import error due to torch.distributed dependency." }, { "left": "issue:44610", "right": "issue:46369", "accept": false, - "reason": "Both are vision-model issues, but one is a processor size mismatch and the other is an OOM report for sam3-litetext." + "reason": "Processor image-size mismatch is not the same as Sam3LiteText memory overflow." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "No shared underlying bug; this is a placeholder/data issue versus a Gemma tokenizer/model roundtrip failure." + "reason": "No meaningful relation: placeholder issue title vs Gemma roundtrip failure on a specific token." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both are Qwen multimodal bugs, but the failing code paths differ: video-input validation vs batched chat-template padding behavior." + "reason": "Different Qwen multimodal bugs: video input ValueError vs batched chat-template crash with padding=False." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Different models and regressions: CLIPVisionModel hidden_states output vs Qwen video-input regression." + "reason": "CLIPVision hidden_states regression is unrelated to Qwen video-input regression." }, { - "left": "issue:45542", - "right": "issue:46234", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "Unrelated backend/import problems: TensorBoard without TensorFlow vs missing import_utils helper for falcon mamba availability." + "reason": "Sam3PixelDecoder dummy Conv2d and Sam3Video OOM are not the same concrete defect." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:45542", + "right": "issue:46234", "accept": false, - "reason": "Same broad SAM3 area, but not the same bug: dummy Conv2d in Sam3PixelDecoder vs OOM in Sam3Video." + "reason": "TensorBoard/TensorFlow backend mismatch is unrelated to missing falcon mamba availability helper." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Both involve pretrained loading, but one is a legacy config field for BEiT and the other is wrong custom-module resolution after save/load." + "reason": "Legacy-field Beit load failure and wrong custom module after save/load are both loading bugs, but not the same code path." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "Completely unrelated: StableLmConfig pad_token_idx regression vs a tensorflow_text backend mapping error." + "reason": "StableLm pad_token_idx regression is unrelated to tensorflow_text backend mapping." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:44792", + "right": "issue:44977", "accept": false, - "reason": "Different training/eval bugs in different stacks: gather_for_metrics label truncation vs auxiliary-loss normalization." + "reason": "Janus test failure and Qwen3.5 flash-attention generation issue are different model behaviors." }, { - "left": "issue:44792", - "right": "issue:44977", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "Different failures in different models: janus image-generation test vs Qwen3.5 flash-attention generation regression." + "reason": "Mask2Former metric gathering bug is unrelated to auxiliary loss normalization in OLMoE/GPT Oss." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "One is a failed generation test, the other is an auto_mappings CI break from removed Sam3LiteText configs." + "reason": "Image generation test failure is unrelated to auto_mappings referencing removed Sam3LiteText configs." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "Different Gemma-related issues: LoRA/extended-vocab collapse vs missing mm_token_type_ids defaulting behavior." + "reason": "LoRA merge collapse with extended vocabulary is unrelated to Gemma 4 mm_token_type_ids defaulting." } ] }, @@ -11564,221 +11526,221 @@ "issue:43673|issue:44743", "issue:43352|issue:46640", "issue:44190|issue:44964", - "issue:41950|issue:43475", "issue:44534|issue:44913", + "issue:41950|issue:43475", "issue:43525|issue:43531", "issue:44734|issue:45406", "issue:35141|issue:42490", "issue:45072|issue:46932", "issue:43827|issue:43994", - "issue:44869|issue:46832", + "issue:44855|issue:45588", "issue:45245|issue:46858", + "issue:44869|issue:46832", "issue:46082|issue:46697", - "issue:44855|issue:45588", "issue:45632|issue:46617", - "issue:43673|issue:46421", "issue:29870|issue:35141", + "issue:43673|issue:46421", "issue:44617|issue:45584", "issue:43931|issue:44060", "issue:44242|issue:46897", "issue:43644|issue:44913", "issue:25251|issue:27957", - "issue:45563|issue:45584", "issue:43335|issue:44855", + "issue:45563|issue:45584", "issue:44933|issue:45362", - "issue:43550|issue:44190", - "issue:44077|issue:44877" + "issue:44077|issue:44877", + "issue:43550|issue:44190" ], - "prepared_review_unit_hash": "fd73615bfc11dd87d3cc6908b34ef1127df2260dc1a3fe60cd6a1f30f9ca56ef", + "prepared_review_unit_hash": "4e1b6ed0f388e63195a8a0cd96ca876919cdc0c7e68a60443a846073589939c0", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35848, - "estimated_input_tokens": 8962, - "estimated_eval_tokens": 18180 + "serialized_chars": 35852, + "estimated_input_tokens": 8963, + "estimated_eval_tokens": 18182 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are superficially similar in places, but they describe distinct bugs or changes across different models, pipelines, and code paths. None of the soft pairs should be merged as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue is safe here: the set spans unrelated bugs, so there is no shared underlying defect to represent the cluster.", + "summary": "This cluster is noisy: the items span many unrelated bugs across generation, caching, model configs, pipelines, and serving. I found no reliable duplicate pair among the soft edges, so this should not be merged into a single canonical triage bucket.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue is a reliable global representative; the closest broad one is `issue:43644`, but it still only covers a specific buffer-init bug and not the rest.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both involve cache behavior, but one is a missing GenerationMixin cache in chunked prefill and the other is recurrent-state reset logic in modular_qwen3_5; different bugs and code paths." + "reason": "Both mention cache/chunked prefill, but one is a missing GenerationMixin cache in v5.0.0 and the other is recurrent state reset in Qwen3 modular code; different bugs and code paths." }, { "left": "issue:43352", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but 43352 is a model-specific unsupported-feature report while 46640 is a different generic FA2 issue; not the same fix." + "reason": "One says NemotronHForCausalLM does not support Flash Attention 2.0; the other is a generic Flash Attention 2 issue. Too vague and not the same concrete failure." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "A local dataset loading failure in an example script is unrelated to a multimodal model load error in latest transformers." + "reason": "Different surfaces entirely: a local dataset loading failure in an example script versus a model loading error for Phi-4 multimodal." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:44534", + "right": "issue:44913", "accept": false, - "reason": "Pipeline processor lookup vs a SAM3 vision encoder attribute error are different subsystems and failures." + "reason": "One is a global non-persistent buffer initialization regression; the other is GPTNeoXConfig serialization/reload losing rotary_pct. Different defects." }, { - "left": "issue:44534", - "right": "issue:44913", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "One is a broad v5 buffer initialization regression; the other is a GPTNeoXConfig reload bug for `rotary_pct`." + "reason": "Video-classification pipeline processor selection and a SAM3 video encoder output attribute error are unrelated model/pipeline bugs." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "A missing `pad_token_id` attribute in Llama4Config is unrelated to Qwen3-MoE sliding-window behavior." + "reason": "Llama4Config pad_token_id absence and Qwen3-MoE sliding_window behavior are separate configuration issues." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Both touch transformers serve, but one is KV-cache indexing and the other is a missing `_tokenizer` on Gemma4Processor; distinct code paths." + "reason": "Both are serve-related, but one crashes on KV-cache continuation indexing while the other lacks a tokenizer on Gemma4Processor. Different code paths." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding reinitialization after resize is unrelated to import-path warning behavior for local models with trailing slashes." + "reason": "Resizing token embeddings reinitializing output embeddings is unrelated to local-path import warnings with a trailing slash." }, { "left": "issue:45072", "right": "issue:46932", "accept": false, - "reason": "Both involve dtype casting, but the affected models and concrete failures differ; this is not one mergeable bug." + "reason": "Both involve dtype behavior, but one is bfloat16 inference mismatches across models and the other is quantized Gemma4 vision/audio embedder casting; not the same bug." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing `pipeline()` is unrelated to a SigLIP2 AutoModel/pipeline inference issue." + "reason": "Docs still referencing removed pipeline APIs is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." }, { - "left": "issue:44869", - "right": "issue:46832", + "left": "issue:44855", + "right": "issue:45588", "accept": false, - "reason": "Whisper trailing replacement-character decoding crash and RT-DETR feature-level mismatch are different model-specific bugs." + "reason": "An import-time Python 3.13 IndentationError in DebertaV2 is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:45245", "right": "issue:46858", "accept": false, - "reason": "Category cardinality limits and cache sequence-length tensor arithmetic are unrelated defects." + "reason": "Category-cardinality overflow and static-cache slice arithmetic with device tensors are distinct runtime errors." }, { - "left": "issue:46082", - "right": "issue:46697", + "left": "issue:44869", + "right": "issue:46832", "accept": false, - "reason": "LlamaConfig head_dim validation is unrelated to Swin2SR image padding behavior." + "reason": "Whisper trailing replacement-character decode failure and RT-DETR feature-level mismatch are unrelated model-specific bugs." }, { - "left": "issue:44855", - "right": "issue:45588", + "left": "issue:46082", + "right": "issue:46697", "accept": false, - "reason": "A Python 3.13 TorchScript parsing issue is not the same as a flash-attention `s_aux=None` crash." + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are different configuration/processing issues." }, { "left": "issue:45632", "right": "issue:46617", "accept": false, - "reason": "Both reference trust_remote_code caching, but one is a leaf-directory path collision and the other is a file-not-found from cache; not the same concrete bug." + "reason": "Both involve trust_remote_code caching, but one is a leaf-directory name collision and the other is a FileNotFoundError when loading from cache; not the same concrete defect." }, { - "left": "issue:43673", - "right": "issue:46421", + "left": "issue:29870", + "right": "issue:35141", "accept": false, - "reason": "Both mention cache, but one is missing cache in prefill and the other is lazy-init causing recompilation; different symptoms and fixes." + "reason": "Bart generate() decoder_attention_mask initialization bug and embedding reinitialization after resize are unrelated." }, { - "left": "issue:29870", - "right": "issue:35141", + "left": "issue:43673", + "right": "issue:46421", "accept": false, - "reason": "Bart generate() decoder mask initialization and token-embedding resize reinit are unrelated." + "reason": "Both mention chunked prefill/cache, but one is a missing cache in v5.0.0 and the other is lazy cache initialization causing recompilation on non-CUDA backends; different symptoms and fixes." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation failure." + "reason": "Sam3Video CUDA OOM and Whisper empty-transcription handling are completely different failures." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "A Qwen3-VL weight-shape mismatch and a Qwen3-Next tied-weights warning are different model-loading issues." + "reason": "Qwen3-VL weight-shape mismatch and Qwen3-Next tied-weights warning are separate model-loading issues." }, { "left": "issue:44242", "right": "issue:46897", "accept": false, - "reason": "Load-balancing loss logic in MoE training is unrelated to Florence2's double-shift training-loss bug." + "reason": "Load-balancing loss omission in MoE routing and Florence2/Moonshine double-shift training loss are unrelated training bugs." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "A broad non-persistent-buffer regression is not the same as GPTNeoXConfig `rotary_pct` not persisting on reload." + "reason": "A global v5 non-persistent buffer regression is not the same as GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:25251", "right": "issue:27957", "accept": false, - "reason": "Pipeline `top_k` nesting behavior and XLMRoberta Flash Attention 2 support are unrelated." + "reason": "Top_k pipeline nesting bug and XLMRoberta Flash Attention 2 support are unrelated." }, { - "left": "issue:45563", - "right": "issue:45584", + "left": "issue:43335", + "right": "issue:44855", "accept": false, - "reason": "A stale warning in paged generate() is unrelated to Whisper's empty-transcription crash." + "reason": "SwitchTransformers sparse-layer creation and DebertaV2 Python parsing/import failure are unrelated." }, { - "left": "issue:43335", - "right": "issue:44855", + "left": "issue:45563", + "right": "issue:45584", "accept": false, - "reason": "SwitchTransformers config creating a sparse layer and a Python parsing/import error have no shared underlying bug." + "reason": "Paged generate warning behavior and Whisper empty transcription failure do not share the same underlying bug." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "A missing import from `image_utils` is unrelated to Qwen3.5 chat crashes." + "reason": "A nonexistent image_utils import and a Qwen3.5 chat crash are unrelated." }, { - "left": "issue:43550", - "right": "issue:44190", + "left": "issue:44077", + "right": "issue:44877", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to local dataset loading in an example script." + "reason": "Optional post_init policy for patchtsmixer and strict config loading for granite_speech are different configuration-policy issues." }, { - "left": "issue:44077", - "right": "issue:44877", + "left": "issue:43550", + "right": "issue:44190", "accept": false, - "reason": "PatchTSMixer post_init policy and strict granite_speech config loading are different configuration issues." + "reason": "Bamba torch.compile/SDPA failure and local dataset loading in an image-classification script are unrelated." } ] }, @@ -11876,224 +11838,221 @@ "issue:44462|issue:45907", "issue:43854|issue:44964", "issue:43673|issue:46858", - "issue:45632|issue:46234", "issue:43643|issue:44496", + "issue:45632|issue:46234", "issue:44246|issue:45003", "issue:44964|issue:45600", "issue:44112|issue:44964" ], - "prepared_review_unit_hash": "8b9f9d78e703d8669eecf0e8564c4723d45c02a90b6858e3a328befabf49553d", + "prepared_review_unit_hash": "343ac7dc26d7be1d1ee8b5abe07b4f57bc07fea46728f9eaf9540ebeb246c421", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36613, - "estimated_input_tokens": 9154, - "estimated_eval_tokens": 18564 + "serialized_chars": 36608, + "estimated_input_tokens": 9152, + "estimated_eval_tokens": 18560 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are only loosely related by subsystem or model family and should stay separate. The only clear duplicate-like pair is the DeepSeek tokenizer regression reports (44779 and 46489).", - "confidence": 0.84, - "canonical_issue_reason": "issue:46489 is the more specific DeepSeek tokenizer regression report and points to the suspected gap in PR #44801, making it the better canonical representative of the tokenizer bug cluster.", + "summary": "The cluster is highly heterogeneous: the issues span unrelated bugs in generation, multimodal preprocessing, model loading, CI, and config handling. None of the proposed soft pairs looks like the same underlying bug, so no merges are warranted.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:46489 has the clearest repro scope and root-cause context, so it is the strongest issue to anchor triage and deduping.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Different model families and failures: Qwen video position IDs vs Sam3LiteText memory overflow." + "reason": "Different bugs: Qwen2.5-VL video position IDs vs Sam3LiteText memory overflow." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Unrelated problems: CUDA assert during Qwen3 auto device mapping vs category-count limit in another path." + "reason": "Unrelated failures: Qwen3 auto device mapping CUDA assert vs category-count limit error." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Different code paths: Qwen video-input regression vs serve/KV-cache tensor indexing crash." + "reason": "Different code paths: multimodal video regression vs serve KV-cache continuation indexing." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "No shared bug: SwitchTransformers sparse-layer config issue vs local dataset loading in an example script." + "reason": "Sparse-layer config bug and local dataset loading are unrelated." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Different concrete issue: dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo processor/model input-size mismatch." + "reason": "Different model issues: dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo input sizing." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "Docs/example problems in SigLIP2 vs local dataset loading failure; not the same underlying bug." + "reason": "SigLIP2 docs example errors are unrelated to dataset loading failure." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Both generation-related, but one is output inconsistency under compile and the other is a stale warning." + "reason": "Chunked-generation inconsistency and stale warning behavior are not the same bug." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Unrelated areas: LoRA/extended-vocab model collapse vs incorrect visual encoder keys on save_pretrained." + "reason": "LoRA/vocab collapse and visual-encoder save_pretrained keys are distinct issues." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both video-related, but different models and failure modes: vision_position_ids vs padding calculation." + "reason": "Both are video-related, but the model/code paths and failure modes differ." }, { "left": "issue:44617", "right": "issue:46369", "accept": false, - "reason": "Same broad Sam3 family, but one is OOM in Sam3Video and the other is memory overflow in Sam3-LiteText; not the same bug." + "reason": "Sam3Video OOM and Sam3LiteText memory overflow are different variants and symptoms." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Different regressions: processor/tokenizer API breakage vs resize_token_embeddings not updating tied embeddings." + "reason": "Processor.tokenizer regression and gemma4 embedding resize behavior are unrelated." }, { "left": "issue:44779", "right": "issue:46489", - "accept": true, - "reason": "Both report the same DeepSeek tokenizer regression in v5+; 46489 is a more specific instance of the same underlying bug." + "accept": false, + "reason": "Related family, but one is a broad tokenizer regression and the other a specific PR gap." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "Both CI-related, but one is a stale GraniteSpeech device-override test and the other is a broken auto_mappings reference." + "reason": "Stale GraniteSpeech CI test vs removed Sam3LiteText config references are unrelated." }, { "left": "issue:45290", "right": "issue:46752", "accept": false, - "reason": "Same API surface, but different failures: tool-call assistant messages vs empty list IndexError." + "reason": "Both hit apply_chat_template, but one is tool-call handling and the other empty-list handling." }, { "left": "issue:43526", "right": "issue:46697", "accept": false, - "reason": "Different image processor bugs: reduced labels vs over-padding when already divisible." + "reason": "BEiT label reduction and Swin2SR over-padding are different image-processor bugs." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "No shared root cause: embedding resize propagation vs Whisper empty-transcription generation failure." + "reason": "Gemma4 embedding resize and Whisper empty transcription are unrelated." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Different subsystems and symptoms: Qwen3OmniMoe video processing ValueError vs serve KV-cache indexing crash." + "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are separate problems." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "Unrelated issues: XLNet CPU tensor allocation vs Fuyu batch conversion swallowing exceptions." + "reason": "XLNet device placement bug and Fuyu bare-except handling are unrelated." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "Both multimodal/chat-template adjacent, but assistant-mask zeros and vision_position_ids are different bugs." + "reason": "Both are multimodal template issues, but they affect different outputs and models." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Different loading problems: Python 3.14 import failure vs TensorFlow/tensorboard backend misdetection." + "reason": "Python 3.14 import failure and tensorboard/TensorFlow backend issue are different environment problems." }, { "left": "issue:25251", "right": "issue:29870", "accept": false, - "reason": "Different pipeline/generation bug vs Bart generate() UnboundLocalError." + "reason": "Pipeline top_k nesting behavior and BART UnboundLocalError are unrelated generate-time bugs." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "Different API regressions: tokenizer.json selection vs list_repo_files revision forwarding." + "reason": "AutoTokenizer tokenizer.json lookup and list_repo_files revision forwarding are unrelated API bugs." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both model-loading failures, but for different models and likely different config/test issues." + "reason": "Different model-load regressions affecting different models and loaders." }, { "left": "issue:43673", "right": "issue:46858", "accept": false, - "reason": "Both cache-related, but one is a missing cache in chunked prefill and the other is a device-tensor arithmetic bug in static cache." + "reason": "Both involve cache, but missing cache in chunked_prefill and device-tensor seq_length arithmetic are distinct." }, { - "left": "issue:45632", - "right": "issue:46234", + "left": "issue:43643", + "right": "issue:44496", "accept": false, - "reason": "Unrelated: trust_remote_code cache-path collision vs missing import_utils helper." + "reason": "trust_remote_code config-field loss and unrecognized model_type are different config-loading failures." }, { - "left": "issue:43643", - "right": "issue:44496", + "left": "issue:45632", + "right": "issue:46234", "accept": false, - "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model due to missing model_type." + "reason": "Local trust_remote_code cache collision and missing import_utils symbol are unrelated." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness and unsafe sys.modules access are related at a high level, but not the same concrete bug." + "reason": "Import latency and unsafe sys.modules access are different import-system issues." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Different models and failure modes: Phi-4 multimodal load failure vs Sam3LiteText CI breakage." + "reason": "Phi-4 multimodal loading and removed Sam3LiteText auto_mappings references are unrelated." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI device-override test failure is unrelated to Phi-4 multimodal loading." + "reason": "GraniteSpeech CI flake and Phi-4 loading failure do not share a concrete bug." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded overall. The only accepted pair (44779/46489) is supported by matching DeepSeek tokenizer regression titles, and the other soft-edge verdicts stay conservative by distinguishing different models, subsystems, or failure modes without overclaiming duplicates." - }, - "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 }, @@ -12167,8 +12126,8 @@ "issue:36246|issue:41628", "issue:32090|issue:35141", "issue:43404|issue:45216", - "issue:36246|issue:38617", "issue:43531|issue:45600", + "issue:36246|issue:38617", "issue:41950|issue:44016", "issue:44295|issue:46506", "issue:43650|issue:44220", @@ -12189,229 +12148,223 @@ "issue:44496|issue:45759", "issue:43976|issue:44279", "issue:44016|issue:44062", - "issue:44792|issue:46424", - "issue:43441|issue:45588" + "issue:43441|issue:45588", + "issue:44792|issue:46424" ], - "prepared_review_unit_hash": "53a15cfe0dd0806a874cbb0965c354bf2994164891f294cddca4368426ec9041", + "prepared_review_unit_hash": "6582b5dbf8c760da96a85912c9ecfb7054a637abc866ba9de31795e661bb294b", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37051, - "estimated_input_tokens": 9263, - "estimated_eval_tokens": 18782 + "serialized_chars": 37053, + "estimated_input_tokens": 9264, + "estimated_eval_tokens": 18784 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are thematically related to Transformers, but the soft pairs mostly share only broad subsystem keywords. I do not see any true duplicate-equivalent pairs; the bugs touch different models, APIs, and failure modes.", - "confidence": 0.59, - "canonical_issue_reason": "No single issue is an obvious duplicate target across this heterogeneous set. If one representative anchor is needed, issue #44521 is the clearest user-facing regression with ongoing discussion and inbound references.", + "summary": "This cluster is not a duplicate set; the items span many unrelated Transformers bugs, regressions, and documentation issues. None of the soft pairs look like the same underlying bug/change, so all are rejected.", + "confidence": 0.93, + "canonical_issue_reason": "No true canonical issue stands out because the cluster is heterogeneous. If one must be chosen, issue:44521 is the most active and broadly referenced, but it is still only a specific multimodal `apply_chat_template` bug and does not represent the rest of the items.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #44521 is the most broadly scoped and well-documented item here, making it the best representative issue for this loose cluster of Transformers regressions.", + "best_issue_reason": "issue:44521 is the best single issue to keep as a representative ticket because it is open, has the most inbound references and discussion, and describes a concrete reproducible bug with clear user impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Different bugs: FlashAttention 2 support for XLMRoberta vs embedding reinitialization after resize_token_embeddings." + "reason": "Both mention embedding/Flash Attention related wording only at a very high level; one is about XLM-RoBERTa with Flash Attention 2, the other about embedding reinitialization after resizing. Different bugs and code paths." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Different scope: missing TensorFlow in a torch install vs a vague dependency problem with no shared code path." + "reason": "One is about TensorFlow not being included when installing the torch extra; the other is a vague dependency issue. No concrete shared bug." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE-marker handling and multimodal assistant-mask generation are unrelated failures." + "reason": "Tokenizer BPE marker handling and multimodal assistant mask generation are unrelated problems in different subsystems." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "Both are config-validation related, but one is Llama head_dim rejection and the other is rope-key list typing; not the same bug." + "reason": "Llama head_dim validation and rope-parameter JSON list handling are both config-validation topics, but they are distinct failures with different inputs and code paths." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Missing Qwen2_5_VLImageProcessor export is different from AutoImageProcessor import failure." + "reason": "Both are import errors, but for different symbols and modules (`Qwen2_5_VLImageProcessor` vs `AutoImageProcessor`). Not the same underlying bug." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding resize/post_init reinitialization." + "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding reinitialization after resize." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both involve model weights/saving, but the concrete failures are different models and different code paths." + "reason": "Both involve model configuration/loading for multimodal models, but one is about untied lm_head weights in Mistral3 and the other about Qwen3.5 checkpoint saving regression. Different concrete defects." }, { - "left": "issue:36246", - "right": "issue:38617", + "left": "issue:43531", + "right": "issue:45600", "accept": false, - "reason": "Both are import errors, but they concern different symbols and different modules." + "reason": "Sliding-window behavior for Qwen3-MoE and auto_mappings referencing removed configs are separate issues with no shared code-path." }, { - "left": "issue:43531", - "right": "issue:45600", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and stale auto_mappings references are separate issues." + "reason": "Both are import-related breakages, but they concern different missing exports from different files." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline image-processor lookup bug and notebook syntax error are unrelated." + "reason": "Video-classification pipeline looking for image processors is unrelated to a notebook syntax error." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "Position_ids buffer access error is unrelated to CPU/CUDA numerical differences in PPDocLayoutV3." + "reason": "Registering `position_ids` as a buffer causing read errors and CPU/CUDA output differences in PPDocLayoutV3 are different behavioral bugs." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "The first item is not a concrete bug report; the second is an audio feature-extraction issue." + "reason": "The titles are generic, but one concerns missing data and the other `_torch_extract_fbank_features()`; there is no evidence they are the same bug." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "Both mention cache, but one is a GenerationMixin chunked-prefill regression and the other is a cache-length config bug." + "reason": "Generation cache missing in v5.0.0 during chunked_prefill and cache max length only honored for Quantized Cache are related to caching, but they are distinct features and failure modes." }, { "left": "issue:43582", "right": "issue:46932", "accept": false, - "reason": "Apple Silicon allocator TypeError and Gemma4 quantized embedder dtype casting are different failures." + "reason": "Apple Silicon TypeError in `caching_allocator_warmup` and Gemma4 dtype casting for quantized embedders are different issues in different subsystems." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor/tokenizer regression and t5gemma decoder embedding resize bug are not the same code-path problem." + "reason": "Processor/tokenizer regression and t5gemma decoder embedding resize behavior are not the same underlying bug." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python 3.9 compatibility breakage is different from a specific model-loading failure on v5.1.0." + "reason": "Python version incompatibility and model-loading failure for `physical-intelligence/fast` are different release/load issues." }, { "left": "issue:45083", "right": "issue:46506", "accept": false, - "reason": "Qwen3_omni_moe feature-length helper behavior is unrelated to PPDocLayoutV3 CPU/CUDA divergence." + "reason": "A helper length calculation bug in qwen3_omni_moe is unrelated to CPU/CUDA result divergence in PPDocLayoutV3." }, { "left": "issue:44016", - "right": "issue:44062", - "accept": false, - "reason": "Notebook syntax error and AddedToken duplicate keyword TypeError are unrelated." - }, - { - "left": "issue:44792", - "right": "issue:46424", + "right": "issue:44220", "accept": false, - "reason": "Failed janus image-generation test is not the same bug as cache max length handling." + "reason": "A notebook syntax error and an fbank feature extraction issue are unrelated." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Unrecognized model config and PIL image processors requiring torchvision are different loading/runtime issues." + "reason": "Unrecognized model type in config loading and PIL image processors requiring torchvision are different bugs." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug and documentation mismatch for return_all_scores do not share a concrete fix." + "reason": "A BEiT image-processor label reduction bug and docs mismatch for `return_all_scores` are unrelated." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both involve flex_attention, but one is a torch>=2.9 compilation failure and the other is an incorrect version gate for AuxRequest import." + "reason": "Compiled flex_attention failing on torch>=2.9 and incorrect PyTorch version check for AuxRequest import are distinct flex_attention-related issues, but not the same code-path problem." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary losses when denoising is off are unrelated model-training bugs." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and FlashAttention failure in Ministral-3 are distinct bugs." + "reason": "Saving dequantized Ministrals/Devstrals and Ministral-3 flash-attention failure are different issues in different parts of the stack." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Docs mentioning return_all_scores is not the same as failing to recognize a model in config.json." + "reason": "Docs mentioning `return_all_scores` and an unrecognized model config error are not the same bug." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Cache missing in GenerationMixin and ANSI codes in loading_report are unrelated." + "reason": "Cache behavior in generation and ANSI codes in loading_report are unrelated." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Default multimodal config initialization and missing deepstack_visual_indexes are different config-schema bugs." + "reason": "Phi4 multimodal default config initialization and missing `deepstack_visual_indexes` in Qwen3_5MoeVisionConfig are different config-construction bugs." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Unrecognized model in config.json is unrelated to AutoModelForCausalLM.from_config not unwrapping text_config." + "reason": "An unrecognized model type error and `from_config` not unwrapping `text_config` for composite Qwen configs are distinct model-loading problems." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python version compatibility and a dependency issue are not the same underlying defect." + "reason": "Python version support and dependency issue are too vague and clearly not the same failure." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Syntax error in a notebook and AddedToken special argument collision are unrelated." + "reason": "Notebook syntax error and `AddedToken()` duplicate `special` keyword error are unrelated." }, { - "left": "issue:44792", - "right": "issue:46424", + "left": "issue:43441", + "right": "issue:45588", "accept": false, - "reason": "Test failure in janus image generation is not the same as cache max length being ignored." + "reason": "Ministral flash-attention failure and `flash_attention.py` crashing on `s_aux=None` are both flash-attention themed but involve different model assumptions and code paths." }, { - "left": "issue:43441", - "right": "issue:45588", + "left": "issue:44792", + "right": "issue:46424", "accept": false, - "reason": "Ministral FlashAttention failure and sink-less-model s_aux=None crash are different FlashAttention-related bugs." + "reason": "A failed janus image-generation test and cache max-length behavior are unrelated." } ] }, @@ -12449,19 +12402,19 @@ "issue:43854", "issue:43901", "issue:43976", - "issue:44016", "issue:44075", "issue:44242", "issue:44279", - "issue:44291", "issue:44327", "issue:44351", "issue:44368", + "issue:44462", "issue:44466", "issue:44521", "issue:44779", "issue:44933", "issue:45020", + "issue:45081", "issue:45200", "issue:45216", "issue:45276", @@ -12499,8 +12452,8 @@ "issue:43577|issue:45216", "issue:42796|issue:43650", "issue:46082|issue:46512", - "issue:43976|issue:45405", "issue:44351|issue:45278", + "issue:43976|issue:45405", "issue:41950|issue:45020", "issue:44779|issue:45812", "issue:43299|issue:45357", @@ -12511,204 +12464,204 @@ "issue:41950|issue:45823", "issue:43650|issue:46209", "issue:43479|issue:43525", - "issue:44016|issue:44291" + "issue:44462|issue:45081" ], - "prepared_review_unit_hash": "7fc1730a5e43d31c1826ba3146ab87857b1374fbbd3e01d1cc3e3b58e89e003c", + "prepared_review_unit_hash": "31b4743f0eb0bdbc4df714c6a8020cc9cefb79b8706ae3efa90fe619776411ac", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35567, - "estimated_input_tokens": 8892, - "estimated_eval_tokens": 18040 + "serialized_chars": 35588, + "estimated_input_tokens": 8897, + "estimated_eval_tokens": 18050 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: many issues share surface themes like regressions, import failures, tokenizer problems, or multimodal/model-loading breakage, but they point to different code paths and concrete failures. I don\u2019t see any soft pair that looks like the same underlying bug/change, so all candidate merges are rejected. The broadest issue here is the remote_code regression report, but the cluster does not form a clean single-topic duplicate set.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 45020 is the broadest and most representative regression report in the set; several others are narrower model-specific breakages, so it is the least-bad cluster anchor.", + "summary": "This cluster is mostly heterogeneous: the soft pairs share broad themes like tokenizers, import errors, configs, and multimodal regressions, but they do not look like the same concrete bug/change. I would use issue 45020 as the best anchor because it is the most general, well-described regression report in the set.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45020 is the broadest and most representative regression report here, with detailed repro context and high discussion activity, so it works best as an anchor for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 has the widest scope and best fits as a representative umbrella for a mixed regression cluster, even though it is not a true duplicate target for the other items.", + "best_issue_reason": "Issue 45020 is the strongest standalone issue in the set: it is detailed, high-signal, and describes a concrete regression with broader applicability than the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but they involve different symbols/modules and likely different fixes." + "reason": "Both are import-related, but they concern different missing symbols/modules and different failure modes; not the same underlying bug." }, { "left": "issue:44521", "right": "issue:46693", "accept": false, - "reason": "Multimodal assistant-mask correctness and FA2 long-context performance are different bugs." + "reason": "One is a multimodal masking bug in chat templating, the other is a long-context FA2 performance regression; unrelated code paths." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "Generic dependency complaints, but one is transformers-related and the other is a torch-version incompatibility." + "reason": "A dependency/install issue is not the same as a torch-version incompatibility; the underlying problems differ." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Resize-token-embeddings propagation and Whisper empty-transcription handling are unrelated code paths." + "reason": "These are different model behaviors in different subsystems; resize_token_embeddings and empty-transcription generation are unrelated." }, { "left": "issue:44327", "right": "issue:46752", "accept": false, - "reason": "QA top-k bounds crash and empty chat-template crash are different pipeline failures." + "reason": "Both are crashes, but they occur in different pipelines with different inputs and stack paths; not a duplicate bug." }, { "left": "issue:46123", "right": "issue:46369", "accept": false, - "reason": "Partial-batch pipeline bookkeeping and model memory overflow are not the same underlying issue." + "reason": "Batch-finalization logic in a pipeline is unrelated to a model memory overflow issue." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both touch Qwen3.5, but one is a load-time model issue and the other is save_pretrained checkpoint corruption." + "reason": "Both involve Qwen models, but one is a load-time regression and the other is a save_pretrained serialization bug; different code paths." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "Config validation for head_dim is unrelated to a missing decorator on rotary kernel registration." + "reason": "Config validation for head_dim is not the same as a kernel registration/decorator crash during rotary embedding setup." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Both mention dependencies, but transformers versioning and unreleased peft requirements are separate problems." + "reason": "Dependency issues and an unreleased minimum-version bump are different release/package problems." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Processor/tokenizer regression and Deepseek tokenization regression affect different APIs and failure modes." + "reason": "Both are tokenizer regressions, but they affect different models and distinct behaviors (processor/tokenizer API vs DeepSeek tokenization output)." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they concern different examples and different API behaviors." + "reason": "Documentation mistakes in different pipelines are not the same issue; the examples and broken behavior differ." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD argument handling and tie_word_embeddings warnings are unrelated." + "reason": "An optimizer argument bug is unrelated to a model warning about tied word embeddings." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "A dtype propagation bug in BLIP2 is distinct from Qwen3.5 save_pretrained key serialization." + "reason": "Loading dtype propagation in BLIP2 is unrelated to Qwen3.5 checkpoint serialization." }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "A tokenizer BPE-marker bug and a placeholder 'ADD THE DATA' issue are not duplicates." + "reason": "Tokenizer BPE marker handling has no relation to the unrelated placeholder issue 'ADD THE DATA'." }, { "left": "issue:46082", "right": "issue:46512", "accept": false, - "reason": "One is a config rejection case; the other is stale config dtype after from_config, so they are different bugs." + "reason": "These both mention config fields, but one is a validation error and the other is stale dtype metadata; different bugs." }, { - "left": "issue:43976", - "right": "issue:45405", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "Python-version support and unreleased dependency versioning are separate compatibility problems." + "reason": "45278 is a broad umbrella over many import errors, but 44351 is a specific missing import; not enough evidence they are the same concrete defect." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:45405", "accept": false, - "reason": "A specific missing import and a broad upgrade-induced import breakage are not the same concrete bug." + "reason": "A Python-version compatibility report and an unreleased dependency version bump are different problems." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Pipeline processor lookup bug and broad remote_code regression are too different to merge." + "reason": "Video pipeline processor selection and remote_code model breakage are unrelated subsystems." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they are model-specific and hit different tokenization behavior." + "reason": "These are different tokenizer regressions affecting different model families and token-ID behaviors." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Loading failure and incorrect save_pretrained visual encoder keys are different Qwen3.5 issues." + "reason": "Both are Qwen3.5-related, but one is a loading failure and the other is incorrect save_pretrained visual-encoder keys; different bugs." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Serialization of tied weights and resize_token_embeddings propagation are distinct problems." + "reason": "Both concern weight/tied-embedding serialization or resizing, but the specific failure modes and affected components differ." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Different model families and different failure modes; no shared code-path evidence." + "reason": "Different models, different loading failures, and no shared concrete code path are evident." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "Tokenizer special-token decoding and D-FINE auxiliary-loss logic are unrelated." + "reason": "A tokenizer special-token registration bug and a missing auxiliary-loss issue are unrelated." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss and Gemma4 token-id defaults are different model-training bugs." + "reason": "These are both model-training logic bugs, but one is router loss handling and the other is defaulting multimodal token type IDs; different fixes." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Video pipeline processor lookup and Gemma4 PLE device placement are unrelated." + "reason": "Video classification processor lookup and Gemma4 device placement are unrelated." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "A placeholder issue cannot be treated as the same underlying bug as FP8 inference on B200." + "reason": "A placeholder/triage issue is not the same as an FP8 inference failure on B200." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Default multimodal config initialization and missing pad_token_id on Llama4Config are separate config issues." + "reason": "Different config classes and different missing/default-field behavior; not the same underlying defect." }, { - "left": "issue:44016", - "right": "issue:44291", + "left": "issue:44462", + "right": "issue:45081", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument mismatch are not the same change or bug." + "reason": "Both touch tokenizer loading, but one is AutoTokenizer repository resolution and the other is a Mistral regex patch crash; different code paths." } ] }, @@ -12754,9 +12707,9 @@ "issue:44246", "issue:44261", "issue:44263", + "issue:44291", "issue:44336", "issue:44361", - "issue:44462", "issue:44589", "issue:44623", "issue:44671", @@ -12781,7 +12734,7 @@ "issue:46851" ], "soft_pairs": [ - "issue:44462|issue:45081", + "issue:44016|issue:44291", "issue:44683|issue:45542", "issue:44361|issue:45593", "issue:43441|issue:45834", @@ -12797,8 +12750,8 @@ "issue:42175|issue:43976", "issue:44671|issue:44779", "issue:43901|issue:43994", - "issue:43673|issue:43761", "issue:44623|issue:45397", + "issue:43673|issue:43761", "issue:43901|issue:44246", "issue:43723|issue:44162", "issue:43525|issue:43881", @@ -12810,202 +12763,202 @@ "issue:43525|issue:43643", "issue:44589|issue:45362" ], - "prepared_review_unit_hash": "48003a776e0040ef57abe661c23d4cebf6ffd203d0d5855b93fa18e142b3ef6d", + "prepared_review_unit_hash": "d76137ee7c6f60395b452c9a4d5bddd9e427506744e3d57526f1d157df51a103", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35353, - "estimated_input_tokens": 8839, - "estimated_eval_tokens": 17934 + "serialized_chars": 35434, + "estimated_input_tokens": 8859, + "estimated_eval_tokens": 17974 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated closed issues. The soft edges are generally only superficial topic similarity (tokenizers, loading, v5 regressions, deepspeed, quantization) and do not describe the same underlying bug or change.", - "confidence": 0.89, - "canonical_issue_reason": "No single issue cleanly serves as a duplicate canonical for the whole set; the items span many distinct code paths and models rather than one shared defect.", + "summary": "This cluster is a loose collection of unrelated Transformers issues, mostly v5 regressions and model-specific bugs. None of the proposed soft pairs look like true duplicates, so they should stay separate.", + "confidence": 0.9, + "canonical_issue_reason": "issue:44291 is a clear, concrete core-library regression with a reproducible loading failure and a well-defined API break, making it the strongest single anchor for this noisy cluster.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue must represent the set, #43723 is the broadest user-facing loading regression, but it is still not a duplicate target for the others.", + "best_issue_reason": "issue:44291 is the best standalone representative: specific, actionable, and centered on a fundamental model-loading path rather than a niche model or docs-only problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44462", - "right": "issue:45081", + "left": "issue:44016", + "right": "issue:44291", "accept": false, - "reason": "Both touch tokenizer loading, but one is AutoTokenizer ignoring tokenizer.json and the other is a Mistral regex patch crash; different root causes and fixes." + "reason": "Notebook syntax error vs init_empty_weights TypeError; different surfaces and failure modes." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Completely different bugs: torch>=2.9 flex_attention compile failure vs packaging/backend install confusion." + "reason": "Compiled flex_attention failure vs missing TensorFlow backend dependency; unrelated bugs." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "Different subsystems and failure modes: MLukeTokenizer AttributeError vs D-FINE auxiliary loss behavior." + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; no shared code path." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "Both are model-specific bugs, but one is FlashAttention on Ministral-3 and the other is Kosmos2.5 OCR index errors; not the same issue." + "reason": "FlashAttention model failure vs long-input OCR index error; different model-specific issues." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex loading crash vs list_repo_files revision forwarding; unrelated code paths." + "reason": "Mistral tokenizer regex patch crash vs repo file revision forwarding; unrelated functionality." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Both involve deepspeed/optimizer behavior, but one is OOM with Adafactor and the other is duplicate optimizer steps; distinct bugs." + "reason": "Both mention DeepSpeed, but one is Adafactor OOM and the other is double optimizer stepping; distinct bugs." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Auto-docstring future-annotations crash and ReasoningEffort import failure are unrelated implementation defects." + "reason": "Auto-docstring annotations crash vs missing ReasoningEffort import; separate failure paths." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Both relate to quantized/dequantized models, but one is saving failure and the other is load/offload with device_map=auto; different paths." + "reason": "Quantized save failure vs dequantized load/offload failure; related theme but different operations and root causes." }, { "left": "issue:44336", "right": "issue:46851", "accept": false, - "reason": "ANSI codes in loading_report and Qwen3.5 variable-length position_ids are unrelated." + "reason": "ANSI code output in loading_report vs Qwen3.5 variable-length position_ids bug; unrelated." }, { "left": "issue:45859", "right": "issue:46851", "accept": false, - "reason": "Same model family, but CPU-offload rotary embedding bug is not the same as variable-length position_ids handling." + "reason": "Same model family, but CPU-offload rotary embedding issue and variable-length position_ids are different code paths." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "processor.save_pretrained missing files and a Chinese report about inference slowdown do not describe the same bug." + "reason": "Processor save_pretrained file emission bug vs performance regression in filtering; unrelated." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "Different model internals: missing rms_norm_eps precision issue vs torch.split return handling in GlmMoeDsaIndexer." + "reason": "MLA layernorm epsilon precision issue vs torch.split return handling in GlmMoeDsaIndexer; different components." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both are repo hygiene/CI-related, but one is stale auto_mappings references and the other is leftover dummy classes; distinct fixes." + "reason": "Stale auto_mappings config references vs leftover dummy classes; both hygiene issues, not the same bug." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "Packaging/backend dependency issue vs Python version compatibility regression; not the same underlying problem." + "reason": "Missing TensorFlow backend in torch install vs Python 3.9/3.10 compatibility; unrelated." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "Both are incorrect outputs in v5, but CamemBERT masked LM predictions and Deepseek tokenizer results are different model/code-path bugs." + "reason": "CamemBERT masked LM regression vs Deepseek tokenizer regression; different models and layers." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs mismatch about return_all_scores vs incorrect SigLIP2 outputs; unrelated." - }, - { - "left": "issue:43673", - "right": "issue:43761", - "accept": false, - "reason": "Both are v5 regressions, but one is GenerationMixin chunked_prefill cache and the other is CLIPVisionModel hidden_states; different fixes." + "reason": "Docs mismatch for return_all_scores vs incorrect model outputs; not the same issue." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different loading/saving issues." + "reason": "Processor save_pretrained file bug vs gemma-4 zero3 from_pretrained bug; different surfaces." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "GenerationMixin cache/chunked_prefill regression vs CLIPVisionModel hidden_states regression; unrelated." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Documentation mismatch vs intermittent slow import; no shared bug." + "reason": "Pipeline docs mismatch vs import slowness; different category entirely." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer loading regression and ESM2 breakage are unrelated." + "reason": "Tokenizer loading failure vs ESM2 model breakage; unrelated." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Different model/config failures: Llama4Config missing pad_token_id vs glm-4v-9b loading failure." + "reason": "Llama4 pad_token_id AttributeError vs glm-4v loading failure; different model-specific bugs." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "ImportError for Qwen2_5_VLForConditionalGeneration and Python version support are unrelated." + "reason": "Qwen2.5-VL import error vs Python version compatibility; no overlap in root cause." }, { "left": "issue:42796", "right": "issue:43441", "accept": false, - "reason": "Tokenizer BPE marker handling and FlashAttention failures are separate model-specific bugs." + "reason": "Ministral tokenizer BPE marker handling vs FlashAttention loading failure; not the same bug." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error and Gemma4 device-mismatch forward bug are unrelated." + "reason": "Notebook syntax error vs Gemma4 device_map auto device mismatch; unrelated." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorBoard/TF backend install issue vs torch 2.4.1 compatibility problem; different layers." + "reason": "Missing TensorFlow backend vs torch/transformers incompatibility; different compatibility problems." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence under torch.compile and ANSI output in loading_report have no common code path." + "reason": "Attention-kernel divergence under torch.compile vs ANSI-code output issue; unrelated." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both mention missing config fields, but one is a model-specific pad_token_id attribute error and the other is trust_remote_code returning incomplete configs." + "reason": "Missing pad_token_id on Llama4Config vs missing fields from AutoConfig with trust_remote_code; different config-loading bugs." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage lookup failure and Qwen3.5 chat crash are unrelated model/runtime issues." + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; unrelated runtime failures." } ] }, @@ -13080,8 +13033,8 @@ ], "soft_pairs": [ "issue:46291|issue:46399", - "issue:45325|issue:46851", "issue:44336|issue:44373", + "issue:45325|issue:46851", "issue:46655|issue:46726", "issue:44246|issue:44496", "issue:41628|issue:43976", @@ -13109,208 +13062,208 @@ "issue:44779|issue:45292", "issue:36246|issue:42757" ], - "prepared_review_unit_hash": "6bf17fc101ea88bb44b603abf9acf05a186d1dec40d5af0be82822c59494cfe7", + "prepared_review_unit_hash": "041b409d2b7c0ba93665fa07ca0819aa72f2a49c40559ec8309e7fdac576647f", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36456, - "estimated_input_tokens": 9114, - "estimated_eval_tokens": 18484 + "serialized_chars": 36460, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab bag of mostly unrelated Transformers bug reports (imports, model-specific regressions, tokenizer issues, and runtime errors). None of the soft pairs look like true duplicates of the same underlying bug/change.", - "confidence": 0.94, - "canonical_issue_reason": "No single duplicate issue clearly unifies the cluster; if one issue must serve as the representative, issue 46399 is the most concrete and actionable standalone bug report (narrow Qwen3.5 code-path failure with a specific registration error).", + "summary": "Mostly unrelated issue reports with a weak Qwen/model-loading overlap; no soft pair looks like the same concrete bug/change, so this cluster should not be merged into a single duplicate bucket.", + "confidence": 0.78, + "canonical_issue_reason": "issue:46851 is the cleanest representative of the strongest recurring theme in the cluster: a Qwen3.5 position_ids / variable-length-inputs regression. It is still not a duplicate of the others, but it is the most central and actionable issue among the set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46399 is the strongest standalone candidate: it describes a specific, reproducible failure in one code path and is easier to act on than the broader or more ambiguous reports.", + "best_issue_reason": "issue:46851 is the best overall issue to anchor the cluster because it is specific, well-scoped, and closest to the cluster\u2019s only repeated technical motif (Qwen position_ids / model-loading regressions).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46291", "right": "issue:46399", "accept": false, - "reason": "Both mention kernels/import behavior, but one is a pipelines import dependency issue and the other is a Qwen3.5 rotary/registration crash; different code paths and fixes." + "reason": "Both touch imports/kernels, but one is a kernels package version/revision dependency failure and the other is a Qwen3.5 rotary registration crash. Different code paths and fixes." }, { - "left": "issue:45325", - "right": "issue:46851", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Both are Qwen-family position-id bugs, but one is Qwen2.5-VL temporal scaling and the other is Qwen3.5 variable-length chat input handling; not the same underlying defect." + "reason": "ANSI codes in loading_report and a wrong docstring for position_ids are unrelated documentation/util issues." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:45325", + "right": "issue:46851", "accept": false, - "reason": "ANSI-code logging behavior and a docstring typo are unrelated." + "reason": "Both mention position_ids, but one is Qwen2.5-VL temporal scaling and the other is Qwen3.5 variable-length inputs. Different models and bugs." }, { "left": "issue:46655", "right": "issue:46726", "accept": false, - "reason": "Both touch vision processing, but one is a SmolVLM video padding bug and the other is an AutoProcessor missing-backend import/error path; different failures." + "reason": "Both are vision/processor issues, but one is a SmolVLM video padding bug and the other is an AutoProcessor missing-backend error." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "A performance complaint on importing transformers is unrelated to an unrecognized-model/config.json error." + "reason": "Import-time slowness is unrelated to an unrecognized model/config loading failure." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "An import error for AutoImageProcessor is not the same as a Python-version compatibility regression." + "reason": "One is an AutoImageProcessor import failure and the other is a Python-version compatibility problem; not the same bug." }, { "left": "issue:45083", "right": "issue:46851", "accept": false, - "reason": "Different Qwen variants and different helpers: output-length calculation vs variable-length position_ids." + "reason": "Both are Qwen-family issues, but they affect different helpers/code paths (_get_feat_extract_output_lengths vs position_ids handling)." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "These are separate import/back-end issues affecting different packages and failure modes." + "reason": "Different import/backend errors in different integrations; no shared underlying defect." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to a gemma-4 from_pretrained/zero3 issue." + "reason": "Tokenizer save metadata mismatch is unrelated to gemma-4 zero3 from_pretrained behavior." }, { "left": "issue:46399", "right": "issue:46851", "accept": false, - "reason": "Same broad model family, but one is a use_kernels registration crash and the other is variable-length input/position_ids behavior; different bugs." + "reason": "Qwen3.5 kernel registration crash vs variable-length position_ids crash: different failure sites and fixes." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "Pytest-xdist file races and a stale generate() warning are unrelated." + "reason": "A pytest-xdist file race and a stale generate warning are unrelated testing/runtime issues." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Auto-docstring annotation handling has no relation to Qwen3.5 kernel registration." + "reason": "Auto-docstring annotation parsing and Qwen3.5 kernel registration are unrelated." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "A BeitImageProcessorFast label-reduction bug is unrelated to unsafe sys.modules access in modeling_utils." + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are different bugs in different subsystems." }, { "left": "issue:43540", "right": "issue:46752", "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe and empty-chat-list handling are different code paths." + "reason": "Video-input processing in Qwen3OmniMoe and empty-list chat templating are unrelated input-validation failures." }, { "left": "issue:45072", "right": "issue:46829", "accept": false, - "reason": "These are distinct model-specific inference/output bugs with no shared underlying change." + "reason": "Different model outputs are mismatched in different classes; these are not the same bug." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both concern Qwen MoE loading/versioning, but one is LoRA portability across hf4.x/hf5.x and the other is a Qwen3VL MoE loading regression in dev0; not the same bug." + "reason": "Both involve Qwen3 MoE loading across versions, but one is LoRA compatibility on hf4.x/hf5.x and the other is Qwen3VL-MoE loading on 5.0.0.dev0. Similar theme, not the same concrete defect." }, { "left": "issue:46614", "right": "issue:46665", "accept": false, - "reason": "Phonemizer language caching and from_pretrained behavior inside Python classes are unrelated." + "reason": "Stale phonemizer backend state and inconsistent from_pretrained behavior inside classes are unrelated." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and a Qwen3.5 chat crash are different failures." + "reason": "Both mention Qwen, but one is sliding_window behavior in Qwen3-MoE and the other is a chat crash in Qwen3.5. Different code paths." }, { "left": "issue:45588", "right": "issue:46851", "accept": false, - "reason": "Flash-attention sinkless-model crash and variable-length position_ids are unrelated." + "reason": "Flash-attention s_aux=None crash is unrelated to Qwen3.5 variable-length position_ids handling." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import performance and Gemma eos_token_id configuration are unrelated." + "reason": "Import latency and eos_token_id config inconsistency are unrelated." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Both mention missing auxiliary loss under a flag, but they are different models and different loss mechanisms, so not the same underlying bug." + "reason": "Load-balancing loss omission and D-FINE auxiliary-loss handling are different model-training bugs." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "A position_ids buffer access error is unrelated to gemma-4 from_pretrained/zero3 behavior." + "reason": "Position_ids buffer access error and gemma-4 zero3 from_pretrained are unrelated." }, { "left": "issue:45003", "right": "issue:46821", "accept": false, - "reason": "Unsafe sys.modules access and a TimesFM window_size AttributeError are separate defects." + "reason": "Unsafe sys.modules access and TimesFM window_size AttributeError are different defects." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "Loading-report ANSI codes and GLM-5 RoPE implementation are unrelated." + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline tokenizer cache loading and AttentionInterface.register behavior changes are different problems." + "reason": "Offline cache loading and AttentionInterface.register behavior changes are separate issues." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and CPU-vs-CUDA output differences in PPDocLayoutV3 are unrelated." + "reason": "Bamba torch.compile/SDPA failure and PPDocLayoutV3 CPU/CUDA divergence are not the same bug." }, { "left": "issue:44451", "right": "issue:46752", "accept": false, - "reason": "Model-loading failure for ScandiBERT and empty chat-template handling are unrelated." + "reason": "ScandiBERT loading failure and empty chat-template input crash are unrelated." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer correctness regression and output-embedding resizing are different code paths." + "reason": "Deepseek tokenizer regression and resize_token_embeddings output mismatch are different problems." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but one is a missing model export and the other is a missing hub utility; unrelated root causes." + "reason": "Two import errors from different packages and APIs; no shared underlying code-path." } ] }, @@ -13391,14 +13344,14 @@ "issue:44496|issue:45003", "issue:43540|issue:46655", "issue:45003|issue:45561", - "issue:45375|issue:45759", "issue:43262|issue:44246", + "issue:45375|issue:45759", "issue:44162|issue:45200", "issue:44479|issue:46655", "issue:45859|issue:46399", "issue:44295|issue:44749", - "issue:43475|issue:46369", "issue:45137|issue:45600", + "issue:43475|issue:46369", "issue:43295|issue:45292", "issue:45292|issue:45584", "issue:44466|issue:45335", @@ -13413,14 +13366,14 @@ "issue:42199|issue:46851", "issue:44479|issue:46170" ], - "prepared_review_unit_hash": "072deda892b764d4f50e40b95e598276fa6676fb56830f51f5dbfcfea3642511", + "prepared_review_unit_hash": "2596e20c4a26a81368cea371f7b04b270127de9f028794eae644f11b11c45ea4", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, @@ -13435,180 +13388,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs look like false positives. The set is highly heterogeneous: unrelated bugs across model loading, video/audio processors, training, serialization, and infrastructure.", - "confidence": 0.97, - "canonical_issue_reason": "No true canonical issue stands out; these items are not a duplicate cluster but a mixed bag of unrelated problems.", + "summary": "No soft pair looks like the same underlying bug/change, so I would keep all issues separate. The set spans unrelated regressions across generation, import paths, video/audio processors, model configs, training, and DeepSpeed behavior.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If forced to choose an anchor, issue:45003 is the broadest core-code bug, but it still does not represent the rest of the set well.", + "best_issue_reason": "Issue #45561 is the best representative to anchor if needed: it is open, has the most inbound references, and reports a concrete, reproducible race in patched testing utilities.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for TextClassificationPipeline vs unsafe sys.modules access in modeling_utils; different bugs." + "reason": "Docs mismatch for TextClassificationPipeline and unsafe sys.modules access are unrelated problems in different code paths." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Different import failures: missing Qwen2_5_VLImageProcessor vs missing HybridCache." + "reason": "Both are import errors, but they refer to different missing symbols and different transformers modules." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MOE weight conversion under accelerate/deepspeed is unrelated to Zero3 deque underflow." + "reason": "Both involve DeepSpeed/large-model startup, but one is a Qwen3 MOE weight conversion issue and the other is a ZeRO3 deque IndexError." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video pipeline image-processor lookup is unrelated to _torch_extract_fbank_features()." + "reason": "Video-classification processor behavior and _torch_extract_fbank_features are distinct subsystems with different failures." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config loading vs sys.modules access bug; not the same code path." + "reason": "Unknown-model config loading and sys.modules import safety are different bugs." }, { "left": "issue:43540", "right": "issue:46655", "accept": false, - "reason": "Qwen3OmniMoe video validation error vs SmolVLM padding-calculation bug." + "reason": "Both mention video inputs, but the Qwen3OmniMoe ValueError and SmolVLM padding bug are different code paths." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to the pytest-xdist captured_info.txt race." + "reason": "Unsafe sys.modules access and pytest-xdist file races are unrelated testing/runtime issues." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:43262", + "right": "issue:44246", "accept": false, - "reason": "Missing config field dropped by strict vs AutoModelForCausalLM unwrap behavior for composite configs." + "reason": "Audio chat-template sample-rate defaults and import latency are not the same bug." }, { - "left": "issue:43262", - "right": "issue:44246", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "Audio chat-template sampling-rate default is unrelated to import-time slowdown." + "reason": "Missing strict config field and AutoModelForCausalLM text_config unwrapping are different composite-config problems." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids default are different model bugs." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific bugs." }, { "left": "issue:44479", "right": "issue:46655", "accept": false, - "reason": "Qwen video-input regression vs SmolVLM video processor padding bug." + "reason": "Both are video-related regressions, but they affect different models and different failure mechanisms." }, { "left": "issue:45859", "right": "issue:46399", "accept": false, - "reason": "Both involve Qwen3.5, but CPU offload incompatibility and missing decorator registration are distinct failures." + "reason": "CPU offload incompatibility in rotary embedding and a missing decorator registration crash are not the same issue." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Buffer position_ids access error vs performance regression after upgrade; different symptoms and causes." + "reason": "Position_ids buffer access and multilingual slowdown after upgrade are unrelated symptoms." }, { - "left": "issue:43475", - "right": "issue:46369", + "left": "issue:45137", + "right": "issue:45600", "accept": false, - "reason": "Missing fpn_position_embeddings attribute vs memory overflow in sam3-litetext are not the same bug." + "reason": "Zero3 deque failure and stale auto_mappings references are different failures." }, { - "left": "issue:45137", - "right": "issue:45600", + "left": "issue:43475", + "right": "issue:46369", "accept": false, - "reason": "DeepSpeed Zero3 deque underflow is unrelated to auto_mappings referencing removed Sam3LiteText configs." + "reason": "Sam3VisionEncoderOutput attribute error and SAM3-LiteText memory overflow are distinct bugs." }, { "left": "issue:43295", "right": "issue:45292", "accept": false, - "reason": "Processor/tokenizer regression differs from resize_token_embeddings not updating output embeddings." + "reason": "Processor.tokenizer regression and resize_token_embeddings output-embedding update are different code paths." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation bug and Whisper empty transcription bug are unrelated." + "reason": "Embedding resize behavior and Whisper empty-transcription generation failure are unrelated." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Serialization inconsistency of tied weights is different from t5gemma resize_token_embeddings propagation." + "reason": "Both concern weight/embedding consistency, but one is lm_head serialization and the other is decoder.embed_tokens resizing." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Strict config loading failure is unrelated to missing deepstack_visual_indexes field handling." + "reason": "Strict config loading and a missing config field are different mechanisms, even though both are config-related." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Different import/runtime failures: missing Qwen2.5 VL image processor vs Zero3/BertModel index error." + "reason": "A missing Qwen2.5-VL image processor import and a ZeRO3 BertModel IndexError are unrelated." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Missing load-balancing loss is unrelated to float16 AMP crash in LwDetrImageLoss." + "reason": "Missing load-balancing loss and an AMP/CUDA crash in LwDetrImageLoss are different training bugs." }, { "left": "issue:44623", "right": "issue:46726", "accept": false, - "reason": "processor.save_pretrained missing files is not the same as AutoProcessor backend-missing image-processor error." + "reason": "processor.save_pretrained missing files and AutoProcessor backend-missing import handling are separate processor issues." }, { "left": "issue:43526", "right": "issue:46897", "accept": false, - "reason": "BEiT image-processor label reduction bug is unrelated to Florence2 double-shift training loss." + "reason": "Reduce-labels returning one label and Florence2 double-shift training loss are different label/loss bugs." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model-loading failure in transformers>=5.1.0 is unrelated to tokenizer error masking when protobuf is absent." + "reason": "Physical-intelligence/fast model loading failure and tokenizer protobuf error masking are unrelated." }, { "left": "issue:44038", "right": "issue:46506", "accept": false, - "reason": "Qwen3-VL-Moe bug and PPDocLayoutV3 CPU/CUDA divergence are different code paths." + "reason": "Qwen3-VL-Moe transformers 5.0 bug and PPDocLayoutV3 CPU/CUDA divergence are different regressions." }, { "left": "issue:29870", "right": "issue:30333", "accept": false, - "reason": "BART generate() UnboundLocalError is unrelated to MLFlow job status reporting." + "reason": "Bart generate() UnboundLocalError and MLFlow job-state reporting are unrelated issues." }, { "left": "issue:42199", "right": "issue:46851", "accept": false, - "reason": "DETR cardinality/background-class issue is unrelated to Qwen3.5 variable-length position_ids handling." + "reason": "DETR cardinality error and Qwen3.5 variable-length position_ids are different model-specific failures." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Qwen video regression and Windows ROCm torch.distributed dependency issue are unrelated." + "reason": "Video-input regression in multimodal models and Windows ROCm torch.distributed import errors are unrelated." } ] }, @@ -13683,21 +13636,21 @@ ], "soft_pairs": [ "issue:44060|issue:44077", - "issue:44368|issue:44977", "issue:43526|issue:46894", - "issue:43653|issue:44162", + "issue:44368|issue:44977", "issue:45412|issue:46123", + "issue:43653|issue:44162", "issue:43352|issue:45250", "issue:42886|issue:45229", - "issue:43475|issue:43540", "issue:44877|issue:46170", + "issue:43475|issue:43540", "issue:46170|issue:46832", "issue:45397|issue:45656", "issue:43065|issue:43408", "issue:43723|issue:46710", "issue:44368|issue:45563", - "issue:43931|issue:44368", "issue:44995|issue:45200", + "issue:43931|issue:44368", "issue:45987|issue:46082", "issue:44188|issue:46202", "issue:43906|issue:46202", @@ -13713,20 +13666,20 @@ "issue:44749|issue:45656", "issue:43701|issue:44749" ], - "prepared_review_unit_hash": "a8749ed612a1117552b7c364bff3c6c23b5ebc6f6b2135d0fbb6a87aa0679777", + "prepared_review_unit_hash": "7328c05bccba0f61f52cc70ba10c820f17181f4b2cef34a44ac8058ce03a18d0", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36787, + "serialized_chars": 36786, "estimated_input_tokens": 9197, "estimated_eval_tokens": 18650 }, @@ -13735,192 +13688,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a collection of unrelated bug reports with a few superficial keyword overlaps (tokenizer, cache, Gemma, SAM3, Flash Attention). I do not see any pair that clearly describes the same underlying bug/change, so no soft edges should be accepted.", - "confidence": 0.87, - "canonical_issue_reason": "issue:46858 is the most concrete and code-path-specific report in the set: it names the failing method, the broken return type, and the backend/torch.compile conditions. If one issue has to anchor the cluster, this is the clearest candidate.", + "summary": "This cluster is not a true duplicate set: it spans many unrelated bugs and regressions across tokenizers, vision models, generation, config loading, cache behavior, and serialization. None of the soft pairs look like the same concrete underlying issue/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are heterogeneous and should be split by subsystem/topic rather than merged as duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46858 is the strongest standalone report because it is narrowly scoped, reproducible, and points to an exact fix area (`cache.get_seq_length()` / static cache arithmetic on non-CUDA backends).", + "best_issue_reason": "No strong global best issue exists here because the soft pairs are mostly superficial matches, not same-bug duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Different model/subsystem and different failure mode; only superficial wording overlap." - }, - { - "left": "issue:44368", - "right": "issue:44977", - "accept": false, - "reason": "One is a tied-embeddings warning, the other is a flash-attention generation bug; not the same issue." + "reason": "Different models and failure modes: tied-weights warning vs optional post_init restriction." }, { "left": "issue:43526", "right": "issue:46894", "accept": false, - "reason": "Image label reduction vs ASR training-loss shifting are unrelated code paths." + "reason": "Different tasks and bugs; one is image processor label reduction, the other ASR training loss shifting." }, { - "left": "issue:43653", - "right": "issue:44162", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are distinct bugs." + "reason": "Both mention Qwen3.5, but one is a config warning and the other is a flash-attention generation regression." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "Memory release on delete vs partial-batch pipeline bookkeeping are unrelated problems." + "reason": "Unrelated: RT-DETR memory cleanup vs MaskGenerationPipeline final-batch handling." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration bug is not the same as the ESM2 breakage." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Model-specific Flash Attention 2 support gating is not the same as the generic Flash Attention 2 topic." + "reason": "One says Flash Attention 2.0 is unsupported for NemotronH; the other is a generic Flash Attention 2.0 topic." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Tokenizer cache/offline loading and Gemma multi-GPU OOM are different failures." + "reason": "Tokenizer cache/offline loading bug is unrelated to Gemma multi-GPU OOM." }, { - "left": "issue:43475", - "right": "issue:43540", + "left": "issue:44877", + "right": "issue:46170", "accept": false, - "reason": "SAM3 vision encoder attribute error and Qwen3OmniMoe video-input ValueError are different multimodal bugs." + "reason": "Strict config loading vs Windows/ROCm import dependency issue are different problems." }, { - "left": "issue:44877", - "right": "issue:46170", + "left": "issue:43475", + "right": "issue:43540", "accept": false, - "reason": "Strict config loading and Windows/ROCm torch.distributed import issues are unrelated." + "reason": "SAM3 missing attribute vs Qwen3OmniMoe video-input ValueError are different code paths." }, { "left": "issue:46170", "right": "issue:46832", "accept": false, - "reason": "Import/dependency failure vs RT-DETR forward-shape TypeError; different code paths." + "reason": "ROCm/Windows import failure is unrelated to RT-DETR forward shape handling." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Gemma-4 zero3 loading and double optimizer-step under deepspeed are not the same bug." + "reason": "Different DeepSpeed-related symptoms; from_pretrained loading vs duplicate optimizer step." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Same SAM3 family, but dummy Conv2d placeholder and model-type warning are different issues." + "reason": "Both are SAM3-related, but one is a dummy layer and the other is a model-type mismatch warning." }, { "left": "issue:43723", "right": "issue:46710", "accept": false, - "reason": "Tokenizer load failure in v5 and accuracy regression after upgrade are different concerns." + "reason": "Tokenizer loading in v5 is not the same as an accuracy regression after version upgrade." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Tie-word-embeddings warning and paged-generate stale warning involve different mechanisms." + "reason": "Tie_word_embeddings warning is unrelated to stale num_return_sequences warning in paged generate()." }, { - "left": "issue:43931", - "right": "issue:44368", + "left": "issue:44995", + "right": "issue:45200", "accept": false, - "reason": "Weight-shape mismatch when loading Qwen3-VL is unrelated to the tied-embeddings warning." + "reason": "GlmMoeDsa stale cache crash is unrelated to Gemma4 mm_token_type_ids defaults." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:43931", + "right": "issue:44368", "accept": false, - "reason": "Stale indexer-cache crash on second forward is not the same as mm_token_type_ids defaults." + "reason": "Weight-shape mismatch in Qwen3-VL is unrelated to Qwen3.5 tie-word-embeddings warnings." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "StaticCache return-type bug and LlamaConfig head_dim validation are unrelated." + "reason": "StaticCache return-type bug and LlamaConfig head_dim validation are different issues." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "torch.compile attention-kernel divergence and Gemma pipeline roundtrip failure are different bugs." + "reason": "Torch.compile branching bug is unrelated to Gemma roundtrip tokenization failure." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of another issue is not enough to treat these as the same bug; the titles do not match a concrete shared failure." + "reason": "Insufficient evidence they share the same underlying bug; one is just an isolated reproduction." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "Chat/completions request rejection and Gemma token roundtrip failure are unrelated." + "reason": "Chat-completions request rejection is unrelated to Gemma token roundtrip issues." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Both mention cache/compile, but one is branching-induced kernel divergence and the other is lazy cache init causing recompilation; different code paths." + "reason": "Both mention cache/compile, but one is divergent attention kernels and the other is recompilation from lazy cache init." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Second-forward stale cache crash and flash_attention.py s_aux=None AttributeError are different failure sites." + "reason": "Different failures: stale indexer cache on second forward vs flash_attention.py AttributeError on s_aux." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "MPS histogram support in MoE and missing chat template in a processor are unrelated." + "reason": "MoE MPS histogram error is unrelated to missing chat template in processor." }, { "left": "issue:45103", "right": "issue:46858", "accept": false, - "reason": "Future-annotations AttributeError in auto-docstring and static-cache seq-length tensor bug are unrelated." + "reason": "Auto-docstring annotations crash is unrelated to static-cache seq-length tensor arithmetic." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "SAM3 decoder placeholder layer and mask-generation partial-batch handling are unrelated." + "reason": "SAM3 pixel-decoder dummy layer and MaskGenerationPipeline batching bug are unrelated." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "Tokenizer AttributeError on tasks and a performance regression after upgrade are different bug classes." + "reason": "Tokenizer AttributeError and slow filtering after upgrade are different problems." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Both are save_pretrained-related, but one is tokenizer metadata mismatch and the other is missing processor files; not the same concrete bug." + "reason": "Tokenizers save_pretrained mismatch is not the same as processor save_pretrained missing files." }, { "left": "issue:43994", "right": "issue:46726", "accept": false, - "reason": "SigLIP2 nonsense outputs with AutoModel/pipeline and missing-vision-backend processor lookup are different issues." + "reason": "Nonsensical outputs with AutoModel/pipeline are different from missing vision-backend error handling." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "Performance regression from upgrade and deepspeed optimizer double-step are unrelated." + "reason": "Transformer filtering slowdown is unrelated to double optimizer stepping under DeepSpeed." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint key mismatch and data-filtering slowdown are different code paths and symptoms." + "reason": "Checkpoint resume key mismatch is unrelated to the filtering performance regression." } ] }, @@ -14006,9 +13959,9 @@ "issue:44188|issue:46851", "issue:44964|issue:45137", "issue:43232|issue:43673", + "issue:39692|issue:46726", "issue:45310|issue:45375", "issue:46463|issue:46877", - "issue:39692|issue:46726", "issue:41628|issue:44933", "issue:44869|issue:45356", "issue:43986|issue:46726", @@ -14019,23 +13972,23 @@ "issue:43723|issue:46326", "issue:42371|issue:45092", "issue:43873|issue:44945", - "issue:45406|issue:46202", - "issue:43784|issue:46399" + "issue:43784|issue:46399", + "issue:45406|issue:46202" ], - "prepared_review_unit_hash": "70dcf96c186dce2bf2058b4d67bae80bc6b403bf321f8cfe9f72373fa2a08b2c", + "prepared_review_unit_hash": "d6008f2aec2093183f403d81a1c6bb3b00088d3b88868ce3b89dc1610f06c8b2", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35432, + "serialized_chars": 35430, "estimated_input_tokens": 8858, "estimated_eval_tokens": 17972 }, @@ -14044,30 +13997,30 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose mix of Transformers v5 regressions, model-loading/API breakages, and a few unrelated docs/runtime issues. Only the two Qwen3.5 Moe from_pretrained reports look like the same underlying bug.", - "confidence": 0.79, - "canonical_issue_reason": "Issue 43646 is the broadest umbrella in the set: Transformers 5.0.0 breaking custom model initialization. Several other items are narrower v5 loading/init regressions that fit under that theme.", + "summary": "This cluster is not a duplicate set; the items span unrelated model-, tokenizer-, processor-, generation-, and training-related bugs. None of the soft pairs look like the same concrete underlying defect, so I would not merge any of them.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the best representative because it captures the shared upstream regression theme most broadly, without being tied to a single model family or API surface.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Both are v5-related regressions, but one is custom model init and the other is ESM2-specific breakage; different failure modes and code paths." + "reason": "Different problems: missing fields from trust_remote_code config loading vs an ESM2 regression/broken workflow." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Processor.save_pretrained missing files is unrelated to a Deepspeed optimizer-step duplication bug." + "reason": "Processor save/export files missing is unrelated to DeepSpeed optimizer step being called twice." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id mismatch and AttentionInterface.register behavior change are distinct issues." + "reason": "Gemma eos_token_id mismatch is a model config issue; AttentionInterface.register changes callback behavior in a different subsystem." }, { "left": "issue:43441", @@ -14079,149 +14032,146 @@ "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM3 attribute failure and trust_remote_code missing-fields behavior are different bugs." + "reason": "Both mention missing attributes, but they affect different objects and code paths: Sam3 vision encoder output vs AutoConfig with trust_remote_code." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure is unrelated to Qwen2.5-GGUF loading under transformers v5." + "reason": "Saving a quantized model and loading Qwen2.5-GGUF are opposite operations with different failure modes." }, { "left": "issue:45375", "right": "issue:45478", - "accept": true, - "reason": "Same reported Qwen3.5 Moe from_pretrained error, same model family, same v5 loading failure." + "accept": false, + "reason": "Same model family, but one is a missing config field dropped by strict loading and the other is a from_pretrained error; not the same concrete bug." }, { "left": "issue:43475", "right": "issue:46821", "accept": false, - "reason": "SAM3 vision encoder attr error and TimesFM window_size AttributeError are unrelated." + "reason": "Sam3 attribute error is unrelated to TimesFM window_size AttributeError." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "PyTorch-missing ImportError and Sentence Transformers loading hang are different problems." + "reason": "Missing PyTorch import dependency is unrelated to Sentence Transformers getting stuck while loading." }, { "left": "issue:44188", "right": "issue:46851", "accept": false, - "reason": "torch.compile attention-kernel divergence is unrelated to Qwen3.5 variable-length position_ids handling." + "reason": "Torch.compile attention-kernel divergence is a different path from Qwen3.5 variable-length position_ids handling." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque error do not share the same bug." + "reason": "Model load failure for Phi-4 multimodal is unrelated to a DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Generation kwargs update after sync_gpus is not the same as missing GenerationMixin cache in chunked_prefill." + "reason": "Both are generation-related, but one is sync_gpus/update_model_kwargs and the other is missing GenerationMixin cache in chunked prefill." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "Documentation example errors are not the same as AutoProcessor backend-missing runtime handling." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Qwen3.5 Moe from_pretrained error and missing deepstack_visual_indexes in Qwen3_5MoeVisionConfig are different failures." + "reason": "Both involve Qwen3.5-Moe, but one is a generic from_pretrained error and the other is a missing deepstack_visual_indexes field silently dropped by strict config handling." }, { "left": "issue:46463", "right": "issue:46877", "accept": false, - "reason": "T5 decoder-input validation messaging and auto_docstring stdout pollution are unrelated." - }, - { - "left": "issue:39692", - "right": "issue:46726", - "accept": false, - "reason": "SigLIP2 doc example errors are unrelated to AutoProcessor backend-missing error handling." + "reason": "Incorrect T5 decoder-input validation messaging is unrelated to auto_docstring printing lint diagnostics at import time." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they involve different symbols and likely different API surfaces." + "reason": "Different import failures: AutoImageProcessor export missing vs a non-existent image_utils import." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are different tokenization issues." + "reason": "Whisper timestamp decoding crash on a trailing replacement character is unrelated to Kimi-K2.5 tokenizer codec regression." }, { "left": "issue:43986", "right": "issue:46726", "accept": false, - "reason": "Missing torchvision during AutoProcessor loading is not the same as the video offloading/quantization issue." + "reason": "Both involve AutoProcessor and vision backends, but one is missing torchvision for video models while the other is unrecognized image processor handling." }, { "left": "issue:44373", "right": "issue:46851", "accept": false, - "reason": "Wrong position_ids docstring is documentation-only; the other issue is an actual runtime behavior bug." + "reason": "A wrong docstring is not the same bug as a position_ids runtime failure." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report and gemma-4 zero3 from_pretrained failure are not the same underlying issue." + "reason": "Generic bug report has no identifiable overlap with gemma-4 ZeRO3 from_pretrained failure." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe bug and processor.save_pretrained missing files are unrelated." + "reason": "Qwen3-VL-Moe model loading is unrelated to missing files from processor.save_pretrained." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized saving failure and missing pad_token_id on Llama4Config are distinct regressions." + "reason": "Quantized model saving failure is unrelated to Llama4Config missing pad_token_id." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "Tokenizer loading in v5 and continue_final_message for mistral_common are not the same bug." + "reason": "AutoTokenizer loading in v5 and continue_final_message for mistral_common tokenizers are different tokenizer paths and failures." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + "reason": "TF32 settings guidance is unrelated to remote-code checkpoint meta-initialization incompatibility." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading and incorrect pipeline-parallel LLM output are separate runtime issues." + "reason": "Quantization/offloading behavior is a different training/runtime issue than incorrect outputs under pipeline parallelism." }, { - "left": "issue:45406", - "right": "issue:46202", + "left": "issue:43784", + "right": "issue:46399", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Gemma \u2581 roundtrip failure are different Gemma bugs with different symptoms." + "reason": "Sentence-transformers import NameError is unrelated to Qwen3.5 kernel registration failure." }, { - "left": "issue:43784", - "right": "issue:46399", + "left": "issue:45406", + "right": "issue:46202", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen3.5 kernel registration ValueError are unrelated." + "reason": "Gemma4Processor missing _tokenizer during serve is not the same as Gemma pipeline roundtrip failure on \u2581." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary stays grounded in the packet and the soft-edge calls are conservative overall. The only accepted pair, issue:45375 with issue:45478, is clearly the same Qwen3.5 Moe from_pretrained regression. The broader clustering remarks are cautious enough and do not overstate unrelated issues as the same bug." - }, - "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 }, @@ -14291,8 +14241,8 @@ "issue:44857|issue:46877", "issue:44945|issue:45753", "issue:41628|issue:44351", - "issue:45593|issue:46877", "issue:45081|issue:45103", + "issue:45593|issue:46877", "issue:36010|issue:41628", "issue:46202|issue:46701", "issue:43278|issue:45137", @@ -14320,229 +14270,229 @@ "issue:43479|issue:44869", "issue:43723|issue:44568" ], - "prepared_review_unit_hash": "b4fed583a66bd31a2613681dbbda7316640e19643b712c317c0c89bc250280e1", + "prepared_review_unit_hash": "4c314dda9b0bae2d60aa2508399ff397a815c054dd2a7111a682fbc7a9685d27", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36873, - "estimated_input_tokens": 9219, - "estimated_eval_tokens": 18694 + "serialized_chars": 36868, + "estimated_input_tokens": 9217, + "estimated_eval_tokens": 18690 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly unrelated issue reports across different bugs. The only clear duplicate pair is the Qwen3ForCausalLM VRAM leak issue; use the earlier, more discussed report as canonical.", - "confidence": 0.95, - "canonical_issue_reason": "issue:42673 is the earlier, higher-activity report for the exact Qwen3ForCausalLM multi-dataloader-thread VRAM leak, making it the best canonical issue.", + "summary": "This cluster is mostly unrelated issue reports. The only clear duplicate pair is the repeated Qwen3ForCausalLM VRAM-leak report (issue:42673 and issue:43704), which share the exact same title and underlying symptom. All other soft edges are broad-theme similarities at best, not the same concrete bug.", + "confidence": 0.97, + "canonical_issue_reason": "issue:42673 is the best canonical representative because it has the exact same title as issue:43704 and is the earlier, fuller report with more discussion.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 has the strongest evidence and discussion for the same concrete bug as issue:43704, so it is the best representative issue in this set.", + "best_issue_reason": "issue:42673 is the strongest issue to keep as the representative target because it is the earliest full report of the same Qwen3ForCausalLM VRAM-leak bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44857", "right": "issue:46877", "accept": false, - "reason": "Different bugs: AMP/CUDA crash in LwDetrImageLoss vs import-time docstring lint noise." + "reason": "Different bugs: AMP/CUDA loss crash vs auto_docstring import-time stdout logging." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Different symptoms and code paths: pipeline parallelism output corruption vs a Qwen3_5 image-specific infinite loop." + "reason": "Different failure modes: pipeline-parallel output corruption vs a specific multimodal infinite loop." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different symbols and modules; not the same underlying bug." + "reason": "Both are import errors, but for different symbols and likely different code paths." }, { - "left": "issue:45593", - "right": "issue:46877", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "Unrelated: D-FINE auxiliary-loss behavior vs auto_docstring import-time logging." + "reason": "Different subsystems: tokenizer regex patching vs auto_docstring annotation handling." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:45593", + "right": "issue:46877", "accept": false, - "reason": "Different crashes in unrelated subsystems: tokenizer regex patching vs auto_docstring with future annotations." + "reason": "Missing auxiliary losses is unrelated to docstring diagnostics printed at import." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import failures, but of different APIs; too broad to treat as one bug." + "reason": "Unrelated import problems: GenerationMixin vs AutoImageProcessor." }, { "left": "issue:46202", "right": "issue:46701", "accept": false, - "reason": "Both mention Gemma, but one is tokenizer roundtrip and the other is an image-text pipeline failure." + "reason": "Both mention Gemma, but the concrete failures are different: token roundtrip vs pipeline integration." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Training/eval dtype drift is unrelated to the ZeRO3 deque IndexError." + "reason": "Different bugs: dtype drift in evaluation vs DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Related model family, but different models and different version-specific from_pretrained failures." + "reason": "Related model family, but different versions/models and no evidence of the same concrete failure path." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact duplicate title and same Qwen3ForCausalLM VRAM leak in multiple dataloader threads." + "reason": "Exact same title and same VRAM-leak symptom; clear duplicate reports of the same bug." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving failure is unrelated to importing a missing class." + "reason": "Quantized save failure vs missing exported class import are unrelated." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Both involve Qwen3.5 MoE/FP8, but one is save_pretrained while the other is from_pretrained loading; not the same bug." + "reason": "Different operations and likely different fixes: FP8 save_pretrained vs from_pretrained load error." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FP8 accuracy regression is different from FP8 MoE save_pretrained failure." + "reason": "Both mention FP8, but one is an accuracy regression and the other is a save/pretrained bug." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwargs/sync_gpus bug is unrelated to a janus image-generation test failure." + "reason": "A generation-kwargs bug and a failed Janus test are not the same concrete issue." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Different areas: model preset loading vs pytest-xdist file race in testing utilities." + "reason": "Pretrained preset loading vs xdist race in testing utilities are unrelated." }, { "left": "issue:45561", "right": "issue:45600", "accept": false, - "reason": "Test-worker file race is unrelated to stale auto_mappings references." + "reason": "A test-runner race on captured_info.txt is unrelated to stale auto_mappings references." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Model config recognition failure is unrelated to captured_info.txt race conditions." + "reason": "Model config loading error vs testing-utils file race; no shared code path." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Package-install backend omission is unrelated to a missing image_utils import." + "reason": "Package/backend dependency issue vs missing image_utils import; unrelated." }, { "left": "issue:45405", "right": "issue:46877", "accept": false, - "reason": "Release/version pinning problem is unrelated to docstring lint output." + "reason": "Release/version pinning issue vs docstring logging problem; unrelated." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed training jobs is unrelated to a DeepSpeed OOM." + "reason": "MLFlow reporting failure and DeepSpeed OOM are different training issues." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version compatibility issue is unrelated to a missing import from image_utils." + "reason": "Python-version incompatibility vs missing import from image_utils; not the same bug." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Loading hang and token-embedding reinit are different model lifecycle bugs." + "reason": "Loading hang and embedding reinitialization are unrelated failures." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different behaviors and are not the same concrete issue." + "reason": "Both are tokenizer regressions, but they affect different behaviors and models." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Backend extra installation issue is unrelated to leftover dummy classes and repo-check failures." + "reason": "Different dependency/packaging problems: missing TensorFlow backend vs dummy classes leaking into namespace." }, { "left": "issue:43479", "right": "issue:46829", "accept": false, - "reason": "Multimodal config defaults and video-classification output wiring are different bugs." + "reason": "Different model/config bugs with different outputs and symptoms." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing processor save files is unrelated to Qwen3.5 MoE loading errors." + "reason": "Processor save_pretrained missing files is a different path than Qwen3.5 MoE from_pretrained loading." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Attention-mask construction bug is unrelated to the partial-batch pipeline state bug." + "reason": "Both touch generation, but one is attention-mask construction and the other is batch result dropping." }, { "left": "issue:42673", "right": "issue:46693", "accept": false, - "reason": "Both are GPU/memory related, but one is a VRAM leak in dataloader threads and the other is a long-context FA2 performance regression." + "reason": "Both involve GPU performance, but the concrete bugs differ: VRAM leak in threads vs FA2 long-context regression." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Different tokenizer regressions: cleanup-space behavior vs wrong token IDs for Granite models." + "reason": "Both are tokenizer regressions, but they affect different APIs and models." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Config initialization bug is unrelated to Whisper word timestamp decode crashing on a trailing replacement character." + "reason": "Different model-family issues: multimodal config defaults vs Whisper timestamp decoding crash." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Both involve tokenizer changes in v5, but one is tokenizer loading and the other is special-token insertion behavior." + "reason": "Both are v5 tokenizer issues, but one is tokenizer loading and the other is add_special_tokens behavior." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet: the two Qwen3ForCausalLM issues have identical titles, the later one has less activity, and the canonical choice is conservative. The soft-edge verdicts also stay cautious and are supported by the issue titles without overclaiming." + "accept": false, + "feedback": "The main duplicate call is plausible, but the canonicalization rationale overstates the packet: issue:42673 and issue:43704 do have the same title and the earlier one is fuller/more discussed, but they do not show more inbound references (both are 0 in the packet). For a precision review, that unsupported detail is enough to reject." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -14617,38 +14567,38 @@ "issue:43479|issue:44079", "issue:43784|issue:43824", "issue:44734|issue:45103", - "issue:43874|issue:45092", "issue:42222|issue:45200", + "issue:43874|issue:45092", "issue:27957|issue:30333", "issue:42898|issue:43295", "issue:30333|issue:34689", "issue:42898|issue:45356", "issue:44295|issue:45656", "issue:30333|issue:36331", - "issue:43116|issue:44190", "issue:45405|issue:45712", + "issue:43116|issue:44190", "issue:22355|issue:28282", "issue:44222|issue:44297", "issue:44075|issue:46231", "issue:44060|issue:45125", - "issue:44448|issue:46132", "issue:44933|issue:45712", + "issue:44448|issue:46132", "issue:42796|issue:43723", "issue:42831|issue:44623", "issue:44297|issue:45656", "issue:43479|issue:43644", - "issue:44360|issue:46762", + "issue:44222|issue:44623", "issue:45245|issue:46132", - "issue:44222|issue:44623" + "issue:44360|issue:46762" ], - "prepared_review_unit_hash": "1a0925d0896f7d125092184c8eaeca17fad2da98a30b1a5a5a8fb946ae1cf4e8", + "prepared_review_unit_hash": "9b647d0f6fd521e66f092127df6f59a61784610b50dc9e47078ad30602798331", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, @@ -14663,192 +14613,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of unrelated Transformers issues spanning tokenizers, multimodal models, training, docs, and runtime regressions. None of the soft-linked pairs look like true duplicates or the same fix target, so no mergeable canonical item emerges.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "These are mostly unrelated issues with superficial title similarity. I did not find any soft pair that clearly describes the same concrete bug/change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 44779 is the most representative user-facing regression in the set: a concrete tokenizer breakage in v5 with clear reproduction impact and multiple references. It is broader and more substantive than the mostly narrow doc/import/config issues around it.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 44779 is the best overall issue candidate because it is a concrete, high-signal regression with broader applicability than the other items, which are mostly model-specific, documentation-only, or isolated API/import errors.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44625", "right": "issue:46877", "accept": false, - "reason": "Different problems: config propagation for Qwen3.5 vs auto_docstring printing diagnostics at import time." + "reason": "Different problems: Qwen3.5 num_labels propagation vs auto_docstring logging diagnostics to stdout." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and a documentation parameter-name mismatch are separate docs issues." + "reason": "A notebook syntax error and a docs parameter-name mismatch are unrelated." }, { "left": "issue:39692", "right": "issue:46909", "accept": false, - "reason": "Different example failures for different models (SigLIP2 vs DIA) with different symptoms." + "reason": "Both are broken examples, but they affect different docs and different model flows; not the same bug." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both tokenizer-related, but one is a v5 tokenization-space behavior change and the other is a Deepseek tokenizer regression." + "reason": "Both involve tokenizer regressions in v5, but one is a specific clean_up_tokenization_spaces behavior change and the other is Deepseek output corruption; too indirect to treat as the same bug." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "One is multimodal config initialization; the other is ModelOutput key assignment logic." + "reason": "Different code paths: multimodal config initialization vs ModelOutput key assignment." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Different import errors with different causes and packages/models." + "reason": "Both are import errors, but they concern different missing symbols and different packages/models." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Server KV-cache indexing crash is unrelated to auto_docstring\u2019s annotation-handling AttributeError." + "reason": "KV-cache indexing in serve is unrelated to the auto_docstring annotations crash." }, { - "left": "issue:43874", - "right": "issue:45092", + "left": "issue:42222", + "right": "issue:45200", "accept": false, - "reason": "Both concern multimodal models, but the code paths differ: missing processor method vs meta-init incompatibility." + "reason": "Different multimodal issues affecting different models and different inputs." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:43874", + "right": "issue:45092", "accept": false, - "reason": "Different model families and distinct bugs: broken VitPose models vs Gemma 4 mm_token_type_ids defaulting." + "reason": "Both touch multimodal model loading, but one is a missing image-processor method and the other is remote-code/meta-init incompatibility; not the same fix." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta is unrelated to MLFlow reporting after failed training jobs." + "reason": "Flash Attention model support and MLflow reporting on training failure are unrelated." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "Both are v5 regressions, but one is tokenization cleanup behavior and the other is processor/tokenizer image-passing behavior." + "reason": "Different regressions: tokenizer cleanup behavior vs processor/tokenizer image-passing compatibility." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLFlow job-status reporting and Llama 3.2 model loading are unrelated issues." + "reason": "Failed MLflow reporting and Llama model loading regression are unrelated." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Different tokenizer regressions affecting different models and code paths." + "reason": "Both are tokenizer regressions, but they appear model/version specific and not the same code-path problem." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Position_ids buffer access bug is unrelated to Deepspeed causing optimizer.step() to run twice." + "reason": "Reading a registered buffer and duplicate optimizer stepping under deepspeed are unrelated." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "Failed-job MLFlow reporting and CustomTrainer compute_loss signature incompatibility are separate training issues." + "reason": "MLflow failure reporting and CustomTrainer signature mismatch are different bugs." }, { - "left": "issue:43116", - "right": "issue:44190", + "left": "issue:45405", + "right": "issue:45712", "accept": false, - "reason": "Different example scripts and failures: empty multi-label outputs vs local dataset loading." + "reason": "A dependency version bump issue and leftover dummy classes are unrelated." }, { - "left": "issue:45405", - "right": "issue:45712", + "left": "issue:43116", + "right": "issue:44190", "accept": false, - "reason": "PEFT version pinning and leftover dummy PT objects are unrelated maintenance issues." + "reason": "Both involve example scripts, but they fail in different scripts for different reasons." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Both are import/dependency errors, but one is missing transformers.onnx while the other is missing PyTorch for AutoModel." + "reason": "Different missing-dependency import errors: transformers.onnx vs torch for AutoModel." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "Both involve save_pretrained, but one is FP8/MoE serialization and the other is tokenizer metadata mismatch." + "reason": "FP8 save_pretrained for MoE and tokenizer_class metadata mismatch are unrelated." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "SGD arguments not being used is an implementation bug; the other is a docs parameter-name typo." + "reason": "Unused SGD args is a runtime optimizer issue, not a documentation typo." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Different Qwen3 issues: incorrect tied-weights warning vs missing tensor-parallel plan." + "reason": "Both are Qwen-family issues, but one is a tied-weights warning and the other is a missing tensor-parallel plan." }, { - "left": "issue:44448", - "right": "issue:46132", + "left": "issue:44933", + "right": "issue:45712", "accept": false, - "reason": "Pegasus v4/v5 output differences are unrelated to AttentionInterface.register behavior changes." + "reason": "Importing a nonexistent image_utils symbol and dummy-class leakage are different maintenance issues." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:44448", + "right": "issue:46132", "accept": false, - "reason": "Nonexistent image_utils import and leftover dummy classes are different cleanup problems." + "reason": "Pegasus v4/v5 output drift and AttentionInterface.register behavior change are not the same bug." }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "Both are tokenizer issues, but one is BPE-marker handling and the other is AutoTokenizer loading." + "reason": "A model tokenizer BPE-marker regression and AutoTokenizer loading failure are different tokenizer problems." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FineGrainedFP8 accuracy and processor.save_pretrained missing files are unrelated." + "reason": "FP8 accuracy degradation and processor.save_pretrained missing files are unrelated." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to Deepspeed optimizer double-stepping." + "reason": "Tokenizer save metadata and deepspeed optimizer-step duplication are unrelated." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Config initialization for Phi4Multimodal is unrelated to non-persistent buffer corruption in v5." + "reason": "Default config initialization and non-persistent buffer corruption are different subsystems." }, { - "left": "issue:44360", - "right": "issue:46762", + "left": "issue:44222", + "right": "issue:44623", "accept": false, - "reason": "DSA indexer/ReLU discussion and Minimax attention numerical drift are not the same bug." + "reason": "MoE FP8 save behavior and processor serialization missing files are unrelated." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-cardinality runtime error is unrelated to AttentionInterface.register semantics." + "reason": "Category cardinality error and AttentionInterface.register behavior change are unrelated." }, { - "left": "issue:44222", - "right": "issue:44623", + "left": "issue:44360", + "right": "issue:46762", "accept": false, - "reason": "Both are save_pretrained-related, but they affect different objects and failure modes." + "reason": "A missing ReLU in one implementation and numerical differences in a different attention backend are not the same concrete bug." } ] }, @@ -14883,7 +14833,6 @@ "issue:43122", "issue:43232", "issue:43475", - "issue:43644", "issue:43723", "issue:44190", "issue:44222", @@ -14923,11 +14872,11 @@ ], "soft_pairs": [ "issue:46132|issue:46710", - "issue:41628|issue:42757", "issue:44373|issue:44485", + "issue:41628|issue:42757", "issue:44297|issue:45310", - "issue:45071|issue:46082", "issue:43723|issue:44977", + "issue:45071|issue:46082", "issue:46614|issue:46932", "issue:44222|issue:44484", "issue:44190|issue:46082", @@ -14950,217 +14899,210 @@ "issue:44279|issue:46328", "issue:44483|issue:45406", "issue:22355|issue:27957", - "issue:44261|issue:45092", - "issue:43644|issue:46305" + "issue:44261|issue:45092" ], - "prepared_review_unit_hash": "f242296e72c5a434914eafeb21446d98166770653243d000702b373259c4ffae", + "prepared_review_unit_hash": "13a97349b6146ba78868b4f27d84fc6198fd5924c093363fefc317816686cd91", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36216, - "estimated_input_tokens": 9054, - "estimated_eval_tokens": 18364 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35353, + "estimated_input_tokens": 8839, + "estimated_eval_tokens": 17934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a noisy mixed cluster of mostly unrelated Transformers issues; none of the soft pairs look like the same underlying bug or change, so all soft edges are rejected.", - "confidence": 0.94, - "canonical_issue_reason": "Issue #44263 is the most concrete, code-path-specific bug in the set and the best representative of the small DSA/indexer subtheme.", + "summary": "The cluster is largely heterogeneous; the soft pairs share only broad topical overlap, not the same concrete bug or change. I rejected all soft edges because none look mergeable as duplicates. I picked 45092 as the closest issue representative of the loose v5-compatibility subset, but there is no strong canonical for the full cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single strong canonical: the set is heterogeneous. If one must be chosen, 45092 is the closest representative because it is a concrete v5 regression with a clear compatibility failure.", "canonical_pr_reason": null, - "best_issue_reason": "#44263 has a narrow repro surface and implementation-level scope, making it the clearest issue to anchor this otherwise heterogeneous cluster.", + "best_issue_reason": "45092 is the best representative for the same reason: it is specific, regression-oriented, and more likely to anchor related compatibility reports than the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46132", "right": "issue:46710", "accept": false, - "reason": "Different problems: attention registration behavior change vs accuracy regression after a version upgrade." + "reason": "Different bugs: attention registration behavior vs an accuracy regression in a specific model." }, { - "left": "issue:41628", - "right": "issue:42757", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Unrelated import failures from different packages/modules." + "reason": "Docstring wording issue vs RoPE implementation bug; not the same code-path failure." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:41628", + "right": "issue:42757", "accept": false, - "reason": "A docstring problem is not the same as a RoPE implementation bug." + "reason": "Separate import errors from different packages/modules and different root causes." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer metadata mismatch in save_pretrained is distinct from a Qwen3.5 from_pretrained failure." + "reason": "Tokenizer save metadata mismatch vs a Qwen3.5 from_pretrained load error; related area but distinct defect." }, { - "left": "issue:45071", - "right": "issue:46082", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "Generic PreTrainedConfig type-checking breakage is not the same as LlamaConfig head_dim validation." + "reason": "Tokenizer loading in v5 vs flash-attention generation for Qwen3.5; unrelated failures." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:45071", + "right": "issue:46082", "accept": false, - "reason": "Tokenizer loading regression and flash-attention generation failure are different code paths." + "reason": "Both are config-related, but one is type-checking regression and the other is LlamaConfig head_dim validation." }, { "left": "issue:46614", "right": "issue:46932", "accept": false, - "reason": "Stale phonemizer backend state is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "Phonemizer backend state bug vs Gemma4 quantized embedder dtype casting; different components." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "FP8 save_pretrained failure is not the same as a question about max_shard_size defaults." + "reason": "FP8 save_pretrained/moe issue vs a max_shard_size default question; no shared bug." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading script issue is unrelated to a config validation error." + "reason": "Dataset loading script failure vs model config validation; unrelated code paths." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenizer behavior regression and FP8 inference failure on B200 are different bugs." + "reason": "Tokenization drift across versions vs FP8 inference on B200; not the same underlying problem." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "DSA indexer split behavior and remote-code/meta-init incompatibility do not share the same failure mode." + "reason": "DSA indexer split behavior vs meta-initialization incompatibility for old checkpoints; distinct issues." }, { "left": "issue:43232", "right": "issue:46424", "accept": false, - "reason": "Generation kwargs update bug and cache-length handling are separate mechanisms." + "reason": "Generation kwargs update bug vs cache max length semantics; same broad area, different failure." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation state update bug and OCR long-input index error are not the same underlying issue." + "reason": "Generation-state update bug vs a long-OCR index error in Kosmos2.5; unrelated." }, { "left": "issue:43475", "right": "issue:46305", "accept": false, - "reason": "Missing attribute on SAM3 video output is unrelated to a sliding-window off-by-one." + "reason": "SAM3 video encoder output attribute issue vs MoonshineStreaming mask off-by-one; different models and logic." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Loading hang and zero-probability outputs are different symptoms with different likely causes." + "reason": "Loading hang vs zero-probability inference bug; not the same defect." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "FP8 inference on B200 and missing chat template are unrelated." + "reason": "FP8 B200 inference failure vs missing chat template error; unrelated." }, { "left": "issue:46305", "right": "issue:46655", "accept": false, - "reason": "Both involve video processing, but the concrete bugs differ: mask off-by-one vs padding calculation." + "reason": "Both are video-related indexing/padding bugs, but in different processors and different computations." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Buffer access error and missing processor save files are different save/load bugs." + "reason": "Buffer access error vs processor.save_pretrained missing files; different save/load paths." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Same subsystem, but different concrete bugs: split return handling vs missing ReLU." + "reason": "Same component family, but one is a torch.split issue and the other is a missing ReLU; not the same bug." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "Different bug from a docstring issue." + "reason": "Indexer return-value bug vs a wrong docstring; no shared failure mode." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Dequantized save failure and broken quantized round-trip are related only broadly, not the same code-path problem." + "reason": "Different quantization/save issues affecting different flows and serialization behaviors." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Different model families and different failure surfaces." + "reason": "Both touch multimodal checkpoint compatibility, but one is meta-init loading and the other is incorrect saved visual encoder keys." }, { "left": "issue:46369", "right": "issue:46909", "accept": false, - "reason": "Model memory overflow is not the same as a broken documentation example." + "reason": "Model memory overflow vs documentation example bug; unrelated." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy regression and Qwen3.5 from_pretrained error are not the same bug." + "reason": "FP8 accuracy regression vs from_pretrained load error; distinct issues despite both being model-related." }, { "left": "issue:45381", "right": "issue:46305", "accept": false, - "reason": "Different model family and different failure mode; not the same underlying issue." + "reason": "Both mention off-by-one style behavior, but they are different models, tensors, and code paths." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "A dependency issue is unrelated to a SequenceFeatureExtractor.pad performance bug." + "reason": "Dependency problem vs pad-performance inefficiency; not the same bug." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "Server request handling failure and Gemma4Processor missing _tokenizer are distinct issues." + "reason": "Chat-completions request handling bug vs Gemma4Processor attribute crash; different API failures." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Missing transformers.onnx module is unrelated to XLMRoberta Flash Attention 2 support." + "reason": "ONNX import error vs Flash Attention 2 support question; unrelated." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "MLA precision error due to rms_norm_eps is unrelated to meta-initialization incompatibility." - }, - { - "left": "issue:43644", - "right": "issue:46305", - "accept": false, - "reason": "Non-persistent buffer junk and a streaming mask off-by-one are different bugs." + "reason": "RMS norm epsilon precision bug vs old checkpoint meta-init incompatibility; different model behavior." } ] }, @@ -15204,11 +15146,11 @@ "issue:43408", "issue:43540", "issue:43576", + "issue:43644", "issue:43646", "issue:43756", "issue:43784", "issue:43824", - "issue:43937", "issue:43976", "issue:44038", "issue:44060", @@ -15227,14 +15169,14 @@ "issue:45593", "issue:45753", "issue:46002", - "issue:46172", + "issue:46305", "issue:46369", "issue:46402", - "issue:46762", "issue:46909" ], "soft_pairs": [ "issue:44829|issue:46002", + "issue:43644|issue:46305", "issue:42831|issue:44038", "issue:41628|issue:43824", "issue:43408|issue:46402", @@ -15258,230 +15200,219 @@ "issue:43976|issue:44945", "issue:39401|issue:44488", "issue:42371|issue:44263", - "issue:43756|issue:44060", "issue:44295|issue:44297", + "issue:43756|issue:44060", "issue:43576|issue:43784", - "issue:41628|issue:45278", - "issue:46172|issue:46762", - "issue:42907|issue:43937" + "issue:41628|issue:45278" ], - "prepared_review_unit_hash": "d153a3c7bb0ca671aa1873489cac63897c8bd7cedec88a304457cc634ce4d728", + "prepared_review_unit_hash": "ecacb1bece2382ab78f15ec9a4f13bb73fc2f04fc79af40c0560682b12ea2f3a", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36061, - "estimated_input_tokens": 9016, - "estimated_eval_tokens": 18288 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35259, + "estimated_input_tokens": 8815, + "estimated_eval_tokens": 17886 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is still mostly heterogeneous, and the packet does not provide explicit duplicate evidence. The closest pair, issue:44829 and issue:46002, looks only tentatively similar at the symptom level, so it should stay unmerged absent stronger evidence.", - "confidence": 0.61, - "canonical_issue_reason": null, + "summary": "The soft pairs are mostly broad-topic similarities rather than true duplicates; I would reject all of them. If one issue must serve as the cluster representative, the broad version-upgrade import regression is the best umbrella candidate.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the most umbrella-like report in this set: it describes many import errors after a transformers upgrade, making it the best representative for a heterogeneous import-regression cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46002 is the closest representative for the NaNs/classification-head symptom family because it is the more concrete report, but the evidence only supports tentative similarity rather than a confirmed duplicate.", + "best_issue_reason": "issue:45278 is the strongest global issue candidate because it is broad, clearly user-facing, and closer to a cluster-level regression than the more model-specific or one-off bug reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Tentatively similar only: both mention classification-head training instability/NaN-like behavior, but the packet gives no explicit duplicate evidence and the failure contexts differ enough to avoid merging." + "reason": "Both are training/initialization bugs, but they describe different failure modes: Flash Attention 3 degenerate training vs NaNs after checkout+init." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "One is a v5 non-persistent-buffer corruption issue; the other is a MoonshineStreaming sliding-window off-by-one. Different code paths and symptoms." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "Both are model-specific regressions, but they describe different bugs and different code paths." + "reason": "FineGrainedFP8 accuracy problems and a Qwen3-VL-Moe bug are unrelated beyond both involving model quality/regression concerns." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but for different symbols/models and likely different root causes." + "reason": "Both are import errors, but for different missing symbols and likely different root causes; not the same concrete bug." }, { "left": "issue:43408", "right": "issue:46402", "accept": false, - "reason": "Same broad SAM3 area, but the warning vs missing-weights failures are different concrete problems." + "reason": "Different model families and different problems: a wrong model-type warning vs missing Sam3 text encoder weights." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Different issues: ONNX import failure vs tokenizer spacing behavior." + "reason": "A missing module import and a tokenizer extra-space bug are unrelated." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "Different model bugs: FP8 accuracy concerns vs an image-specific infinite loop." + "reason": "Accuracy regression in FP8 is not the same as an infinite loop on a specific image." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer-related, but the symptoms and models differ." + "reason": "Both involve tokenization behavior, but one is offset mapping for Qwen3 and the other is BOS/EOS insertion for mDeBERTa; different bugs." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Offline cache loading and quantized-model saving are different code-path failures." + "reason": "Offline cache loading failure and quantized model saving failure are different lifecycle stages and code paths." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Tokenizer/loading related only at a broad level; the concrete failures differ." + "reason": "Offline cache loading and wrong tokenizer decoder type are unrelated issues." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Different model/runtime failures with different symptoms." + "reason": "Flash Attention 2 support for XLMRoberta and a Sentence Transformers loading hang are different problems." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Attention-kernel divergence under torch.compile is unrelated to a position_ids docstring issue." + "reason": "A torch.compile attention-kernel divergence bug is unrelated to a docstring issue." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "Both involve imports after an upgrade, but they are too broad and not the same concrete failure." + "reason": "Both mention import breakage, but one is a specific huggingface_hub symbol and the other is a broad transformers upgrade regression; not enough to call them the same bug." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "Same family (SAM3) but different failures: model-type warning vs memory overflow." + "reason": "Model-type warning and memory overflow are different failure modes." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation-state handling and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:42796", "right": "issue:43122", "accept": false, - "reason": "Both are tokenization regressions, but the target models and symptoms differ." + "reason": "Tokenizer BPE-marker handling for one model and a general tokenization difference across versions are not clearly the same bug." }, { "left": "issue:45381", "right": "issue:46909", "accept": false, - "reason": "Different domains entirely: video position IDs vs a documentation example bug." + "reason": "A Qwen2.5-VL video-position-id bug and a broken documentation example for DIA are unrelated." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage is not the same bug as NaNs in classification heads." + "reason": "Custom model initialization breakage and NaNs in classification heads both touch init, but they are not the same concrete failure." }, { "left": "issue:43540", "right": "issue:46909", "accept": false, - "reason": "Video-processing ValueError is unrelated to a documentation example bug." + "reason": "Video input ValueError in Qwen3OmniMoe is unrelated to a DIA docs example." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Different models/platforms with different runtime symptoms." + "reason": "Loading hang for Sentence Transformers and a MacOS bus error with a CLIP model are distinct issues." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Model-saving failure and pipeline-parallel output divergence are different problems." + "reason": "Quantized model saving failure and incorrect pipeline-parallel outputs are different bugs." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving/dequantization issues are unrelated to a multimodal model load failure." + "reason": "Quantized model saving failure and inability to load Phi-4 multimodal are unrelated." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python-version incompatibility is unrelated to pipeline-parallel correctness." + "reason": "Python version compatibility regression and pipeline parallelism output corruption are different problems." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer/loading-related, but one is wrong offsets and the other is a load failure." + "reason": "Tokenizer offset mapping for Qwen3 and failure to load a specific model are not the same underlying bug." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 API guidance and torch.split return behavior are unrelated." + "reason": "TF32 configuration guidance and torch.split return-value handling are unrelated." }, { - "left": "issue:43756", - "right": "issue:44060", + "left": "issue:44295", + "right": "issue:44297", "accept": false, - "reason": "Different model-specific architecture bugs." + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are different tokenizer/model serialization bugs." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "A registered-buffer read issue and tokenizer metadata mismatch are not the same bug." + "reason": "RoPE layer dropping in Smollm3 and a tied-weights warning in Qwen3-Next are different model-definition issues." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "Environment command regression and an import NameError are separate issues." + "reason": "A broken CLI command and a NameError during import are unrelated." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "The latter is broader, but this is not the same concrete import failure." - }, - { - "left": "issue:46172", - "right": "issue:46762", - "accept": false, - "reason": "Different numerical/transpose behavior in unrelated model families." - }, - { - "left": "issue:42907", - "right": "issue:43937", - "accept": false, - "reason": "Model-saving failure is unrelated to invalid GenerationConfig validation." + "reason": "Both concern imports after upgrades, but the specific missing symbol and the broad multi-import regression are too different to merge." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The summary is too confident relative to the packet. There is no explicit duplicate evidence, and all soft pairs are marked deterministic_accept=false. In particular, treating issue:44829 and issue:46002 as the \"same underlying bug\" and calling issue:46002 the \"best representative\" because of the \"strongest duplicate evidence\" overstates what the packet supports. Please soften that verdict to a tentative similarity at most, or leave it unmerged unless there is stronger evidence." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -15503,7 +15434,6 @@ "issue:29127", "issue:29159", "issue:36296", - "issue:38617", "issue:39401", "issue:42175", "issue:42757", @@ -15537,18 +15467,21 @@ "issue:44945", "issue:45200", "issue:45216", - "issue:45278", "issue:45290", "issue:45478", "issue:45600", "issue:45753", + "issue:46172", "issue:46231", "issue:46326", "issue:46328", "issue:46614", + "issue:46762", "issue:46877" ], "soft_pairs": [ + "issue:46172|issue:46762", + "issue:42907|issue:43937", "issue:43784|issue:45600", "issue:39401|issue:44521", "issue:44117|issue:44704", @@ -15574,220 +15507,218 @@ "issue:22355|issue:29127", "issue:42175|issue:44938", "issue:43493|issue:45753", - "issue:46231|issue:46877", "issue:44188|issue:46328", - "issue:42898|issue:46614", - "issue:38617|issue:45278", - "issue:43937|issue:44485" + "issue:46231|issue:46877", + "issue:42898|issue:46614" ], - "prepared_review_unit_hash": "3c678673e588c0b28ee5991ccdf865831b8f0acc1150962482aaed9acda93a3b", + "prepared_review_unit_hash": "82ef5647af3de684fed8e06aad34ac96dbb2d5b70f5a66484f8eedb5350b5f28", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36465, - "estimated_input_tokens": 9117, - "estimated_eval_tokens": 18490 + "serialized_chars": 36475, + "estimated_input_tokens": 9119, + "estimated_eval_tokens": 18494 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is not a true duplicate set: the items span unrelated import errors, tokenizer regressions, multimodal chat-template bugs, model-specific failures, training issues, and docs/perf reports. All soft edges should be rejected.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue fits this cluster; the apparent matches are mostly lexical, not the same underlying bug.", + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly share only broad subsystem terms (tokenizers, save/load, multimodal, import errors) but not the same concrete bug. I found no duplicate-worthy pair and no PRs.", + "confidence": 0.92, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global canonical. 45278 is the broadest import-regression report, but it does not subsume the many unrelated tokenizer/model/training bugs here.", + "best_issue_reason": "issue:44521 is the strongest standalone issue to anchor on operationally: it is open, has multiple inbound references, and describes a concrete user-facing regression in multimodal chat templating. It is not, however, a duplicate target for the other items here.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "Different model families and different symptoms: deepseek-v4 tensor conversion bug vs Minimax MSA numerical drift." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Unrelated failures: save_pretrained for dequantized models vs invalid GenerationConfig for GLM-5." + }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Different failures: sentence-transformers import error vs removed SAM3 Lite config references breaking CI." + "reason": "Importing sentence-transformers breaks due to missing nn, while the other is a CI break from removed Sam3LiteText config refs." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset_mapping bug is unrelated to multimodal assistant_masks being all zeros." + "reason": "Different code paths: tokenizer offset mapping for Qwen3 vs assistant_masks generation in multimodal chat templates." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES/from_pretrained logic is unrelated to AutoProcessor kwargs forwarding." + "reason": "TOKENIZER_MAPPING_NAMES None handling is a tokenizer registry bug; AutoProcessor kwargs forwarding is a cached_file plumbing issue." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Qwen3.5 num_labels propagation and flash_attention_3 degenerate training are different code paths." + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are unrelated bugs." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Both are SAM3-related, but one is a model-type warning and the other is a missing attribute in vision output." + "reason": "Both are SAM 3 related, but one is a model-type warning and the other is a missing output attribute; different failures." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "AutoProcessor cached_file kwargs bug is unrelated to apply_chat_template crashing on tool-call messages." + "reason": "AutoProcessor kwargs forwarding and apply_chat_template tool-call crash affect different paths and would not merge into one fix." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "General v4-to-v5 tokenization differences are not the same as multimodal assistant mask generation." + "reason": "Tokenizer-version regression is not the same bug as multimodal assistant_mask generation." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer/ReLU discussion is unrelated to GLM-5 RoPE implementation." + "reason": "DSA indexer activation discussion and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:43122", "right": "issue:46326", "accept": false, - "reason": "Both mention tokenization, but one is a broad v4/v5 regression and the other is a mistral_common continue_final_message bug." + "reason": "Both involve tokenizer behavior, but they describe different regressions in different APIs and code paths." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is unrelated to gguf_file breaking AutoTokenizer/AutoModel loading." + "reason": "Save/pretrained failure for dequantized models is unrelated to gguf_file load issues in AutoTokenizer/AutoModel." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe bug and tokenizer.save_pretrained config mismatch are different issues." + "reason": "Qwen3-VL-Moe bug and tokenizer_class metadata mismatch are different problems." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Missing transformers.onnx module is unrelated to slow/fast tokenizer inconsistency." + "reason": "Missing transformers.onnx module import and slow/fast tokenizer inconsistency are unrelated." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5, but save_pretrained checkpoint corruption and from_pretrained failure are not the same bug." + "reason": "Both mention Qwen3.5, but one is a bad saved checkpoint while the other is a from_pretrained import error in newer transformers." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Qwen3-VL-Moe regression is unrelated to position_ids buffer access errors." + "reason": "Qwen3-VL-Moe bug is unrelated to the position_ids buffer access error." }, { "left": "issue:39401", "right": "issue:46614", "accept": false, - "reason": "Qwen3 offset_mapping and Wav2Vec2 phonemizer backend caching are unrelated tokenizer/audio bugs." + "reason": "Qwen3 tokenizer offset mapping and stale phonemizer language state are different tokenizer subsystems with different root causes." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Qwen3-VL-Moe import/runtime issue is unrelated to FP8 save_pretrained behavior." + "reason": "Qwen3-VL-Moe issue is not the same as FP8 save_pretrained for MoE models." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different root causes." + "reason": "Different missing exports: Qwen2_5_VLForConditionalGeneration vs HybridCache." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example-script empty multi-label results and mm_token_type_ids defaulting are different problems." + "reason": "Example script multi-label inference returning empty results is not the same bug as mm_token_type_ids defaults for Gemma 4 fine-tuning." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but not the same tokenizer/model or code path." + "reason": "General tokenizer regression and Deepseek-specific incorrect tokenization are not clearly the same underlying defect." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "BF16-to-FP32 eval dtype change is unrelated to incorrect pipeline-parallel LLM output." + "reason": "Embedding dtype drift between train/eval is unrelated to incorrect pipeline-parallel LLM outputs." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bug and classification example empty outputs are unrelated." + "reason": "Tensor parallel training bug and example-script classification emptiness are different failure modes." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "huggingface_hub missing is_offline_mode and transformers missing HybridCache are separate import regressions." + "reason": "Both are import errors, but they refer to different symbols from different packages and different breakages." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "onnx import failure and LayoutLMv3 error-message clarity are unrelated." + "reason": "Transformers.onnx import failure and LayoutLMv3 error messaging are unrelated." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Missing TensorFlow extra in pip install and Python 3.14 load failure are different environment issues." + "reason": "Package extras missing tensorflow_backend and Python 3.14 import/load failure are unrelated environment issues." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "SigLIP2 implementation discrepancy and Qwen3_5 image infinite loop are different multimodal bugs." - }, - { - "left": "issue:46231", - "right": "issue:46877", - "accept": false, - "reason": "Docs parameter-name mismatch is unrelated to import-time docstring lint output." + "reason": "SigLIP2 numerical discrepancy and Qwen3_5 image-triggered infinite loop are different model-specific bugs." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "torch.compile attention-kernel divergence and SequenceFeatureExtractor.pad performance issue are unrelated." - }, - { - "left": "issue:42898", - "right": "issue:46614", - "accept": false, - "reason": "v5 tokenization-space cleanup behavior and stale phonemizer language caching are unrelated." + "reason": "torch.compile attention-kernel divergence and SequenceFeatureExtractor.pad performance regression are unrelated." }, { - "left": "issue:38617", - "right": "issue:45278", + "left": "issue:46231", + "right": "issue:46877", "accept": false, - "reason": "Specific layer_type_validation import error and broad import failures after upgrade are not the same concrete bug." + "reason": "Documentation mismatch in set_training and auto_docstring stdout diagnostics are separate doc-tooling issues." }, { - "left": "issue:43937", - "right": "issue:44485", + "left": "issue:42898", + "right": "issue:46614", "accept": false, - "reason": "GLM-5 generation-config validation and GLM-5 RoPE implementation are different defects." + "reason": "Tokenizer cleanup behavior change and stale phonemizer language state do not share a concrete bug path." } ] }, @@ -15812,29 +15743,28 @@ "cluster_id": "cluster-10000-503", "nodes": [ "issue:28348", - "issue:33453", - "issue:39401", + "issue:38617", "issue:42222", "issue:42371", "issue:42673", "issue:42796", - "issue:43064", "issue:43066", "issue:43116", "issue:43441", "issue:43493", "issue:43611", "issue:43618", - "issue:43673", "issue:43825", "issue:43827", "issue:43874", + "issue:43937", "issue:43976", "issue:43992", "issue:44038", "issue:44075", "issue:44263", "issue:44355", + "issue:44485", "issue:44610", "issue:44625", "issue:44704", @@ -15844,6 +15774,7 @@ "issue:45072", "issue:45083", "issue:45092", + "issue:45278", "issue:45357", "issue:45446", "issue:45478", @@ -15861,6 +15792,8 @@ "issue:46981" ], "soft_pairs": [ + "issue:38617|issue:45278", + "issue:43937|issue:44485", "issue:43066|issue:44779", "issue:43611|issue:46002", "issue:42673|issue:45072", @@ -15885,207 +15818,205 @@ "issue:28348|issue:46640", "issue:45092|issue:46762", "issue:42673|issue:42796", - "issue:42222|issue:43441", "issue:43827|issue:45446", - "issue:33453|issue:39401", - "issue:43064|issue:43673" + "issue:42222|issue:43441" ], - "prepared_review_unit_hash": "5b4c93500326ffb37e24391bd8756d9a2bbd4436c5fd776ae139d78266c61c5a", + "prepared_review_unit_hash": "2d937999e793d0b13da86904f910d16f7dce735aadd704c33bd008d49077a2cd", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35843, - "estimated_input_tokens": 8961, - "estimated_eval_tokens": 18178 + "serialized_chars": 35794, + "estimated_input_tokens": 8949, + "estimated_eval_tokens": 18154 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose similarity set of mostly unrelated Transformers issues. None of the soft pairs are close enough to treat as the same underlying bug/change, so all edges are rejected.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44779 is the clearest concrete bug report in the set: it has a specific regression claim, version split (v4 vs v5), and an identifiable tokenizer failure mode.", + "summary": "These items mostly look like separate regression reports across different models, tokenizer paths, and runtime behaviors. None of the soft pairs appear to be the same underlying bug/change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45278 is the broadest, most umbrella-style report in the set (\u201cmany import errors after update\u201d), so it is the best single issue to anchor the cluster even though most other items are not duplicates of it.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44779 is the strongest representative issue because it is specific, user-facing, and well-scoped compared with the other mostly broad, model-specific, or docs-only reports.", + "best_issue_reason": "issue:45278 is the clearest cluster representative: general regression wording, broad scope, and strongest fit as an umbrella issue among otherwise unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but one is a specific missing symbol and the other is a broad upgrade complaint; not the same concrete bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Different GLM-5 failures: invalid generation config vs RoPE implementation concerns." + }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Both involve tokenizers in v5, but the failures differ: decoder-type mismatch vs incorrect DeepSeek tokenization results. Different bug surfaces and likely different fixes." + "reason": "Both are tokenizer regressions in v5, but they affect different models and likely different code paths." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Loading via base_model_prefix and NaNs in classification heads are unrelated failure modes and code paths." + "reason": "Loading via base_model_prefix and NaNs in classification heads are unrelated bugs." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leakage across dataloader threads is unrelated to bfloat16 dtype mismatches during inference." + "reason": "VRAM leak in multithreaded dataloading is unrelated to bfloat16 dtype mismatch in inference." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "A Python-version compatibility issue is not the same as a torch-version gate for a flex_attention import." + "reason": "Different environment/version-check issues in separate code paths." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Different model families and different symptoms: Qwen3-VL-Moe bug vs Qwen3.5 MoE from_pretrained failure." + "reason": "Both mention Qwen MoE, but they describe different failures and model variants." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Multi-label example script output is unrelated to SigLIP2 implementation numerical discrepancy." + "reason": "Example-script multi-label behavior and SigLIP2 implementation discrepancy are not the same bug." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both touch Qwen3.5, but one is about save_pretrained visual encoder keys and the other about from_pretrained loading errors; not the same bug." + "reason": "save_pretrained key mismatch and from_pretrained loading error are different problems." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors are unrelated to a documentation parameter-name mismatch." + "reason": "Compiled-file execution errors and a docs parameter-name mismatch are unrelated." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "num_labels propagation and NaNs in classification heads are different issues with different likely fixes." + "reason": "num_labels propagation is a config wiring bug, not the same as NaNs in heads." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 API guidance and a missing image patch helper are unrelated." + "reason": "TF32 API guidance and GLM46V image-patch helper absence are unrelated." }, { "left": "issue:44610", "right": "issue:46909", "accept": false, - "reason": "Processor image-size mismatch is not the same as a broken documentation example." + "reason": "Processor image-size mismatch and a broken documentation example are different issues." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "A qwen3_omni_moe feature-length helper bug is unrelated to a flex_attention import version check." + "reason": "Feature-extraction length helper behavior and a PyTorch import version check are unrelated." }, { "left": "issue:46726", "right": "issue:46981", "accept": false, - "reason": "Missing vision backend handling and repeated HTTP 429 retries are different AutoProcessor problems." + "reason": "Missing vision backend handling and HTTP 429 retry behavior in AutoProcessor are different failure modes." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Optimizer argument handling and trainer evaluate-after-train behavior are distinct code paths and bugs." + "reason": "Optimizer argument handling and trainer evaluate/train interaction are separate bugs." }, { "left": "issue:43618", "right": "issue:46829", "accept": false, - "reason": "Both are output-field issues, but they affect different model outputs and fields; not the same concrete bug." + "reason": "Both affect model outputs, but they involve different output fields and different models." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script multi-label behavior is unrelated to a pipeline() error message/docs deprecation issue." + "reason": "Example-script classification behavior and pipeline error messaging are not the same code-path problem." }, { "left": "issue:43066", "right": "issue:46710", "accept": false, - "reason": "These are both regression reports around DeepSeek/tokenization, but one is a tokenizer decoder-type problem and the other is a broad accuracy regression; too weak to merge." + "reason": "Tokenizer decoder-type bug and DeepSeek accuracy regression are too different to merge." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "UMT5 encoder weight loading and AutoProcessor kwargs forwarding are separate loading bugs." + "reason": "A missing weight in UMT5 loading is unrelated to AutoProcessor kwargs forwarding." }, { "left": "issue:46614", "right": "issue:46752", "accept": false, - "reason": "A stale phonemizer backend after language switching is unrelated to empty-list chat templating crashes." + "reason": "Tokenizer phonemizer backend state and empty-chat-template handling are unrelated." }, { "left": "issue:44995", "right": "issue:46697", "accept": false, - "reason": "A stale indexer cache crash and image over-padding are different subsystems and symptoms." + "reason": "GLM-MoE stale cache crash and Swin2SR over-padding are unrelated model/component bugs." }, { "left": "issue:44263", "right": "issue:46762", "accept": false, - "reason": "A torch.split return-value bug in an indexer is not the same as numerical divergence in Minimax MSA attention." + "reason": "GlmMoeDsa torch.split handling and Minimax attention numerical differences are different problems." }, { "left": "issue:28348", "right": "issue:46640", "accept": false, - "reason": "One is a specific GPT2LMHeadModel FlashAttention 2 support request, the other a generic Flash Attention 2 issue; not enough evidence of the same bug/change." + "reason": "A flash-attention feature request for GPT-2 is not the same as a flash-attention bug report." }, { "left": "issue:45092", "right": "issue:46762", "accept": false, - "reason": "InternVL2 meta-init incompatibility and Minimax attention numerical differences are unrelated." + "reason": "Meta-init incompatibility for InternVL2 and Minimax attention numerics are unrelated." }, { "left": "issue:42673", "right": "issue:42796", "accept": false, - "reason": "VRAM leakage in threaded loading and tokenizer BPE-marker handling are unrelated." - }, - { - "left": "issue:42222", - "right": "issue:43441", - "accept": false, - "reason": "Different model families and different failures: vitpose breakage vs Ministral-3 FlashAttention incompatibility." + "reason": "VRAM leak in Qwen3ForCausalLM and Ministral tokenizer marker handling are unrelated." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() and a flex_attention version check are unrelated." - }, - { - "left": "issue:33453", - "right": "issue:39401", - "accept": false, - "reason": "A generic tokenizer-loading regression and a Qwen3 offset_mapping bug are not the same underlying issue." + "reason": "Docs still mentioning pipeline() and a flex_attention import version check are unrelated." }, { - "left": "issue:43064", - "right": "issue:43673", + "left": "issue:42222", + "right": "issue:43441", "accept": false, - "reason": "Optimizer state corruption under FSDP2/PEFT is unrelated to missing GenerationMixin cache during chunked_prefill." + "reason": "Vitpose breakage and Ministral FlashAttention failure are different model-specific regressions." } ] }, @@ -16109,12 +16040,14 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-503", "nodes": [ + "issue:33453", "issue:38617", "issue:39401", "issue:41628", "issue:42222", "issue:42617", "issue:42673", + "issue:43064", "issue:43122", "issue:43408", "issue:43450", @@ -16148,7 +16081,6 @@ "issue:46002", "issue:46077", "issue:46178", - "issue:46209", "issue:46231", "issue:46298", "issue:46328", @@ -16158,6 +16090,8 @@ "issue:46981" ], "soft_pairs": [ + "issue:33453|issue:39401", + "issue:43064|issue:43673", "issue:43937|issue:45563", "issue:43824|issue:46981", "issue:43673|issue:46328", @@ -16172,9 +16106,9 @@ "issue:39401|issue:44779", "issue:44410|issue:45310", "issue:45071|issue:46512", + "issue:43450|issue:43986", "issue:45632|issue:46981", "issue:45341|issue:45561", - "issue:43450|issue:43986", "issue:45341|issue:46328", "issue:44829|issue:45216", "issue:44117|issue:45081", @@ -16185,211 +16119,216 @@ "issue:45362|issue:46328", "issue:42222|issue:46178", "issue:44877|issue:46231", - "issue:44484|issue:45310", - "issue:39401|issue:46209" + "issue:44484|issue:45310" ], - "prepared_review_unit_hash": "6baa9b8a24c38163c73a56fbd3beaac71451fda9f0eee39d45c929456c1b145f", + "prepared_review_unit_hash": "599915ded169037d6984fb02b2166d65841521f36c280c3924cbdebc6729a9b9", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35756, - "estimated_input_tokens": 8939, - "estimated_eval_tokens": 18134 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36669, + "estimated_input_tokens": 9168, + "estimated_eval_tokens": 18592 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated Transformers issues grouped by superficial similarity; none of the soft pairs look like the same underlying bug/change, so no duplicates should be accepted.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue is a true duplicate hub; the cluster is heterogeneous. If a representative must be chosen, issue:43937 is the broadest user-facing regression report, but it does not subsume the others.", + "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, config/import, processor, training, and logging issues. The soft edges are superficial title similarity only; none appear to be the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits this cluster because the items span multiple unrelated failure modes and code paths.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:43986 is one of the clearest actionable bug reports in the set (explicit crash path and missing-dependency condition), but it still is not representative of the whole cluster.", + "best_issue_reason": "No issue is a strong global representative; if forced, 43122 is the broadest tokenizer-regression report, but it still does not unify the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both mention tokenizer problems, but one is generic loading regression and the other is a Qwen3 offset_mapping bug; different code paths and symptoms." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "One is FSDP/PEFT optimizer-state corruption, the other is a generation cache issue in chunked prefill; not the same bug." + }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "GenerationConfig validation failure vs a stale warning in paged generate(); different behavior and code paths." + "reason": "GLM generation config validation and a stale paged-generate warning are unrelated." }, { "left": "issue:43824", "right": "issue:46981", "accept": false, - "reason": "ImportError for a missing model class vs repeated HTTP 429 retries in AutoProcessor; unrelated failure modes." + "reason": "Both involve loading/auto APIs, but one is an import error for a specific model class and the other is repeated HTTP 429 retries in AutoProcessor." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Chunked prefill cache regression vs SequenceFeatureExtractor.pad performance issue; not the same bug." + "reason": "Both touch generation/padding-ish utilities, but one is cache handling and the other is a performance issue in SequenceFeatureExtractor.pad." }, { "left": "issue:42673", "right": "issue:46655", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM multithreading vs incorrect video padding calculation; different subsystems." + "reason": "VRAM leakage in Qwen3ForCausalLM across dataloader threads is unrelated to a video processor padding bug." }, { "left": "issue:42222", "right": "issue:46752", "accept": false, - "reason": "Broken vitpose models vs apply_chat_template on empty list; unrelated model families and code paths." + "reason": "Vitpose model breakage and apply_chat_template empty-list IndexError are different subsystems and failure modes." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Unexpected _is_hf_initialized arg during init_empty_weights vs PreTrainedConfig type-checking regression; distinct defects." + "reason": "init_empty_weights argument mismatch and PreTrainedConfig type checking regression are separate v5 API issues." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "Version pin bump to unreleased PEFT vs wrong PyTorch version check for AuxRequest import; not a duplicate." + "reason": "One is a PEFT version pin problem; the other is an incorrect PyTorch version check for flex_attention." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None handling vs flash_attention crash on s_aux=None; different components and triggers." + "reason": "TOKENIZER_MAPPING_NAMES None handling and flash_attention crash on s_aux=None are unrelated." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "GPT2 attention scaling ignored under SDPA/FlashAttention vs Gemma4 bidirectional mask construction; separate model-specific bugs." + "reason": "GPT2 attention scaling under SDPA/FlashAttention and Gemma4 bidirectional-mask construction affect different models and logic." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "Serialization of tied lm_head weights vs NaNs in classification heads; unrelated symptoms and fixes." + "reason": "Serialization of lm_head.weight/tied weights is distinct from NaNs in classification heads after checkout/init." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GLM-5 GenerationConfig invalid vs Qwen3.5 flash-attention generation failure; different models and root causes." + "reason": "Both are generation-related, but one is invalid GenerationConfig handling and the other is Qwen3.5 flash-attention generation behavior." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Qwen3 tokenizer offset_mapping bug vs Deepseek tokenization regression in v5; same broad area, different concrete issue." + "reason": "Both are tokenizer regressions, but the first is offset mapping and the second is Deepseek tokenization output; different concrete defects." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing qkv/gate projections in qwen3next layer 0 vs Qwen3.5 Moe from_pretrained error; not the same failure." + "reason": "Missing qkv/gate projections in qwen3next layer 0 is unrelated to Qwen3.5 Moe from_pretrained loading errors." }, { "left": "issue:45071", "right": "issue:46512", "accept": false, - "reason": "Type checking break in v5.4.0 vs stale config.dtype after from_config(dtype=...); different API regressions." + "reason": "Type checking regression in PreTrainedConfig is not the same as stale dtype metadata after from_config(dtype=...)." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Video processor shape bug and AutoProcessor crash without torchvision are both video-related but are different code paths and root causes." }, { "left": "issue:45632", "right": "issue:46981", "accept": false, - "reason": "Local trust_remote_code cache-path collision vs repeated retry behavior on 429 responses; unrelated loader issues." + "reason": "trust_remote_code cache-path collision is unrelated to repeated retry behavior on HTTP 429 responses." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Minor testing_utils bug vs xdist race on captured_info.txt; different test infrastructure problems." - }, - { - "left": "issue:43450", - "right": "issue:43986", - "accept": false, - "reason": "Batched video processor shape bug vs crash loading video model without torchvision; distinct preprocessing vs dependency handling." + "reason": "A bug in testing_utils.py is not the same as pytest-xdist workers racing on captured_info.txt." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "testing_utils bug vs numpy-to-list conversion slowdown in SequenceFeatureExtractor.pad; not a duplicate." + "reason": "Testing-utils bug and SequenceFeatureExtractor.pad performance issue do not share a concrete underlying change." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "FlashAttention-3 training degeneration vs Qwen3.5 save_pretrained checkpoint regression; different code paths." + "reason": "FlashAttention-3 degenerate training in sequence classification is unrelated to Qwen3.5 checkpoint save_pretrained regression." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping None handling vs Mistral regex patch crash on backend_tokenizer access; unrelated causes." + "reason": "Mapping-name fallback and Mistral regex patch crash are separate tokenizer-loading failures." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "Cannot run 3d_parallel.py vs a minor testing_utils bug; no shared underlying bug." + "reason": "Running 3d_parallel.py and a testing_utils.py bug are unrelated." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "sam3_video/sam3_tracker type mismatch warning vs MobileLLM key error on load; different model-loading issues." + "reason": "sam3_video/sam3_tracker config mismatch is distinct from a key error loading facebook/MobileLLM-125M." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenization changes across versions vs noisy logging message; totally different severity and behavior." + "reason": "Tokenization differences across versions and a noisy logging message are not the same bug." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "ImportError for layer_type_validation vs ImportError for AutoImageProcessor; both import-related but distinct missing symbols." + "reason": "Both are import errors, but they involve different symbols and modules (`layer_type_validation` vs `AutoImageProcessor`)." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen3.5 chat crash vs padding performance inefficiency; unrelated functionality." + "reason": "Qwen3.5 chat crash and SequenceFeatureExtractor.pad performance issue are unrelated." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "Broken vitpose models vs backend mapping rejection for tensorflow_text; different areas and fixes." + "reason": "Vitpose breakage and a BACKENDS_MAPPING tensorflow_text error are different loading/backend issues." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config blocking granite_speech load vs docstring/parameter-name mismatch in set_training(); not a duplicate." + "reason": "Strict config preventing granite_speech loading is unrelated to a docs/parameter-name mismatch in set_training()." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "save_pretrained shard-size question vs Qwen3.5 Moe loading error; not the same bug." - }, - { - "left": "issue:39401", - "right": "issue:46209", - "accept": false, - "reason": "Tokenizer offset mapping regression vs FP8 inference failure on B200; completely different problem class." + "reason": "save_pretrained shard-size question and Qwen3.5 Moe from_pretrained error are unrelated." } ] }, @@ -16415,10 +16354,10 @@ "nodes": [ "issue:27957", "issue:38617", + "issue:39401", "issue:41628", "issue:42222", "issue:42371", - "issue:42886", "issue:43012", "issue:43066", "issue:43097", @@ -16432,7 +16371,6 @@ "issue:43874", "issue:43931", "issue:43937", - "issue:43986", "issue:44038", "issue:44075", "issue:44077", @@ -16457,12 +16395,14 @@ "issue:45542", "issue:45656", "issue:46172", + "issue:46209", "issue:46617", "issue:46762", "issue:46894", "issue:46981" ], "soft_pairs": [ + "issue:39401|issue:46209", "issue:42371|issue:44261", "issue:43937|issue:44496", "issue:44261|issue:46762", @@ -16491,117 +16431,120 @@ "issue:43441|issue:44829", "issue:43012|issue:43577", "issue:43097|issue:43452", - "issue:45003|issue:45341", - "issue:43986|issue:45341", - "issue:42886|issue:46981", - "issue:44261|issue:44485" + "issue:45003|issue:45341" ], - "prepared_review_unit_hash": "d4e9052fab9de40a8db375afc57004d0dde5b6681fd4ccf354fd2e62cdc0b739", + "prepared_review_unit_hash": "e19bdb7bc7cc15ce84d93efcfdc9638f742e0e9851e438ca9976ce7952d93631", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36828, - "estimated_input_tokens": 9207, - "estimated_eval_tokens": 18670 + "soft_pair_count": 30, + "serialized_chars": 36157, + "estimated_input_tokens": 9040, + "estimated_eval_tokens": 18336 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated Transformers issues, with a small import-regression subgroup. Only one soft edge looks duplicate-like: a specific import failure and a broader post-upgrade import-error umbrella.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45278 is the best umbrella issue: it broadly covers the same post-upgrade import breakage pattern seen in the import-related subset, and is more representative than the narrow single-failure reports.", + "summary": "This cluster is mostly heterogeneous: many issues share broad vocabulary around imports, tokenization, model loading, or attention/precision, but they do not look like the same underlying bug. I would not merge any of the soft-edge pairs; the strongest representative is a broad import-regression report.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella here: it reports multiple import errors after a Transformers upgrade, so it best represents the cluster\u2019s general regression theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the strongest representative for this cluster because it is the broadest, most central issue around import regressions after upgrading Transformers.", + "best_issue_reason": "45278 is the most suitable triage representative because it is the most general and could subsume narrower import-related reports, even though it is not a true duplicate of most others.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset-mapping bug vs. FP8 inference failure are unrelated." + }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 configuration behavior and MLA layernorm epsilon precision are different bugs in different code paths." + "reason": "Both are precision-related, but one is TF32 policy guidance and the other is a model-specific RMSNorm epsilon bug." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Both are loading/config errors, but one is a GLM generation-config validation problem and the other is an unrecognized model/config issue." + "reason": "Invalid GenerationConfig vs. unrecognized model/config loading are different failure modes." }, { "left": "issue:44261", "right": "issue:46762", "accept": false, - "reason": "Both mention numerical issues, but they concern different models and different mechanisms (layernorm epsilon vs MSA attention mismatch)." + "reason": "Both concern numerical differences, but they affect different models and code paths." }, { "left": "issue:42371", "right": "issue:46617", "accept": false, - "reason": "TF32 settings and offline trust_remote_code cache loading are unrelated." + "reason": "TF32 behavior settings and cache FileNotFoundError are unrelated." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen-related, but one is missing projections in qwen3next and the other is a from_pretrained error for Qwen3.5 MoE." + "reason": "Different Qwen model families and different loading/projection problems." }, { "left": "issue:44933", "right": "issue:45278", - "accept": true, - "reason": "This looks like the same upgrade-driven import regression: 44933 is a specific missing import, while 45278 is the broader report of many import errors after the version update." + "accept": false, + "reason": "A specific missing import is not the same as a broad upgrade-wide import-regression report." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "Both involve optimizers, but one is ignored SGD args and the other is a double optimizer step under deepspeed." + "reason": "One is unused SGD arguments; the other is double optimizer stepping under deepspeed." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "An import error from configuration_utils and a feature-extraction bug are not the same underlying defect." + "reason": "ImportError for a utility symbol and a fbank feature issue are unrelated." }, { "left": "issue:43208", "right": "issue:46894", "accept": false, - "reason": "These are different model-specific training bugs with no clear shared code-path." + "reason": "Different model/training bugs; only the vague word \"training\" overlaps." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both relate to tokenizer behavior changes in v5, but one is a wrong decoder type and the other is a tokenization-output difference." + "reason": "Tokenizer decoder type mismatch and different tokenization output are related thematically but not the same concrete bug." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "Different GLM issues: torch.split handling in a DSA indexer versus a RoPE implementation bug." + "reason": "Moe indexer torch.split semantics and GLM-5 RoPE implementation are different code paths." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "A removed config field and unexpected position-id keys are different regression reports." + "reason": "Removed config API and unexpected position-id keys are different regressions." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Both are chat/completions failures, but one is a generic API rejection and the other is a model-specific streaming bug." + "reason": "Both mention chat/completions, but one is a general endpoint acceptance issue and the other is a model-specific streaming failure." }, { "left": "issue:43673", @@ -16613,113 +16556,92 @@ "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "A broken vitpose model and a TensorFlow backend detection error are unrelated." + "reason": "VitPose model breakage and TensorFlow backend detection are unrelated." }, { "left": "issue:44077", "right": "issue:46981", "accept": false, - "reason": "Optional post_init validation and repeated 429 retry behavior are unrelated." + "reason": "PatchTSMixer post_init validation and repeated 429 retries are unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Both are Qwen vision-family issues, but they are different concrete failures: weight-shape mismatch versus missing projections." + "reason": "Weight-shape mismatch and missing projections are different model-loading failures." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "An AutoImageProcessor import failure and PretrainedConfig type-checking regression are different problems." + "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression are unrelated." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "PretrainedConfig type-checking and a missing-TensorFlow backend error do not share the same bug." + "reason": "Type-checking breakage and missing TensorFlow backend handling are different issues." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing multimodal-token helper method and MLA rms_norm_eps precision behavior are unrelated." + "reason": "Image-patch counting bug and RMSNorm epsilon mismatch are different model-component errors." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Shard-size behavior in save_pretrained and a Qwen3.5 MoE from_pretrained error are unrelated." + "reason": "Save-pretrained sharding size question and Qwen3.5-Moe loading error are unrelated." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "Different model internals: GlmMoeDsaIndexer split handling versus deepseek-v4 tensor transpose logic." + "reason": "Different GLM/DeepSeek tensor-shape issues with no shared concrete code path." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "A missing ReLU in the DSA indexer and a deepseek-v4 dtype/transpose issue are not the same defect." + "reason": "DSA ReLU bug and DeepSeek transpose bug are unrelated." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Both are Flash Attention 2 support complaints, but they are model-specific and not the same code-path bug." + "reason": "Both mention FlashAttention 2, but they target different models and different support gaps." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "FlashAttention v5 RC failure and flash_attention_3 degenerate training are different failure modes." + "reason": "FlashAttention failures on different models with different symptoms; not a single fixable bug." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both concern dtype/precision, but one is a PyTorch warning during compile and the other is incorrect model dtype loading." + "reason": "Compile-time float32 warning and BLIP2 dtype staying float32 are different dtype issues." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "A removed config option and a gguf_file load-path break are unrelated." + "reason": "Removed config API and gguf_file/pretrained loading breakage are unrelated." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access and a testing_utils bug are unrelated." - }, - { - "left": "issue:43986", - "right": "issue:45341", - "accept": false, - "reason": "AutoProcessor/video-model torchvision handling and a testing_utils bug are unrelated." - }, - { - "left": "issue:42886", - "right": "issue:46981", - "accept": false, - "reason": "Offline cache loading and repeated 429 retries are different request/path handling bugs." - }, - { - "left": "issue:44261", - "right": "issue:44485", - "accept": false, - "reason": "A layernorm epsilon precision issue and a RoPE implementation issue are not the same underlying bug." + "reason": "Sys.modules access in modeling_utils and a small testing_utils bug are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary stays within the packet evidence: it correctly treats the cluster as mostly heterogeneous, identifies only the import-regression pair as duplicate-like, and does not overclaim similarity on the other soft edges. The canonical issue choice is conservative and supported by the titles." - }, - "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 }, @@ -16738,6 +16660,7 @@ "nodes": [ "issue:38617", "issue:42757", + "issue:42886", "issue:42890", "issue:43012", "issue:43064", @@ -16753,16 +16676,15 @@ "issue:44038", "issue:44062", "issue:44117", + "issue:44261", "issue:44263", "issue:44315", "issue:44351", - "issue:44380", "issue:44410", "issue:44484", - "issue:44704", + "issue:44485", "issue:44936", "issue:44987", - "issue:45103", "issue:45341", "issue:45372", "issue:45375", @@ -16775,16 +16697,19 @@ "issue:46298", "issue:46365", "issue:46369", - "issue:46421", "issue:46655", "issue:46664", "issue:46701", "issue:46710", "issue:46752", "issue:46821", - "issue:46897" + "issue:46897", + "issue:46981" ], "soft_pairs": [ + "issue:43986|issue:45341", + "issue:42886|issue:46981", + "issue:44261|issue:44485", "issue:45753|issue:46369", "issue:44117|issue:44987", "issue:43531|issue:45341", @@ -16810,206 +16735,210 @@ "issue:43012|issue:46821", "issue:44117|issue:46701", "issue:43502|issue:43992", - "issue:44936|issue:46365", - "issue:44704|issue:45103", - "issue:44380|issue:46421" + "issue:44936|issue:46365" ], - "prepared_review_unit_hash": "bc1c23594c165cbebd2878ed8a661428406f170647c1040e20ed52d7c66e82c7", + "prepared_review_unit_hash": "d815a8b4832a59ddac4e65fb9cb1218887decff71b6a93551f764fb6d362f842", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35341, - "estimated_input_tokens": 8836, - "estimated_eval_tokens": 17928 + "soft_pair_count": 29, + "serialized_chars": 35604, + "estimated_input_tokens": 8901, + "estimated_eval_tokens": 18058 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose mix of unrelated Transformers issues: import/regression errors, model-specific bugs, training problems, and a few loading/processor crashes. None of the proposed soft pairs look like true duplicates of the same underlying bug or change.", - "confidence": 0.88, - "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated import errors, model-specific bugs, training regressions, and test flakiness rather than one shared defect.", + "summary": "This cluster is noisy and mostly consists of unrelated Transformer/HF issues that only share broad topical similarity. I do not see any soft pair that is clearly the same underlying bug/change, so I would not merge any of them.", + "confidence": 0.69, + "canonical_issue_reason": "If a representative issue must be chosen, issue:43873 has the broadest cross-cutting impact and the most discussion/references; however, the set does not really form a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:38617 is the closest representative of the small import-regression subtheme (missing exported symbol from transformers internals), but it still does not explain the rest of the cluster.", + "best_issue_reason": "Issue:43873 is the strongest single representative by activity and scope, but only as a weak canonical pick because the cluster is heterogeneous.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Both mention crashes, but one is an AutoProcessor/video-model dependency failure and the other is a testing_utils bug; different code paths and symptoms." + }, + { + "left": "issue:42886", + "right": "issue:46981", + "accept": false, + "reason": "Both involve hub/network behavior, but one is offline cache loading and the other is repeated 429 retries via HF_ENDPOINT; not the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "One is a missing rms_norm_eps config causing precision mismatch; the other is a GLM RoPE implementation issue. Different model components." + }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Different model families and failure modes: an infinite-loop image bug vs a memory overflow report." + "reason": "Both are model runtime failures, but one is an infinite loop on a specific image and the other is a memory overflow in sam3-litetext." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Both involve loading, but one is a tokenizer mapping None case and the other is a package loading failure; not the same bug." + "reason": "The first is a tokenizer-mapping None assumption bug; the second is a model-specific load failure on transformers>=5.1.0. Too broad to be the same defect." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to a testing_utils bug." + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug are unrelated." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Both are GLM/MoE-related, but one is a missing image-patch helper and the other is a torch.split indexing issue; different code paths." + "reason": "Both touch GLM-related code, but one is an image processor missing a method and the other is a torch.split return-value issue; different failures." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "Different DeepSeek models and distinct failures; no clear shared underlying defect." + "reason": "DeepSeekV3 divergence from remote implementation is not the same as the deepseek-v4 comb.transpose bug." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "Removed config API vs flash-attention crash; only a broad transformer-v5 theme, not the same bug." + "reason": "Removed tie_embeddings_and_encoder_decoder and flash_attention s_aux=None crash are unrelated change/bug paths." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Chunked-prefill cache initialization and partial-batch masking are separate pipeline/runtime issues." + "reason": "One is missing GenerationMixin cache during chunked_prefill; the other is a MaskGenerationPipeline partial-batch bookkeeping bug." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Both are Qwen-related, but one is missing projections in layer 0 and the other is a missing config field silently dropped." + "reason": "Both are Qwen3_5/Qwen3-VL-Moe related, but one is missing projections in layer 0 and the other is a missing config field dropped by strict loading." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Both are processor-loading failures, but they stem from different missing dependencies/imports and different models." + "reason": "Both are import-time crashes, but they fail on different missing dependencies/symbols and in different processor-loading flows." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Same broad Qwen family, but one is a transformers 5.0 model bug and the other is a specific layer-0 projection omission." + "reason": "General Qwen3-VL-Moe bug report versus a specific missing-attention-projection defect; not the same concrete issue." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "A removed config attribute and a broken sinq save/load path are unrelated." + "reason": "A removed method in v5 and a broken quantized save/load round-trip are different regressions." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are ImportErrors, but one is a transformers internal symbol and the other is a huggingface_hub symbol; different packages and fixes." + "reason": "Different missing imports from different modules/symbols; superficial import-error similarity only." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Distributed optimizer-state corruption is unrelated to Liger Kernel not applying with model_init." + "reason": "Optimizer-state corruption with FSDP2/PEFT is unrelated to Liger Kernel not being applied under model_init." }, { "left": "issue:43208", "right": "issue:46897", "accept": false, - "reason": "Both mention training, but they concern different models and different failure mechanisms." + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are different model-specific training issues." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "AutoProcessor dependency crash vs save_pretrained shard-size question; no shared bug." + "reason": "AutoProcessor dependency crash and max_shard_size question are not the same bug." }, { "left": "issue:46655", "right": "issue:46664", "accept": false, - "reason": "Same SmolVLM video processor area, but one is padding calculation and the other is num_frames semantics." + "reason": "Both are SmolVLM video-processor issues, but one is padding calculation and the other is num_frames semantics; related subsystem, different defect." }, { "left": "issue:46298", "right": "issue:46752", "accept": false, - "reason": "Noisy logging and empty-list chat-template IndexError are unrelated." + "reason": "A noisy logging complaint is unrelated to an apply_chat_template empty-list crash." }, { "left": "issue:46172", "right": "issue:46710", "accept": false, - "reason": "DeepSeek-v4 tensor conversion bug vs downstream accuracy regression after upgrade; not the same concrete issue." + "reason": "A deepseek-v4 tensor transform bug is unrelated to an accuracy regression after upgrading Transformers." }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "Test seed flakiness and a logging message issue are unrelated." + "reason": "Seed-related test flakiness and a noisy logging message do not describe the same problem." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Tokenizer AddedToken argument duplication is a different failure from TOKENIZER_MAPPING_NAMES returning None." + "reason": "Tokenizers.AddedToken keyword duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer initialization bugs." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Removed tie_embeddings API and quantization offloading problems are distinct." + "reason": "Removed API in v5 and quantization/offloading issues are unrelated." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are import-related, but they involve different missing symbols and different import surfaces; not one bug." + "reason": "Both are import errors, but they involve different symbols and modules and likely different release regressions." }, { "left": "issue:43012", "right": "issue:46821", "accept": false, - "reason": "A PyTorch bfloat16 warning and a TimesFM window_size AttributeError do not share an underlying code path." + "reason": "A PyTorch warning during bfloat16 compilation is unrelated to a window_size AttributeError in TimesFM 2.5." }, { "left": "issue:44117", "right": "issue:46701", "accept": false, - "reason": "Tokenizer mapping None and DiffusionGemma pipeline failure are separate loading/inference problems." + "reason": "Tokenizer mapping assumptions and a diffusion-gemma image-text pipeline failure are different code paths." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "local_files_only API leakage and a missing embed_tokens.weight in from_pretrained are different load-time bugs." + "reason": "local_files_only still hitting API requests and UMT5Encoder missing embed_tokens.weight are unrelated loading bugs." }, { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "Trainer evaluation failure after train is unrelated to importing Trainer only after uninstalling datasets." - }, - { - "left": "issue:44704", - "right": "issue:45103", - "accept": false, - "reason": "Passing kwargs to cached_file and an auto_docstring annotation crash are unrelated." - }, - { - "left": "issue:44380", - "right": "issue:46421", - "accept": false, - "reason": "Attention scaling under SDPA/FlashAttention is unrelated to chunked-prefill cache lazy initialization." + "reason": "Trainer.evaluate after train and inability to import Trainer with datasets installed are not the same issue." } ] }, @@ -17039,7 +16968,6 @@ "issue:42490", "issue:42757", "issue:42907", - "issue:42994", "issue:43012", "issue:43064", "issue:43066", @@ -17051,7 +16979,6 @@ "issue:43550", "issue:43582", "issue:43618", - "issue:43756", "issue:43761", "issue:43824", "issue:43825", @@ -17067,10 +16994,11 @@ "issue:44380", "issue:44483", "issue:44484", - "issue:44485", "issue:44514", "issue:44655", + "issue:44704", "issue:44938", + "issue:45103", "issue:45478", "issue:45479", "issue:45584", @@ -17080,9 +17008,12 @@ "issue:46172", "issue:46178", "issue:46298", - "issue:46365" + "issue:46365", + "issue:46421" ], "soft_pairs": [ + "issue:44704|issue:45103", + "issue:44380|issue:46421", "issue:44938|issue:46178", "issue:43986|issue:44016", "issue:42907|issue:44655", @@ -17090,8 +17021,8 @@ "issue:43825|issue:44655", "issue:43012|issue:43550", "issue:45712|issue:46298", - "issue:43208|issue:43550", "issue:43012|issue:44368", + "issue:43208|issue:43550", "issue:43992|issue:44484", "issue:43352|issue:43976", "issue:43352|issue:44380", @@ -17109,240 +17040,235 @@ "issue:43064|issue:44380", "issue:42490|issue:43012", "issue:44483|issue:44514", - "issue:43208|issue:43526", "issue:43618|issue:44380", + "issue:43208|issue:43526", "issue:38617|issue:43502", - "issue:43618|issue:43761", - "issue:43756|issue:44485", - "issue:42994|issue:44655" + "issue:43618|issue:43761" ], - "prepared_review_unit_hash": "b106326efc311f6d13a54bb749a3e563a25a68f04fce3f95b457d2d2d76b21a9", + "prepared_review_unit_hash": "8dd5521e5b39940ec85fc6b432ecf7dffeb43aa3136cfaff9dbd62de351c2c38", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37088, - "estimated_input_tokens": 9272, - "estimated_eval_tokens": 18800 + "serialized_chars": 37196, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is heterogeneous: the soft-similarity pairs mostly connect unrelated regression reports that share only broad subsystem keywords (imports, saving, warnings, models, or v5 regressions). I would not collapse any of the suggested pairs as duplicates. The best single issue to anchor the cluster is the broad open import regression in issue 46365, but it is still not a true duplicate of the others.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 46365 is the broadest, most generally user-facing import regression in the set, so it is the least-specific representative candidate; however, the cluster is too mixed to treat it as a true duplicate anchor.", + "summary": "This cluster is mostly heterogeneous and does not form a single duplicate set. The only plausible duplicate-like subcluster is the CLIP v5 regression pair around missing outputs from CLIPVisionModel forward.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43761 is the best anchor for the only near-duplicate subcluster: it clearly describes a v5 regression in CLIPVisionModel.forward, and 43618 looks like a narrower symptom of the same output-population bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46365 is the most general and high-impact issue here, with a clear runtime failure and broad import surface, making it the best representative issue if one must be chosen.", + "best_issue_reason": "43761 is the most actionable and specific report in the set, with a concrete regression label, model path, and expected-vs-actual behavior that make it a strong triage target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different bugs: AutoProcessor kwarg forwarding vs auto_docstring crash under future annotations." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Different code paths: GPT2 attention scaling/backend behavior vs cache lazy initialization recompilation." + }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import failure is unrelated to a missing backend mapping for tensorflow_text." + "reason": "Unrelated import/runtime failures: Python 3.14 load issue vs tensorflow_text backend mapping error." }, { "left": "issue:43986", "right": "issue:44016", "accept": false, - "reason": "A video-model/torchvision import crash is unrelated to a notebook syntax error." + "reason": "Video AutoProcessor/torchvision crash is unrelated to a notebook syntax error." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals is a different failure mode from saving Pipeline objects." + "reason": "Both involve saving, but one is dequantized model persistence and the other is Pipeline.save_pretrained; too different." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Quantized model saving and a Qwen image infinite-loop bug are different code paths and symptoms." + "reason": "Dequantized save failure is unrelated to a Qwen image-generation infinite loop." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "An error-message regression in pipeline() is not the same as save_pretrained failing for Pipeline objects." + "reason": "Error-message wording for pipeline tasks is unrelated to pipeline serialization." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "A PyTorch precision warning during compile is not the same as a Bamba SDPA torch.compile failure." + "reason": "A precision warning during bfloat16 compilation is not the same as a Bamba torch.compile SDPA failure." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Left is a leftover dummy-class cleanup problem; right is a noisy logging message." + "reason": "Dummy class cleanup/import leakage is unrelated to a noisy logging message." }, { - "left": "issue:43208", - "right": "issue:43550", + "left": "issue:43012", + "right": "issue:44368", "accept": false, - "reason": "xLSTM training bugs are unrelated to Bamba SDPA compile failures." + "reason": "Different warnings with different causes: float32 precision during compile vs tie_word_embeddings warning." }, { - "left": "issue:43012", - "right": "issue:44368", + "left": "issue:43208", + "right": "issue:43550", "accept": false, - "reason": "A compile-time float32 warning is unrelated to the tie_word_embeddings warning in Qwen3.5 LoRA fine-tuning." + "reason": "xLSTM training blockers are unrelated to Bamba torch.compile SDPA." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens weights on load is unrelated to the max_shard_size default in save_pretrained." + "reason": "Missing embed_tokens.weight in UMT5Encoder is unrelated to max_shard_size defaults." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "FlashAttention support for Nemotron is unrelated to a Python version compatibility issue." + "reason": "FlashAttention support in Nemotron is unrelated to the Python version compatibility issue." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Lack of FA2 support for Nemotron is not the same bug as GPT2 attention scaling being ignored under SDPA/FlashAttention." - }, - { - "left": "issue:43066", - "right": "issue:43906", - "accept": false, - "reason": "A tokenizer decoder-type regression is not evidenced as the same bug as an isolated reproduction of another issue." + "reason": "Different features and failures: Nemotron FlashAttention support vs GPT2 attention scaling ignored." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 training breakage is unrelated to the DeepSeek tensor transpose/dtype issue." + "reason": "ESM2 breakage is unrelated to the deepseek-v4 dtype/transpose bug." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "A removed config field is not the same as a generic noisy logging message." + "reason": "Removed config API behavior is unrelated to a noisy logging message." }, { "left": "issue:45584", "right": "issue:46365", "accept": false, - "reason": "Whisper empty-transcription generation failure is unrelated to importing Trainer being blocked by datasets." + "reason": "Whisper empty-transcription generation failure is unrelated to Trainer import interference from datasets." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "Ignoring GPT2 attention scaling under backend dispatch is different from AttentionInterface.register behavior changes." + "reason": "GPT2 attention scaling bug is unrelated to AttentionInterface.register behavior changes." }, { "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "PatchTSMixer post_init validation and Trainer import failures are unrelated." + "reason": "patchtsmixer post_init validation is unrelated to importing Trainer." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained failures and Qwen3.5 MoE from_pretrained errors are different concrete failure points." + "reason": "Both mention MoE/FP8/Qwen, but one is save_pretrained FP8 and the other is a from_pretrained regression; not the same bug." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "A Gemma3Config attribute error is unrelated to an Apple Silicon allocator TypeError." + "reason": "Gemma3Config vocab_size AttributeError is unrelated to Apple Silicon caching_allocator_warmup TypeError." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention backend scaling is unrelated to the degenerate zero-loss classification bug." + "reason": "GPT2 attention backend scaling is unrelated to the zero-loss classification bug." }, { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "Backend packaging composition is not the same as a Trainer import conflict with datasets." + "reason": "Backend packaging/install issue is unrelated to Trainer import problems." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Two different import errors for different symbols and subsystems are not the same underlying bug." + "reason": "Different missing-symbol import errors from different packages/models." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer state corruption is unrelated to GPT2 attention scaling behavior." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "A local-path trailing-slash warning is unrelated to a float32 precision warning during compile." + "reason": "Local path warning on trailing slash is unrelated to a bfloat16 compile precision warning." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "A chat/completions acceptance failure is not the same bug as Qwen2_5_VLProcessor batching crashing." + "reason": "Chat completions API rejection is unrelated to Qwen2_5_VLProcessor batched padding behavior." }, { - "left": "issue:43208", - "right": "issue:43526", + "left": "issue:43618", + "right": "issue:44380", "accept": false, - "reason": "xLSTM training bugs are unrelated to BeitImageProcessorFast reduce_labels behavior." + "reason": "CLIP output regression and GPT2 attention scaling are different model/code paths." }, { - "left": "issue:43618", - "right": "issue:44380", + "left": "issue:43208", + "right": "issue:43526", "accept": false, - "reason": "CLIP attentions not being assigned is a different regression from GPT2 attention scaling under backend dispatch." + "reason": "xLSTM training blockers are unrelated to BeitImageProcessorFast label reduction." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "An import removed from configuration_utils is unrelated to local_files_only still allowing API requests." + "reason": "Missing configuration_utils symbol import is unrelated to local_files_only API requests." }, { "left": "issue:43618", "right": "issue:43761", - "accept": false, - "reason": "Both are CLIP regressions, but one is about attentions and the other about hidden_states; not enough evidence they are the same concrete bug." - }, - { - "left": "issue:43756", - "right": "issue:44485", - "accept": false, - "reason": "Both mention RoPE, but they concern different models and likely different implementation bugs." - }, - { - "left": "issue:42994", - "right": "issue:44655", - "accept": false, - "reason": "Quantized model saving failure and Pipeline.save_pretrained failure are distinct save-path issues." + "accept": true, + "reason": "Both describe the same CLIP v5 output regression in the CLIP vision forward path, with attentions/hidden_states no longer populated as expected." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The single accepted soft pair (43618/43761) is plausibly the same CLIP v5 output-regression cluster, and the rejected pairs are justified without overstating evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -17365,7 +17291,7 @@ "issue:36296", "issue:36683", "issue:42757", - "issue:42890", + "issue:42994", "issue:43010", "issue:43064", "issue:43208", @@ -17375,6 +17301,7 @@ "issue:43525", "issue:43611", "issue:43618", + "issue:43756", "issue:43867", "issue:43937", "issue:43976", @@ -17385,6 +17312,7 @@ "issue:44230", "issue:44355", "issue:44483", + "issue:44485", "issue:44492", "issue:44625", "issue:44655", @@ -17395,27 +17323,27 @@ "issue:45216", "issue:45276", "issue:45278", + "issue:45290", "issue:45310", - "issue:45375", - "issue:45561", "issue:46050", "issue:46077", "issue:46082", "issue:46202", "issue:46365", - "issue:46369", "issue:46829", "issue:46851", "issue:46877" ], "soft_pairs": [ + "issue:43756|issue:44485", + "issue:42994|issue:44655", "issue:44038|issue:44230", - "issue:43299|issue:43867", "issue:36010|issue:42757", + "issue:43299|issue:43867", "issue:45276|issue:46077", "issue:44625|issue:46365", "issue:43611|issue:44655", - "issue:44483|issue:46369", + "issue:44483|issue:45290", "issue:43352|issue:43937", "issue:36683|issue:43475", "issue:44355|issue:46877", @@ -17439,218 +17367,234 @@ "issue:44938|issue:45071", "issue:44355|issue:44855", "issue:43867|issue:46202", - "issue:36683|issue:43064", - "issue:43867|issue:45375", - "issue:42890|issue:45561" + "issue:36683|issue:43064" ], - "prepared_review_unit_hash": "fe0cb272c68eb4a69a0382033521d312ece23f03d02e9950e5d541c83562358b", + "prepared_review_unit_hash": "484f0bcc6f3764fadab56912c5904ab231f96330283afd8da7172909c193f8e3", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37233, - "estimated_input_tokens": 9309, - "estimated_eval_tokens": 18874 + "serialized_chars": 37133, + "estimated_input_tokens": 9284, + "estimated_eval_tokens": 18824 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the soft pairs share broad themes like loading, saving, configs, or Python compatibility, but they do not look like the same underlying bug/change. No PRs are present, so triage is issue-only.", - "confidence": 0.79, - "canonical_issue_reason": "issue:36683 is the most active, concrete runtime regression here and the best single anchor if one must choose a canonical issue.", + "summary": "This cluster is a heterogeneous mix of unrelated regression reports, model-specific bugs, and API/import issues. The titles share broad themes like loading/saving/imports, but none of the soft pairs look like the same concrete bug or change, so I would not merge any of them.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 36683 is the highest-signal current report in the set: it is open, has the most discussion, and describes a concrete model/config regression. There is no true duplicate cluster here, but it is a reasonable representative issue for triage.", "canonical_pr_reason": null, - "best_issue_reason": "issue:36683 has the richest discussion and a specific reproducible failure, making it the most representative standalone issue in this mixed cluster.", + "best_issue_reason": "Issue 36683 is the strongest standalone issue to anchor the cluster because it is active, well-discussed, and has a specific reproducible failure. It is not a duplicate of the others, just the best representative among unrelated issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44038", - "right": "issue:44230", + "left": "issue:43756", + "right": "issue:44485", "accept": false, - "reason": "One is a Qwen3-VL-Moe loading regression; the other is fp8 inference support for Qwen3.5 MoE. Different bugs/features." + "reason": "Different model families and failures: Smollm3 RoPE layer loss vs GLM-5 RoPE implementation discussion." }, { - "left": "issue:43299", - "right": "issue:43867", + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both mention save_pretrained, but one is quantized model saving and the other is Pipeline object serialization; different code paths." + }, + { + "left": "issue:44038", + "right": "issue:44230", "accept": false, - "reason": "Both mention model loading, but one is Qwen3VL-Moe breakage and the other is a sorted-state_dict load error. Different failure modes." + "reason": "Qwen3-VL-Moe loading bug vs fp8 inference support request; related model name, but not the same bug." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they involve different symbols and packages (GenerationMixin vs is_offline_mode)." + "reason": "Unrelated import errors from different dependencies: GenerationMixin from transformers.generation vs is_offline_mode from huggingface_hub." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are loading regressions, but one is Qwen3-VL-Moe model loading and the other is state_dict ordering during load_model." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Same model family, but one is token-embedding resizing propagation and the other is bidirectional attention masking. Separate code paths." + "reason": "Both are Gemma4 issues, but one is resizing embeddings and the other is attention-mask construction; separate bugs." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Unrelated: Qwen3.5 num_labels propagation vs importing Trainer with datasets installed." + "reason": "Qwen3.5 num_labels propagation bug vs a Trainer import conflict with datasets; unrelated." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "Different serialization problems: base_model_prefix load behavior vs Pipeline save_pretrained." + "reason": "base_model_prefix loading regression vs Pipeline save_pretrained failure; different APIs and failure modes." }, { "left": "issue:44483", - "right": "issue:46369", + "right": "issue:45290", "accept": false, - "reason": "API request rejection and SAM3 memory overflow are unrelated." + "reason": "Both concern chat-related behavior, but one is /v1/chat/completions request handling and the other is apply_chat_template crashing on tool-call assistant messages." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "FlashAttention support gate vs invalid GenerationConfig; different causes." + "reason": "Nemotron flash-attention support limitation vs GLM-5 GenerationConfig validation error; unrelated." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Different model configs/outputs and different missing attributes." + "reason": "Gemma3Config missing vocab_size vs SAM 3 Video output missing fpn_position_embeddings; different models and attributes." }, { "left": "issue:44355", "right": "issue:46877", "accept": false, - "reason": "Compiled-file import errors vs import-time docstring lint stdout noise; unrelated." + "reason": "Compiled Python file import errors vs auto_docstring printing doc-lint diagnostics at import time; not the same issue." }, { "left": "issue:44655", "right": "issue:46050", "accept": false, - "reason": "Both involve save/load, but one is Pipeline serialization and the other is a sinq quantized model roundtrip." + "reason": "Both involve save_pretrained/from_pretrained, but one is Pipeline serialization and the other is quantized model roundtrip corruption." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Different Gemma4/Qwen3.5 issues and different config fields." + "reason": "Qwen3.5 num_labels propagation vs Gemma4 bidirectional-attention mask logic; unrelated." }, { "left": "issue:36683", - "right": "issue:43525", + "right": "issue:44077", "accept": false, - "reason": "Different config attributes on different models." + "reason": "Gemma3Config missing vocab_size vs patchtsmixer post_init restriction; no shared code path." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are Python-version compatibility reports, but they target different versions and likely different breakages." + "reason": "Both are Python-version compatibility reports, but they target different versions and different breakages." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Both concern saving/loading checkpoints, but one is sorted state_dict loading and the other is Qwen3.5 save_pretrained corruption." + "reason": "State_dict sorting/load ordering bug vs Qwen3.5 save_pretrained regression; both serialization-related but not the same defect." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Attention-mask behavior vs Llama head_dim validation are unrelated." + "reason": "Gemma4 attention-mask bug vs LlamaConfig head_dim validation error; unrelated model/config issues." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP output field regression vs Qwen3.5 checkpoint save corruption." + "reason": "CLIPOutput attentions field regression vs Qwen3.5 checkpoint save corruption; separate bugs." }, { "left": "issue:43208", "right": "issue:46829", "accept": false, - "reason": "xLSTM training bugs and VideoPrism output misassignment are unrelated." + "reason": "xLSTM training blockers vs VideoPrism hidden_states field mismatch; different subsystems and symptoms." }, { "left": "issue:36683", - "right": "issue:44077", + "right": "issue:43525", "accept": false, - "reason": "Gemma3 missing vocab_size and optimizer SGD args not used are unrelated." + "reason": "Gemma3Config missing vocab_size vs Llama4Config missing pad_token_id; similar shape, but different models and config fields." }, { - "left": "issue:44355", - "right": "issue:46877", + "left": "issue:36683", + "right": "issue:44625", "accept": false, - "reason": "Compiled Python import errors and auto_docstring import-time logging are unrelated." + "reason": "Gemma3Config missing vocab_size vs Qwen3.5 num_labels propagation; unrelated config regressions." }, { "left": "issue:29870", "right": "issue:36296", "accept": false, - "reason": "Bart generate decoder_attention_mask bug vs tensor-parallel training bug; different subsystems." + "reason": "BART generate decoder_attention_mask UnboundLocalError vs tensor-parallel training bug; no shared concrete code path." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Distributed optimizer-state corruption vs a cache strategy typo." + "reason": "FSDP2/PEFT optimizer-state corruption vs a typo in cache strategy docs/code; not the same bug." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache update decoration and broad import failures are not the same bug." + "reason": "no_grad decoration for cache update vs broad import failures after upgrade; unrelated." }, { "left": "issue:44075", "right": "issue:46365", "accept": false, - "reason": "Optimizer argument handling vs Trainer import shadowing by datasets; unrelated." + "reason": "SGD optimizer args ignored vs inability to import Trainer until datasets is removed; different components." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Both are Qwen3.5 from_pretrained issues, but the missing embed_tokens.weight bug is distinct from the generic MoE load error." + "reason": "UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error; separate model loading bugs." }, { "left": "issue:44908", "right": "issue:46851", "accept": false, - "reason": "Scheduler kwargs omission vs variable-length position_ids handling; unrelated." + "reason": "Scheduler kwargs bug vs Qwen3.5 variable-length position_ids failure; unrelated." }, { "left": "issue:29870", "right": "issue:33453", "accept": false, - "reason": "Generation-time Bart decoder bug vs tokenizer loading regression." + "reason": "BART generate error vs tokenizer loading regression; both are regressions but not the same bug." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 load failure vs PretrainedConfig type-checking regression." + "reason": "Python 3.14 import/load failure vs PretrainedConfig type-checking breakage in v5.4.0; different failures." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled Python import errors and a torch.jit decorator parsing issue are different failure modes." + "reason": "Compiled-file import errors vs DebertaV2Model import failure from torch.jit.script parsing; different root causes." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "Both mention roundtrip/loading, but one is state_dict ordering and the other is Gemma pipeline token roundtrip." + "reason": "State_dict load ordering issue vs Gemma pipeline roundtrip failure on \u2581; unrelated serialization problems." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs Trainer FSDP2/PEFT optimizer-state issue; different layers of the stack." } ] }, @@ -17709,13 +17653,13 @@ "issue:44556", "issue:44734", "issue:44749", - "issue:44961", "issue:44964", "issue:45125", "issue:45357", "issue:45375", "issue:45440", "issue:45478", + "issue:45561", "issue:45600", "issue:45850", "issue:46132", @@ -17724,6 +17668,8 @@ "issue:46664" ], "soft_pairs": [ + "issue:43867|issue:45375", + "issue:42890|issue:45561", "issue:43010|issue:45850", "issue:36296|issue:43575", "issue:41669|issue:44246", @@ -17745,8 +17691,8 @@ "issue:43824|issue:46617", "issue:43502|issue:44485", "issue:43874|issue:46617", - "issue:44387|issue:44556", "issue:43010|issue:43065", + "issue:44387|issue:44556", "issue:43576|issue:46132", "issue:30333|issue:33453", "issue:45125|issue:45478", @@ -17754,239 +17700,229 @@ "issue:45125|issue:45375", "issue:30990|issue:31795", "issue:42548|issue:44351", - "issue:43502|issue:44222", - "issue:43502|issue:44961" + "issue:43502|issue:44222" ], - "prepared_review_unit_hash": "a3ae8ff5ffdfc5fb2592fe468f3412bb0a2258b97dcdf9ec18eab23257f04458", + "prepared_review_unit_hash": "808c3b99c4795793dbf1ca5513341c372984fa87d472a6fdcbf3e53248033b08", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36847, - "estimated_input_tokens": 9212, - "estimated_eval_tokens": 18680 + "soft_pair_count": 33, + "serialized_chars": 37214, + "estimated_input_tokens": 9304, + "estimated_eval_tokens": 18864 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of distinct v5 regressions, import failures, and model-specific bugs. Only a couple of soft pairs look like true duplicates; the rest are related only at a high subsystem level and should stay separate.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43723 is the broadest, most central user-facing loading regression here: it targets `AutoTokenizer.from_pretrained` in v5 and best represents the cluster\u2019s repeated load/import breakages.", + "summary": "This set is mostly a loose collection of unrelated Transformers regressions: import/export errors, model-loading issues, tensor-parallel bugs, test flakes, and a few docs/behavior changes. The soft links look like keyword overlap rather than true duplicates, so I would not merge them into one cluster.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43723 is the broadest user-facing loading regression in the set and the best available anchor, though the overall cluster is too heterogeneous to have a strong true canonical.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best single issue to anchor this cluster because it is broad, clearly broken, and close to several of the other loading-related reports.", + "best_issue_reason": "43723 is the most general, central loading failure among the issues; it is the closest thing to a representative duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Different bugs: state_dict loading order vs missing Qwen3.5 vision config field." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated test issues: missing set_seed vs xdist capture-file race." + }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Cache `update()` no-grad behavior and a general \u201cv5 breaks too many models\u201d complaint are not the same bug." + "reason": "Specific no_grad cache-update bug vs a broad, vague compatibility complaint." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Both mention tensor parallelism/OOM, but one is a training bug and the other is a model load OOM; too different to merge." + "reason": "Both involve tensor parallelism, but different failures and code paths." }, { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Both point to slow `transformers` import, and the explicit `import *` cleanup is a plausible root cause of the same performance problem." + "accept": false, + "reason": "Import slowness overlap only; one is a specific import-* change, the other is generic." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Different model/code-path problems: CLIP attentions assignment vs `patchtsmixer` post-init policy." + "reason": "Different model/API regressions; no shared concrete bug." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "No shared failure mode; cache update semantics are unrelated to the Python-version compatibility issue." + "reason": "Completely different: cache update semantics vs Python version support." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Model load failure from sorted state_dict and incorrect visual encoder save keys are different serialization bugs." + "reason": "Both are loading-related, but one is state_dict ordering and the other is save_pretrained key mapping." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Different model families and different loading failures; both are from_pretrained problems but not the same one." + "reason": "Different model-specific from_pretrained failures, not one shared defect." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "ImportError for PyTorch availability vs tokenizer loading regression are unrelated." + "reason": "PyTorch-missing import error vs tokenizer loading regression; unrelated." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "Both involve CLIP, but one is a missing attentions assignment and the other is a structural API inconsistency; not the same concrete bug." + "reason": "Both are CLIP-related, but the specific regressions differ." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA/ReLU discussion and DeepSeekV3MoE implementation divergence are different code paths and symptoms." + "reason": "DSA indexer discussion vs DeepSeekV3 implementation divergence; not the same bug." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "A flaky test due to missing seeds is not the same as a model-loading failure." + "reason": "Test seeding flake vs model-load failure for Phi-4 multimodal." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "One is test determinism, the other is a CI failure in a different model; no shared root cause." + "reason": "Both are test failures, but they have different root causes." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Seed-related test flakiness and stale auto_mapping references are separate issues." + "reason": "Missing set_seed in tests vs stale auto_mappings config references." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Offline-request leakage and `AddedToken` argument duplication are unrelated." + "reason": "Network request bug under local_files_only vs AddedToken constructor argument error." }, { "left": "issue:42898", "right": "issue:46664", "accept": false, - "reason": "Tokenization-space behavior and video-frame semantics are unrelated API changes." + "reason": "Tokenization-space behavior change vs video num_frames semantics; unrelated." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "KV-cache continuation crash and trust_remote_code cache FileNotFoundError are different failure modes." + "reason": "Serve KV-cache indexing crash vs cache FileNotFoundError on trust_remote_code." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer loading and the other is checkpoint reload compatibility; too broad to treat as duplicates." + "reason": "Tokenizer load regression vs checkpoint reload compatibility after upgrade." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained behavior and missing tensor-parallel plan are separate Qwen/MoE issues." + "reason": "FP8 save_pretrained bug vs missing tensor-parallel plan; different fixes." }, { "left": "issue:43824", "right": "issue:46617", "accept": false, - "reason": "Missing top-level model import and cache FileNotFoundError are unrelated." + "reason": "Import error for a model class vs cache loading FileNotFoundError." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Local-files-only network leakage is unrelated to GLM-5 RoPE implementation." + "reason": "local_files_only networking bug vs GLM-5 RoPE discussion." }, { "left": "issue:43874", "right": "issue:46617", "accept": false, - "reason": "Missing image-patch helper and cache FileNotFoundError are different bugs." + "reason": "Missing image-patch helper vs cache FileNotFoundError; not the same bug." }, { - "left": "issue:44387", - "right": "issue:44556", + "left": "issue:43010", + "right": "issue:43065", "accept": false, - "reason": "Quantization memory growth and checkpoint reload incompatibility are not the same issue." + "reason": "Cache no_grad issue vs dummy Conv2d placeholder; unrelated." }, { - "left": "issue:43010", - "right": "issue:43065", + "left": "issue:44387", + "right": "issue:44556", "accept": false, - "reason": "Cache `update()` decoration and a dummy Conv2d in Sam3PixelDecoder are unrelated." + "reason": "Different version regressions: int4 memory growth vs checkpoint reload failure." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "Broken env command and AttentionInterface behavior change are separate API regressions." + "reason": "Broken env command vs AttentionInterface behavior change; separate issues." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLFlow reporting on failure and tokenizer loading regression are unrelated." + "reason": "MLflow reporting bug vs tokenizer loading regression; no overlap." }, { "left": "issue:45125", "right": "issue:45478", - "accept": true, - "reason": "Both are Qwen3.5 MoE loading/tensor-parallel regressions; the missing `_tp_plan` is a plausible direct cause of the reported from_pretrained error." + "accept": false, + "reason": "Related model family, but not enough to say it is the same concrete bug." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python-file errors and a Chinese report about slower filtering after upgrade are not the same bug." - }, - { - "left": "issue:45125", - "right": "issue:45375", - "accept": false, - "reason": "Missing tensor-parallel plan and missing `deepstack_visual_indexes` are different Qwen3.5 vision/MoE fields." - }, - { - "left": "issue:30990", - "right": "issue:31795", - "accept": false, - "reason": "Sentence Transformers loading and forward-doc confusion are unrelated." + "reason": "Compiled-file errors vs slowdown after upgrade; unrelated." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different fixes." + "reason": "Different missing top-level imports for different symbols." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Offline request leakage and FP8 save_pretrained behavior are different bugs." - }, - { - "left": "issue:43502", - "right": "issue:44961", - "accept": false, - "reason": "No meaningful relation between offline networking behavior and a placeholder issue." + "reason": "local_files_only network bug vs FP8 save_pretrained issue; unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The cluster summary is grounded in the packet: it accurately characterizes most items as separate loading/import/model-specific regressions, and the two accepted soft pairs are phrased conservatively enough to look mergeable into single bugfix PRs. I don\u2019t see an overstatement of evidence in the canonical choice or the reject/accept calls." - }, - "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 }, @@ -18018,12 +17954,12 @@ "issue:43299", "issue:43352", "issue:43475", + "issue:43502", "issue:43576", "issue:43673", "issue:43824", "issue:43827", "issue:43901", - "issue:44038", "issue:44222", "issue:44230", "issue:44273", @@ -18039,7 +17975,6 @@ "issue:44961", "issue:45003", "issue:45071", - "issue:45125", "issue:45230", "issue:45362", "issue:45468", @@ -18053,9 +17988,10 @@ "issue:46726" ], "soft_pairs": [ + "issue:43502|issue:44961", "issue:46664|issue:46726", - "issue:43010|issue:43232", "issue:43827|issue:44961", + "issue:43010|issue:43232", "issue:30333|issue:36296", "issue:29870|issue:31795", "issue:43299|issue:44556", @@ -18064,8 +18000,8 @@ "issue:43576|issue:45071", "issue:42757|issue:43576", "issue:44315|issue:44360", - "issue:43576|issue:43824", "issue:31795|issue:38617", + "issue:43576|issue:43824", "issue:44961|issue:45468", "issue:44485|issue:44961", "issue:44351|issue:46617", @@ -18074,8 +18010,8 @@ "issue:43576|issue:44938", "issue:43475|issue:46617", "issue:46143|issue:46144", - "issue:44315|issue:46209", "issue:44222|issue:44230", + "issue:44315|issue:46209", "issue:41628|issue:42548", "issue:43673|issue:44492", "issue:44485|issue:46664", @@ -18083,229 +18019,234 @@ "issue:44273|issue:45542", "issue:44623|issue:45230", "issue:44273|issue:45003", - "issue:44360|issue:46664", - "issue:44038|issue:45125" + "issue:44360|issue:46664" ], - "prepared_review_unit_hash": "cbc30e8b0ac3faed2e597dbf972188c9e6db473dd64300f6504dd009e4987dc0", + "prepared_review_unit_hash": "008c8bef7d00ca82705fc89a173399a3512d786c1cfca4e9be80098d0915021d", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 32, - "serialized_chars": 36400, - "estimated_input_tokens": 9100, - "estimated_eval_tokens": 18456 + "serialized_chars": 35851, + "estimated_input_tokens": 8963, + "estimated_eval_tokens": 18182 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of mostly unrelated Transformers issues: generation/cache bugs, import/load regressions, documentation mismatches, and model-specific failures. None of the soft pairs look like the same underlying bug or change.", - "confidence": 0.92, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: most items are unrelated loading/import/regression/docs/model-specific issues, so the soft edges do not look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue is a reliable canonical because the set spans unrelated bugs and doc regressions with different code paths and symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good canonical representative because the cluster is heterogeneous; the closest broad item is still too generic to subsume the others.", + "best_issue_reason": "No issue is a strong global representative for deduping; the cluster does not form one coherent underlying bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46664", - "right": "issue:46726", + "left": "issue:43502", + "right": "issue:44961", "accept": false, - "reason": "Both are vision/processor-related, but one is a SmolVLM video-frames semantic issue and the other is an AutoProcessor backend-missing import/message; different failure modes." + "reason": "Local-files-only/network behavior and an unrelated placeholder issue are not the same bug." }, { - "left": "issue:43010", - "right": "issue:43232", + "left": "issue:46664", + "right": "issue:46726", "accept": false, - "reason": "Both touch generation/cache internals, but one is about @torch.no_grad on cache updates and the other is about _update_model_kwargs_for_generation after sync_gpus; not the same bug." + "reason": "Video frame-count semantics and missing-vision-backend import handling are different processor problems." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Unrelated: one is a v5 docs cleanup about pipeline() removals, the other is a nonspecific issue titled 'racoon' with no clear technical overlap." + "reason": "Docs about pipeline removals vs an unrelated issue; no shared code-path bug." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache no-grad decoration and generation kwargs sync_gpus ordering are distinct generation issues." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLFlow reporting on failed jobs and tensor-parallel training bugs are distinct subsystems and symptoms." + "reason": "MLflow failure reporting and tensor-parallel training are unrelated." }, { "left": "issue:29870", "right": "issue:31795", "accept": false, - "reason": "One is a concrete generate() UnboundLocalError in BART, the other is documentation confusion about forward arguments." + "reason": "A generate() UnboundLocalError is not the same as confusing forward docs." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both mention v5 model loading, but one is a Qwen3VL MoE regression and the other is a checkpoint reload compatibility issue; too different to merge." + "reason": "Both are loading regressions, but for different versions/models and different failure modes." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading regressions, but they affect different model/tokenizer paths and have different reported breakages." + "reason": "Tokenizer loading regression and Llama 3.2 model-loading breakage are not the same concrete defect." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled Python-file errors and PretrainedConfig type-checking regressions are unrelated." + "reason": "Compiled-Python-file loading errors and PretrainedConfig type checking are unrelated." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Both are v5 breakages, but one is the transformers env command and the other is PretrainedConfig typing; different code paths." + "reason": "Broken env CLI and config type-checking are different code paths." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Importing is_offline_mode from huggingface_hub is a dependency/import API issue, not the transformers env command bug." + "reason": "Missing huggingface_hub symbol import and the env command failure are not clearly the same bug." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application in model_init and a DSA ReLU discussion are unrelated." + "reason": "Liger Kernel/model_init behavior and the DSA ReLU discussion are unrelated." }, { - "left": "issue:43576", - "right": "issue:43824", + "left": "issue:31795", + "right": "issue:38617", "accept": false, - "reason": "Transformers env command failure and missing Qwen2_5_VLForConditionalGeneration import are different regressions." + "reason": "Forward-doc confusion and an ImportError for layer_type_validation are different issues." }, { - "left": "issue:31795", - "right": "issue:38617", + "left": "issue:43576", + "right": "issue:43824", "accept": false, - "reason": "Docs confusion about model.forward args is not the same as an ImportError for layer_type_validation." + "reason": "CLI env breakage and missing Qwen2_5_VL export are unrelated." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No meaningful overlap: 'racoon' is not a technical report, while the Gemma-4 issue is a specific audio positional encoding bug." + "reason": "No meaningful shared bug; the titles do not describe the same failure." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "A GLM-5 RoPE implementation discussion is unrelated to the 'racoon' issue." + "reason": "RoPE implementation discussion and an unrelated placeholder issue are not duplicates." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "HybridCache import failure and trust_remote_code cache FileNotFoundError are both loading-related but arise from different mechanisms." + "reason": "Missing HybridCache export and trust_remote_code cache FileNotFoundError are different loading failures." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training and a docs note about return_all_scores are unrelated." + "reason": "Tensor-parallel training and pipeline docs behavior are unrelated." }, { "left": "issue:43352", "right": "issue:44961", "accept": false, - "reason": "Flash Attention support for Nemotron is a concrete model capability issue, not related to 'racoon'." + "reason": "Flash-Attention support for Nemotron and an unrelated placeholder issue are not the same bug." }, { "left": "issue:43576", "right": "issue:44938", "accept": false, - "reason": "The env command bug and Python 3.14 load failures are different environment regressions." + "reason": "Broken env CLI and Python 3.14 import/load failure are different problems." }, { "left": "issue:43475", "right": "issue:46617", "accept": false, - "reason": "SAM 3 Video missing an attribute and cache loading FileNotFoundError are different bugs." + "reason": "SAM3 attribute failure and cache FileNotFoundError do not share a concrete code-path bug." }, { "left": "issue:46143", "right": "issue:46144", "accept": false, - "reason": "Closely related RoFormer issues, but one is about kwargs not being forwarded and the other is about attention-interface usage; not the same concrete defect." + "reason": "Both touch RoFormer, but one is kwargs propagation and the other is attention-interface usage; related but not one underlying defect." }, { - "left": "issue:44315", - "right": "issue:46209", + "left": "issue:44222", + "right": "issue:44230", "accept": false, - "reason": "Liger Kernel/model_init and FP8 inference on B200 are unrelated." + "reason": "FP8 save_pretrained behavior and FP8 inference support are different operations." }, { - "left": "issue:44222", - "right": "issue:44230", + "left": "issue:44315", + "right": "issue:46209", "accept": false, - "reason": "Both mention FP8 and MoE/Qwen, but one is about save_pretrained and the other about inference support; different fixes." + "reason": "Model_init/Liger issues and B200 FP8 inference failure are unrelated." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing imports: AutoImageProcessor vs PreTrainedModel." + "reason": "Both are import errors from transformers, but for different symbols and likely different export regressions." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill and a typo in cache strategies are not the same issue." + "reason": "Generation cache missing and a typo in cache strategies are not the same bug." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "RoPE implementation discussion and a video processor frame-count semantics bug are unrelated." + "reason": "GLM-5 RoPE implementation and SmolVLM num_frames semantics are unrelated." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "A specific Qwen3.5 chat crash is not the same as a broad complaint that v5 breaks many models." + "reason": "A specific Qwen3.5 chat crash and a broad complaint about many models are not the same concrete issue." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading behavior and undefined tf backend errors are different subsystems." + "reason": "Lazy loading and tensorboard-without-TensorFlow backend errors are different failures." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Missing processor.save_pretrained files and a generic 'Bug report' are not a duplicate pair." + "reason": "Processor.save_pretrained missing files and an empty bug report are unrelated." }, { - "left": "issue:44273", - "right": "issue:45003", + "left": "issue:43824", + "right": "issue:44961", "accept": false, - "reason": "Lazy loading and unsafe sys.modules access are different implementation problems." + "reason": "Missing Qwen2_5_VL export and an unrelated placeholder issue are not duplicates." }, { - "left": "issue:44360", - "right": "issue:46664", + "left": "issue:45003", + "right": "issue:44355", "accept": false, - "reason": "A DSA indexer ReLU discussion is unrelated to a video processor num_frames bug." + "reason": "Unsafe sys.modules access and compiled-Python-file load errors are different underlying problems." }, { - "left": "issue:44038", - "right": "issue:45125", + "left": "issue:44360", + "right": "issue:46664", "accept": false, - "reason": "Both concern Qwen3/Qwen3.5 MoE, but one is a loading breakage and the other is a missing tensor-parallel plan; different concrete code paths." + "reason": "A DSA/ReLU discussion and a video processor num_frames bug are unrelated." } ] }, @@ -18331,7 +18272,6 @@ "nodes": [ "issue:31795", "issue:33453", - "issue:34689", "issue:41669", "issue:42371", "issue:42890", @@ -18342,8 +18282,8 @@ "issue:43572", "issue:43576", "issue:43673", - "issue:43704", "issue:43976", + "issue:44038", "issue:44077", "issue:44230", "issue:44246", @@ -18364,6 +18304,7 @@ "issue:44908", "issue:45071", "issue:45092", + "issue:45125", "issue:45230", "issue:45278", "issue:45310", @@ -18379,9 +18320,10 @@ "issue:46894" ], "soft_pairs": [ + "issue:44038|issue:45125", "issue:44393|issue:45071", - "issue:44393|issue:45753", "issue:41669|issue:43673", + "issue:44393|issue:45753", "issue:43576|issue:44556", "issue:43525|issue:45850", "issue:43572|issue:45850", @@ -18400,8 +18342,8 @@ "issue:44908|issue:45071", "issue:44295|issue:45230", "issue:44492|issue:46328", - "issue:44704|issue:44908", "issue:44297|issue:45230", + "issue:44704|issue:44908", "issue:41669|issue:44351", "issue:41669|issue:45341", "issue:43976|issue:45850", @@ -18417,312 +18359,288 @@ "issue:45468|issue:46894", "issue:45230|issue:45310", "issue:44655|issue:44861", - "issue:43576|issue:46218", "issue:43519|issue:44485", - "issue:43519|issue:44908", - "issue:41669|issue:43704", - "issue:31795|issue:34689", - "issue:43704|issue:44485" + "issue:43576|issue:46218", + "issue:43519|issue:44908" ], - "prepared_review_unit_hash": "a479be4880db55ecb0f666d1c53967187f6a6855b1c56af0e49e1b00e28e558d", + "prepared_review_unit_hash": "257d8464f74cd5ca33c3b7de0bee042c4ba4975dcca9da90399d6a6044cfa4cb", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 44, - "serialized_chars": 40292, - "estimated_input_tokens": 10073, - "estimated_eval_tokens": 20402 + "soft_pair_count": 42, + "serialized_chars": 39670, + "estimated_input_tokens": 9918, + "estimated_eval_tokens": 20092 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose theme cluster around Transformers v5/v5.4+ regressions, model/config loading failures, and import/lazy-loading issues, but the individual reports describe different code paths and are not true duplicates.", - "confidence": 0.73, - "canonical_issue_reason": "issue:45278 is the best anchor because it is a broad, concrete upgrade-regression report (many import errors after moving from 4.57.0 to 5.5.0) and sits near the center of the shared v5 breakage theme.", + "summary": "This cluster is mostly a grab bag of Transformers v5 regressions and model-specific bugs, with only one clear duplicate-like pair around import/lazy-loading slowdown. No pull requests are present in the cluster.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 44273 is the clearest specific report in the import/lazy-loading subcluster and is the best duplicate target among the provided issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most suitable representative issue for deduping this cluster: it is specific enough to be actionable, yet broad enough to cover the common version-upgrade failure pattern seen across the group.", + "best_issue_reason": "Issue 45278 is the strongest overall umbrella for the cluster because it captures the broad post-upgrade import/load breakage theme, even though it is still broader than the individual bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44393", - "right": "issue:45071", + "left": "issue:44038", + "right": "issue:45125", "accept": false, - "reason": "Different problems: Qwen3-VL 2D box output bug vs PretrainedConfig type-checking regression." + "reason": "Different Qwen3.5-related problems: one is a model/support bug, the other is missing tensor-parallel metadata." }, { "left": "issue:44393", - "right": "issue:45753", + "right": "issue:45071", "accept": false, - "reason": "Different failure modes: bounding-box hallucination vs infinite loop on a specific image." + "reason": "Different code paths and symptoms: Qwen3-VL 2D box output issue vs PretrainedConfig type-checking regression." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "One is import-time overhead from import-* usage; the other is chunked_prefill cache absence." + "reason": "Import-time performance from wildcard imports is not the same bug as missing GenerationMixin cache during chunked prefill." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but not the same concrete bug: bbox output error vs infinite loop on one image." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Different symptoms: broken `transformers env` command vs checkpoint reload incompatibility." + "reason": "CLI/environment command breakage is unrelated to checkpoint reload incompatibility after v5 upgrade." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Llama4Config missing `pad_token_id` is a specific config bug, not the generic v5 breakage complaint." + "reason": "Specific Llama4 pad_token_id regression is not the same as a generic 'v5 breaks many models' complaint." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "StableLmConfig missing `pad_token_idx` is a separate config-field regression." + "reason": "Specific StableLm pad_token_idx regression is not the same underlying issue as the broad v5 breakage report." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are v5 compatibility issues, but they affect different loading paths and models." + "reason": "Both are v5 compatibility complaints, but they point to different concrete failures and code paths." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "Cache-strategy typo is unrelated to the broad model-breakage report." + "reason": "A typo in cache strategies is unrelated to a broad version-breakage report." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Same general upgrade theme, but not the same underlying bug or code path." + "reason": "Both are broad v5 complaints, but one is specifically about import errors after upgrade and the other is a generic statement." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 settings API guidance is unrelated to a cache-strategy typo." + "reason": "TF32 API guidance is unrelated to the cache-strategy typo." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-star cleanup and a specific missing `Qwen3_5ForCausalLM` import are different issues." + "reason": "Import performance regression is different from a specific missing model import/module error." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type warning vs Liger Kernel not applied with `model_init`; different behaviors." + "reason": "Wrong-model warning and Liger Kernel application under model_init are unrelated issues." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 model support request vs deepseekv4 weight-conversion question are unrelated." + "reason": "FP8 inference support request and weight-conversion-script question are different feature areas." }, { "left": "issue:44246", "right": "issue:44273", - "accept": false, - "reason": "Import-time slowness and lazy-loading malfunction are related only broadly, not the same bug." + "accept": true, + "reason": "Both describe the same lazy-loading/import slowdown problem; one reports the symptom and the other names the underlying lazy-loading failure." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test seed flakiness is unrelated to scheduler kwargs handling." + "reason": "Test flakiness from missing seeds is unrelated to inverse_sqrt scheduler kwargs handling." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken env command and scheduler kwargs ignoring are different subsystems." + "reason": "Broken env CLI and scheduler kwargs are unrelated bugs." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash-Attention support gap is unrelated to Qwen3-VL bounding-box output errors." + "reason": "Flash Attention 2 support gating for Nemotron is unrelated to Qwen3-VL bbox output errors." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version support regression is not the same as a Qwen3-VL hallucination bug." + "reason": "Python-version compatibility and Qwen3-VL hallucination/bbox errors are unrelated." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs handling and PretrainedConfig type checking are unrelated." + "reason": "Scheduler kwargs handling is unrelated to PretrainedConfig type checking." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Position_ids buffer access bug and a generic bug report are not duplicates." + "reason": "Specific position_ids buffer access bug is not the same as a generic bug report." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance issue are different code paths." + "reason": "Cache strategy typo and SequenceFeatureExtractor.pad inefficiency are different issues." }, { - "left": "issue:44704", - "right": "issue:44908", + "left": "issue:44297", + "right": "issue:45230", "accept": false, - "reason": "AutoProcessor kwargs forwarding and scheduler kwargs forwarding are different APIs." + "reason": "Tokenizer save_pretrained mismatch is unrelated to a generic bug report." }, { - "left": "issue:44297", - "right": "issue:45230", + "left": "issue:44704", + "right": "issue:44908", "accept": false, - "reason": "Tokenizer class mismatch on save is unrelated to the generic bug report." + "reason": "AutoProcessor kwargs passthrough and scheduler kwargs are unrelated." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star cleanup and missing `HybridCache` export are not the same bug." + "reason": "Wildcard-import slowdown is not the same as a HybridCache import failure." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-star cleanup and a testing_utils typo are unrelated." + "reason": "Import performance regression is unrelated to a testing_utils bug." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python support regression is broader release compatibility, not the same as generic model breakage." + "reason": "Python-version compatibility is not the same as a generic 'many models break' complaint." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python version support and missing `Qwen3_5ForCausalLM` import are different problems." + "reason": "Python-version compatibility is unrelated to a missing Qwen3_5ForCausalLM module import." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different model families and failures: Qwen3-VL bbox issue vs Qwen3.5 MoE from_pretrained error." + "reason": "Qwen3-VL bbox hallucination and Qwen3.5 Moe from_pretrained failure are different bugs." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "Optional `post_init` policy and `_tied_weight_keys` AttributeError are unrelated." + "reason": "PatchTSMixer post_init validation and tied-weight-key AttributeError are different code paths." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion vs tokenizer-loading regression are not duplicates." + "reason": "Documentation confusion about inputs/cache is unrelated to tokenizer loading regression." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env command and lazy-loading malfunction are different symptoms." + "reason": "Broken env CLI and lazy-loading/import slowdown are different bugs." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application bug and flash_attention_3 training degeneration are different code-path failures." + "reason": "Liger Kernel application under model_init is unrelated to flash_attention_3 degenerate training." }, { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Both involve processors, but timestamp calculation and video padding math are distinct bugs." + "reason": "Qwen3VL timestamp calculation and SmolVLM video padding are different processor bugs." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer discussion is unrelated to Gemma-4 audio positional encoding." - }, - { - "left": "issue:44273", - "right": "issue:44908", - "accept": false, - "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated." - }, - { - "left": "issue:45468", - "right": "issue:46894", - "accept": false, - "reason": "Gemma-4 audio positional encoding and CohereASR training-loss shifting are different model bugs." + "reason": "DSA indexer/ReLU discussion and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report and Qwen3.5 MoE from_pretrained error are not the same underlying issue." + "reason": "Generic bug report and Qwen3.5 Moe from_pretrained regression are not duplicates." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Pipeline save_pretrained failure and tied-weight key AttributeError are separate problems." - }, - { - "left": "issue:43576", - "right": "issue:46218", - "accept": false, - "reason": "Broken env command and missing Qwen3_5 module import are different failures." + "reason": "Pipeline save_pretrained failure is unrelated to tied-weight-key AttributeError." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp math and GLM-5 RoPE implementation are unrelated." - }, - { - "left": "issue:43519", - "right": "issue:44908", - "accept": false, - "reason": "Timestamp calculation bug and scheduler kwargs bug are different subsystems." - }, - { - "left": "issue:41669", - "right": "issue:43704", - "accept": false, - "reason": "Import-star cleanup and multi-thread VRAM leak are unrelated." + "reason": "Timestamp calculation in Qwen3VL Processor is unrelated to GLM-5 RoPE implementation." }, { - "left": "issue:31795", - "right": "issue:34689", + "left": "issue:43576", + "right": "issue:46218", "accept": false, - "reason": "Documentation confusion and Llama 3.2 Vision model-loading regression are not the same bug." + "reason": "Env command breakage and missing Qwen3_5ForCausalLM import are unrelated." }, { - "left": "issue:43704", - "right": "issue:44485", + "left": "issue:43519", + "right": "issue:44908", "accept": false, - "reason": "VRAM leak in multi-thread use and GLM-5 RoPE implementation are unrelated." + "reason": "Processor timestamp logic is unrelated to inverse_sqrt scheduler kwargs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The chosen umbrella issue and the lazy-loading/import duplicate pair are both supported by the titles and soft-pair evidence, and the note that there are no PRs is correct." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -18744,9 +18662,11 @@ "issue:30333", "issue:31515", "issue:31795", + "issue:34689", "issue:36296", "issue:37934", "issue:38617", + "issue:41669", "issue:42371", "issue:42913", "issue:42936", @@ -18787,6 +18707,9 @@ "issue:46218" ], "soft_pairs": [ + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:43704|issue:44485", "issue:46018|issue:46172", "issue:44485|issue:45468", "issue:42371|issue:43704", @@ -18831,286 +18754,310 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "3047b7939abeff0f123d6deb29cf55d2fcc70d044f6ee4744ab9d2868605ec59", + "prepared_review_unit_hash": "6432d40ee442d1f271ca9a99550eff4c416d4b24fdcc27d6f288ac417db9b82c", "component_budget": { "node_count": 503, "item_count": 503, "soft_pair_count": 1851, - "serialized_chars": 960124, - "estimated_input_tokens": 240031, - "estimated_eval_tokens": 480318 + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 }, "budget": { - "node_count": 46, - "item_count": 46, - "soft_pair_count": 43, - "serialized_chars": 38567, - "estimated_input_tokens": 9642, - "estimated_eval_tokens": 19540 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 46, + "serialized_chars": 40596, + "estimated_input_tokens": 10149, + "estimated_eval_tokens": 20554 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft-similarity links mostly connect unrelated issues across different models, features, and regressions. I don\u2019t see a single true duplicate target, so I\u2019d avoid collapsing them into one canonical issue.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "This set is mostly a loose similarity cluster, not a true duplicate group. The strongest common thread is broad Transformers 5.x regression/compatibility complaints, but the items span unrelated model bugs, import errors, docs, and support requests, so I would not merge them as duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:45850 is the broadest umbrella (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the shared regression theme, though the cluster is too heterogeneous for a strict duplicate canonical.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45850 is the most representative high-level regression report in the set, but it is still too vague to serve as a true dedupe anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46018", - "right": "issue:46172", + "left": "issue:41669", + "right": "issue:43704", "accept": false, - "reason": "Both mention DeepSeek-V4, but one is about shared-expert gating and the other a specific tensor operation in the model code; they are not clearly the same bug." + "reason": "Import-* performance regression and multi-thread VRAM leak are different failure modes." }, { - "left": "issue:44485", - "right": "issue:45468", + "left": "issue:31795", + "right": "issue:34689", "accept": false, - "reason": "Different models and different code paths: GLM-5 RoPE implementation vs Gemma-4 audio positional encoding." + "reason": "Documentation confusion is not the same bug as a model-loading breakage." }, { - "left": "issue:42371", - "right": "issue:43704", + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in dataloader threads and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:46018", + "right": "issue:46172", "accept": false, - "reason": "TF32 configuration behavior and a VRAM leak in Qwen3ForCausalLM are unrelated issues." + "reason": "Shared-expert gating logic and DeepSeek-V4 weight transposition are different code paths." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "RoPE implementation discussion and Gemma-4 audio positional encoding are not the same issue." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API guidance and a VRAM leak bug do not share the same root cause." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only API requests and Qwen3VL timestamp calculation are separate problems." + "reason": "Local-files-only network requests and Qwen3VL timestamp math are unrelated." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad v5 regression complaint is not the same as an import failure for Qwen3_5ForCausalLM." + "reason": "A broad v5 regression report is not the same as a specific Qwen3_5 import failure." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation in a processor and missing files from save_pretrained do not describe the same bug." + "reason": "Timestamp calculation and missing save_pretrained files are different processor bugs." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "GlmMoeDsaIndexer torch.split behavior and Gemma-4 audio positional encoding are unrelated." + "reason": "torch.split output handling and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch in bark-small is unrelated to Gemma-4 audio positional encoding." + "reason": "CPU offload device mismatch is a different bug from audio positional encoding." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "A timestamp bug in Qwen3VL and a VRAM leak in multi-threaded Qwen3ForCausalLM are different failures." + "reason": "Timestamp calculation and multi-thread VRAM leak do not match." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching for audio models and DeepSeek-V4 weight conversion are unrelated." + "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are different requests." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Both are import errors, but they concern different symbols/modules and different underlying breakages." + "reason": "Both are import failures, but for different missing symbols/modules and not one concrete bug." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash-Attention support for NemotronHForCausalLM is a feature gap, not the same issue as a Qwen3_5 import failure." + "reason": "Flash Attention support for NemotronH and Qwen3_5 import failure are unrelated." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both touch tokenizer v5 behavior, but one is a broad behavior-change report and the other is a specific decoder-type bug; not clearly one bug." + "reason": "Both mention tokenizer v5, but one is a general behavior change and the other a specific decoder-type bug." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "Both mention Flash Attention 2.0, but one is specific to GPT2LMHeadModel and the other is a vague general request; not enough to treat as the same underlying change." + "reason": "Specific GPT2 flash-attention support and a generic Flash Attention 2.0 request are not the same concrete issue." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs handling and an unrelated 'Unemployment' issue are not the same." + "reason": "Scheduler kwargs handling and the unrelated 'Unemployment' issue are clearly different." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A typo in cache strategies is unrelated to the placeholder-like 'Unemployment' issue." + "reason": "Cache-strategy typo and the unrelated 'Unemployment' issue are not duplicates." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions assignment and 'Unemployment' are unrelated." + "reason": "CLIP attentions regression and the unrelated 'Unemployment' issue are different." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "A nonsensical title and a broad v5 regression report do not match the same bug." + "reason": "The unrelated 'Unemployment' issue is not the same as the broad v5 breakage report." }, { "left": "issue:44938", "right": "issue:44998", "accept": false, - "reason": "Python 3.14 import failure and 'Unemployment' are unrelated." + "reason": "Python 3.14 load failure and the unrelated 'Unemployment' issue do not match." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "_get_tied_weight_keys list handling and chat-template crashes are different code paths." + "reason": "Tied-weight key crash and chat-template tool-call crash are different bugs." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "from_pretrained slowness and a missing layer_type_validation import are unrelated." + "reason": "Slow from_pretrained loading and a missing import symbol are unrelated." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and tied-weight key handling are unrelated." + "reason": "Missing config pad token index and tied-weight key crash are different problems." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is not the same issue as _get_tied_weight_keys crashing." + "reason": "Missing Llama4Config pad_token_id and tied-weight key crash are not the same issue." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support and processor.save_pretrained missing files are unrelated." + "reason": "FP8 support request and missing processor.save_pretrained files are unrelated." }, { "left": "issue:28348", "right": "issue:43352", "accept": false, - "reason": "Both are Flash Attention 2.0 requests, but they target different models and are not the same concrete change." + "reason": "Both mention Flash Attention 2.0, but they target different models and code paths." }, { "left": "issue:29870", "right": "issue:31515", "accept": false, - "reason": "A BART generate() mask bug and slow checkpoint loading are unrelated." + "reason": "BART generate() mask bug and slow checkpoint loading are unrelated." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting on training failure and from_pretrained slowness are unrelated." + "reason": "MLFlow reporting on failure and slow from_pretrained loading are different concerns." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint load performance and tensor-parallel training bugs are different problems." + "reason": "Checkpoint loading speed and tensor-parallel training bug are unrelated." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint loading performance and forward() documentation confusion are unrelated." + "reason": "Loading performance and documentation clarity are not the same bug." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 support for qwen3-vl/qwen3.5 moe inference and a from_pretrained error in Qwen3.5 Moe are different issues." + "reason": "FP8 inference support and a Qwen3.5 MoE from_pretrained error are different issues." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio-model continuous batching and FP8 support are unrelated." + "reason": "Continuous batching for audio models and FP8 support are separate feature requests." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights usage and RoFormer attention-interface usage are unrelated." + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared expert gating are unrelated." + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are different bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "A ModernBERT type-hint issue and a tokenizer regex warning regression are not the same bug." + "reason": "A config type-hint issue is not the same as a tokenizer regex warning." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value handling and kwargs not passed to loss calculation are unrelated." + "reason": "Unused ignore_value and missing kwargs in loss calculation are different code paths." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss-function kwargs passing and RoFormer attention-interface usage are different code paths." + "reason": "Loss-function kwargs handling and RoFormer attention implementation are unrelated." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2Former ignore_value handling and tokenizer regex warnings are unrelated." + "reason": "Mask2former ignore_value and tokenizer regex warnings are not duplicates." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2former ignore_value handling and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former ignore_value handling and a typo in quantization_operation are unrelated." + "reason": "Model bug and a typo in quantization-operation naming are different." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "A type-hint correction and a typo in conversion code are unrelated." + "reason": "Type-hint correctness and a typo fix are not the same issue." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Missing kwargs in loss calculation and a ModernBERT type hint are unrelated." + "reason": "Loss kwargs propagation and ModernBERT type hints are unrelated." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "A norm-layer design question and a type-hint bug are unrelated." + "reason": "Llama4 normalization semantics and ModernBERT type hints are different topics." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2Former ignore_value handling and a ModernBERT type-hint issue are unrelated." + "reason": "Mask2former ignore_value handling and ModernBERT config typing are unrelated." } ] }, @@ -19143,12 +19090,12 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "ca11fa07ad7479793e4bbedce194d3dafa6fb583d30de179080a991e1e74f4e7", + "prepared_review_unit_hash": "56160a4ca2bae4cf4f36cc3eef09302bd1586bf40f028fe05bda7618f1b672d5", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2875, + "serialized_chars": 2876, "estimated_input_tokens": 719, "estimated_eval_tokens": 1694 }, @@ -19156,7 +19103,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2875, + "serialized_chars": 2876, "estimated_input_tokens": 719, "estimated_eval_tokens": 1694 }, @@ -19165,30 +19112,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues are not duplicates of each other: one is about tokenizer truncation direction, one about ReduceLROnPlateau-like scheduling, and one about Wav2Vec2 word time offsets. The cluster looks like a false positive from lexical similarity rather than a shared underlying bug/feature request.", + "summary": "These three issues are not duplicates: one is about tokenizer truncation behavior, one requests ReduceLROnPlateau-like scheduler behavior, and one asks for Wav2Vec2 word time offsets. They are separate feature requests in different components.", "confidence": 0.98, - "canonical_issue_reason": "If a representative is needed, #10519 is the clearest concrete feature request and the only one with a well-scoped technical topic; the other two are unrelated.", + "canonical_issue_reason": "issue:10519 is the most concrete and self-contained feature request, with a clearly stated desired behavior and scope.", "canonical_pr_reason": null, - "best_issue_reason": "#10519 is the best issue to stand in for this cluster only in the weak sense that it is the most explicit and actionable request, but it should not be treated as a true canonical duplicate for the other two.", + "best_issue_reason": "issue:10519 is the best representative of the cluster because it is the most specific and actionable issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", - "right": "issue:11005", + "right": "issue:11307", "accept": false, - "reason": "Different domains and problems: tokenizer truncation behavior vs learning-rate scheduler functionality." + "reason": "Different topics and code paths: tokenizer truncation control vs Wav2Vec2 word timing metadata." }, { "left": "issue:10519", - "right": "issue:11307", + "right": "issue:11005", "accept": false, - "reason": "Unrelated requests: truncation direction vs extracting word time offsets from Wav2Vec2." + "reason": "Unrelated feature requests: truncation direction option vs learning-rate scheduler behavior." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "No shared bug or feature: scheduler behavior request vs Wav2Vec2 alignment/timing metadata." + "reason": "Separate requests in different subsystems: ReduceLROnPlateau-like scheduling vs Wav2Vec2 alignment/time offsets." } ] }, @@ -19220,7 +19167,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "37b0f5f79b328598d647b2cfe643499537d734762283c8945344a5848c709afa", + "prepared_review_unit_hash": "2a9bd14cb250ded5b293b7203536ad019b7351fe6129598f3d9b589d62920175", "component_budget": { "node_count": 3, "item_count": 3, @@ -19242,24 +19189,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is not a true duplicate set: one issue is about Tapas tokenization, while the other two concern model download/load failures. Of the two model-related issues, they are both user-facing loading problems but still appear to describe different failure modes.", - "confidence": 0.78, - "canonical_issue_reason": "issue:16618 is the broadest, most central user-visible model-loading report in the cluster; it is closer in theme to the other loading/error issue than the Tapas tokenization report.", + "summary": "Only one issue appears to be about a model-loading failure; the other two are unrelated topics (Tapas tokenization and private-model auth error messaging). No soft pair should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 16618 is the clearest representative of the model-loading problem ('Can't load the model for bert-base-uncased') and is more general than the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:16618 is the best representative for the cluster because it captures the main model-loading/problem theme and is the most general of the three.", + "best_issue_reason": "16618 best captures the underlying user-visible failure around loading a Hugging Face model; 15980 is specifically about an authentication/error-message edge case, and 13244 is about tokenizer behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Different bugs: Tapas tokenization behavior versus a model loading failure for bert-base-uncased. No evidence they share the same code path." + "reason": "Different problems: Tapas tokenization mismatch vs. model loading failure for bert-base-uncased." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Both relate to model download/load errors, but one is specifically a bad auth/error message for private models while the other is a generic inability to load bert-base-uncased; not clearly the same underlying bug." + "reason": "Related only at a high level to model download/load errors, but 15980 is specifically about private model download without login and error messaging, not the same concrete bug." } ] }, @@ -19289,12 +19236,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "c86c59177e33f9035819f5909dd984ba4ebcf0e7f1ca9570e1c8fdc6fdb9ff64", + "prepared_review_unit_hash": "72a87f88f84f9289c8b321cd757fa3edb42240cd2b6bb14565d1e263ea7a3d38", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -19302,7 +19249,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -19311,8 +19258,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate community contribution requests for different capabilities: Flash Attention 2 support vs GGUF support. They are not duplicates and should not be merged.", - "confidence": 0.98, + "summary": "These are two separate community-contribution feature requests in different areas (Flash Attention 2 vs GGUF support). They share a generic framing but not the same underlying change, so they should not be merged as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -19322,7 +19269,7 @@ "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different feature requests with different technical targets and code paths: Flash Attention 2 architecture support versus GGUF architecture support. They are related only at a broad \u2018community contribution\u2019 level, not the same underlying change." + "reason": "Not the same bug/change: one asks for Flash Attention 2 support for more architectures, the other asks for GGUF support for more architectures. They are unrelated feature requests with no shared targets/files or concrete common code path." } ] }, @@ -19387,214 +19334,214 @@ "issue:46222|issue:46931", "issue:46226|issue:46931" ], - "prepared_review_unit_hash": "21ba274cb989b740878670701cd60d97754489fb0a8b6dd436559069abf2ae62", + "prepared_review_unit_hash": "d5c1638b212a772c3447e004d9d13e31e36ab3904e1011d2ae8b29b41303c810", "component_budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15980, - "estimated_input_tokens": 3995, - "estimated_eval_tokens": 8246 + "serialized_chars": 15974, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 }, "budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15980, - "estimated_input_tokens": 3995, - "estimated_eval_tokens": 8246 + "serialized_chars": 15974, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly a set of similar i18n translation-request issues, but they are not all the same underlying request. The only clear duplicates are the exact same generic-template issues, while the language-specific translation issues should stay separate.", - "confidence": 0.76, - "canonical_issue_reason": "issue:43834 is the best representative of the main exact-title duplicate subgroup ([i18n-] Translating docs to ) and is the earliest of that set; the other items are either later copies of that template or distinct language-specific requests.", + "summary": "Mostly not duplicates: several issues are separate docs-translation requests for different languages. The only clear duplicate groups are the identical placeholder-template issues and the identical Thai pair; the rest should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43834 is the best anchor among the duplicate-like items because it is the earliest of the identical placeholder-template translation issues and sits at the center of that exact-duplicate subgroup.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43834 is the most central and earliest issue in the largest exact-match subgroup, so it is the least ambiguous representative of the cluster.", + "best_issue_reason": "issue:43834 is the most representative issue in the cluster, but only for the repeated placeholder-template translation-request subgroup, not for the language-specific requests.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language request (Turkish vs generic template); similar template text is not enough to call them duplicates." + "reason": "Different language requests (Turkish vs generic template); same theme but not the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different language request (Turkish vs generic template); this looks like a similar i18n template, not the same issue." + "reason": "Different language requests; these are separate translation tickets, not one bug/change." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish-specific issue versus generic placeholder title; too coarse to treat as the same underlying request." + "reason": "Turkish request vs generic placeholder template; related topic, but not the same issue." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different language request (Turkish vs Romanian), so these are separate translation issues." + "reason": "Different target languages and separate translation work items." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different language request (Turkish vs Spanish), not a duplicate." + "reason": "Turkish and Spanish translation requests are distinct issues." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish-specific request versus generic placeholder template; not the same issue." + "reason": "Spanish request is separate from the generic placeholder-template issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish-specific request versus generic placeholder template; similar wording only." + "reason": "Different language translation requests; not the same underlying ticket." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Different language request (Spanish vs Romanian), so these are distinct issues." + "reason": "Spanish vs Romanian translation requests are separate." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish-specific request versus generic placeholder template; not a clear duplicate." + "reason": "Spanish request vs generic template issue; same subsystem, different issue." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian-specific request versus generic placeholder template; similar theme but different underlying request." + "reason": "Romanian request and generic template issue are not the same underlying task." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian-specific request versus generic placeholder template; not enough to merge." + "reason": "Romanian request vs generic placeholder-template issue; related, but distinct." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian-specific request versus generic placeholder template; different underlying language request." + "reason": "Romanian request is separate from the repeated generic template issues." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different language request; titles/body similarity appears template-driven rather than a true duplicate." + "reason": "Romanian translation issue vs unrelated Thai-titled issue." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different language request; similar template text does not establish the same issue." + "reason": "Romanian translation issue vs unrelated Thai-titled issue." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same generic placeholder title and near-identical issue text; these look like duplicate copies of the same template issue." + "reason": "Same placeholder title and near-identical body; looks like an exact duplicate of the same translation request." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same generic placeholder title and near-identical issue text; likely duplicate template submissions." + "reason": "Same placeholder-template issue; the bodies/titles are effectively the same." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Different title/language content; similarity is too template-like to prove the same issue." + "reason": "Very similar body text, but the titles are unrelated and this does not clearly look like the same issue." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Different title/language content; not clearly the same underlying request." + "reason": "Same as above: high textual similarity, but not enough evidence that it is the same underlying issue." }, { "left": "issue:43834", "right": "issue:46931", "accept": false, - "reason": "Different title/content; the high similarity appears to come from shared template structure, not a true duplicate." + "reason": "Different Thai-language title and not clearly the same request as the placeholder template." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Exact same generic placeholder title and highly similar bodies; this is the strongest duplicate candidate in that subgroup." + "reason": "Identical placeholder-template translation issue; this is a clear duplicate." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Different title/language content; not enough evidence they are the same issue." + "reason": "Text similarity is high, but the titles point to different issues and the match is not clear enough." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Different title/language content; looks like a similar template, not a duplicate." + "reason": "High similarity alone is not enough; the issue titles are unrelated." }, { "left": "issue:44365", "right": "issue:46931", "accept": false, - "reason": "Different title/content; too risky to merge as a duplicate." + "reason": "Different Thai-titled issue; not clearly the same underlying task." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Different title/language content; similar template text is insufficient for duplicate status." + "reason": "Very similar bodies, but different titles and no clear evidence they are the same issue." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Different title/language content; not the same underlying request." + "reason": "Similar content, but not enough to treat as the same issue." }, { "left": "issue:44370", "right": "issue:46931", "accept": false, - "reason": "Different title/content; similarity is likely due to the shared issue template." + "reason": "Different Thai-titled issue; not the same request." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "Identical title and essentially identical issue content; clear duplicate pair." + "reason": "Same Thai title/body, so this is an exact duplicate pair." }, { "left": "issue:46222", "right": "issue:46931", "accept": false, - "reason": "Different title/content; similar template structure does not prove a duplicate." + "reason": "The titles differ and the evidence does not show a clear one-to-one duplicate." }, { "left": "issue:46226", "right": "issue:46931", "accept": false, - "reason": "Different title/content; not enough evidence to merge." + "reason": "Different titles and no strong proof they are the same underlying issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is conservative and grounded in the packet: the language-specific i18n issues are plausibly separate, while the exact-title placeholder pairs ([i18n-] and the identical Thai-title pair) are the only clearly duplicate-looking groups. No soft-edge verdict appears overstated." + "feedback": "Grounded and conservative overall. The summary correctly separates the language-specific translation requests from the clear duplicate subgroups, and the accepted pairs are supported by nearly/fully identical titles and bodies. No obvious overstatement beyond the evidence in the packet." }, "evaluator_used": true, "retried": false, @@ -19677,12 +19624,12 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "4a97892d8333918461f5e8364ed07eaf70d96f0beeb9f8b30334425688cac171", + "prepared_review_unit_hash": "05774a1c79042f3e14b8d00c26f65df4a79a30feb7b840f6b042ed74b82e4b6d", "component_budget": { "node_count": 19, "item_count": 19, "soft_pair_count": 40, - "serialized_chars": 24674, + "serialized_chars": 24676, "estimated_input_tokens": 6169, "estimated_eval_tokens": 12594 }, @@ -19690,7 +19637,7 @@ "node_count": 19, "item_count": 19, "soft_pair_count": 40, - "serialized_chars": 24674, + "serialized_chars": 24676, "estimated_input_tokens": 6169, "estimated_eval_tokens": 12594 }, @@ -19699,174 +19646,174 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a loose theme of model-support/request issues, not a set of true duplicates. The soft links are mostly driven by generic wording like \"Add\" or \"Support\" rather than the same concrete model or code-path change.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 46133 is the strongest representative because it is a substantive, active model-request thread with the most discussion/inbound activity in this set, but the cluster is too heterogeneous to treat as one duplicate issue.", + "summary": "These issues are all about adding or supporting different models/features, but they are not duplicates of one another; the cluster is a loose thematic group, not a single underlying request.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits: the items request distinct models or capabilities (e.g. RoMa, EuroBert, RF-DETR, Qwen3-TTS/ASR, Voxtral, etc.) rather than the same bug/change.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46133 is the best anchor for this cluster: it is detailed, active, and clearly scoped as a model-support request. Even so, it should be treated as a thematic representative, not a true duplicate canonical.", + "best_issue_reason": "None of the issues is a strong global canonical. If a representative had to be picked, issue 43671 is a concrete model-support request, but it is not a duplicate of the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36718", "right": "issue:41115", "accept": false, - "reason": "Different requests: RoMa keypoint matcher vs MiniCPM3 architecture; only share a generic model-add pattern." + "reason": "Different model families: RoMa keypoint matcher vs MiniCPM3 architecture." }, { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config addition and DEIMv2 model support are unrelated model requests." + "reason": "EuroBert config addition and DEIMv2 are unrelated model requests." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config vs Qwen3-TTS support are different model families and different kinds of changes." + "reason": "EuroBert config work is unrelated to Qwen3-TTS support." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config and RF-DETR model support are separate model additions, not the same issue." + "reason": "EuroBert and RF-DETR are different models with different integration needs." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are distinct model requests with no shared bug/change." + "reason": "RF-DETR and DEIMv2 are separate detector requests, not the same change." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR detection support and LFM2.5 Audio support target different models and modalities." + "reason": "RF-DETR and LFM2.5 Audio are different model domains." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Adding RF-DETR and adding OpenAI weight-sparse transformer support are different changes." + "reason": "RF-DETR detection support is unrelated to the weight-sparse transformer request." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR model support and Qwen3-ASR support are unrelated requests." + "reason": "RF-DETR and Qwen3-ASR are different model integrations." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are different model families and different functionality." + "reason": "RF-DETR and Qwen3-TTS are separate model requests." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are distinct model requests with no common code-path." + "reason": "NeoBERT and Qwen3-ASR are unrelated models." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT addition and a weight-sparse transformer request are different underlying changes." + "reason": "NeoBERT support and circuitgpt weight-sparse transformer are different changes." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are unrelated model-support requests." + "reason": "NeoBERT and LFM2.5 Audio target different architectures and tasks." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR model support and NeoBERT addition are separate model requests." + "reason": "RF-DETR and NeoBERT are unrelated model additions." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different model additions, not the same issue." + "reason": "NeoBERT and ModernVBERT are distinct model support requests." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture and Qwen3-ASR support are unrelated requests." + "reason": "MiniCPM3 architecture and Qwen3-ASR are not the same model or code path." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 architecture work is not the same as adding a weight-sparse transformer model." + "reason": "MiniCPM3 support is unrelated to the OpenAI sparse transformer request." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 architecture and ModernVBERT models are different model-support tasks." + "reason": "MiniCPM3 and ModernVBERT are separate model integrations." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 architecture and Voxtral TTS support are different model families and modalities." + "reason": "MiniCPM3 architecture work is unrelated to Voxtral TTS support." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + "reason": "DEIMv2 and Qwen3-TTS are different models and tasks." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR support are different model additions." + "reason": "ModernVBERT and Qwen3-ASR are different model requests." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT models and Qwen3-TTS support are not the same underlying change." + "reason": "ModernVBERT models and Qwen3-TTS are unrelated." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT model addition and a RoPE-related BERT-like discussion are different scopes." + "reason": "ModernVBERT and BERT-like models with RoPE are only broadly related by theme, not by the same underlying change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "RoPE for BERT-like models and Voxtral TTS support are unrelated changes." + "reason": "BERT-like RoPE support and Voxtral TTS are unrelated changes." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "BERT-like models with RoPE and a user-name issue are clearly unrelated." + "reason": "BERT-like model support and the unrelated person-only issue are not duplicates." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS support and Qwen3-ASR support are related by family name but are different modalities and separate requests." + "reason": "Qwen3-TTS and Qwen3-ASR are related by branding but different modalities and code paths." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support and weight-sparse transformer support are different code paths and model types." + "reason": "Qwen3-TTS support and the sparse transformer request are unrelated." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral-4B-TTS are different model-specific support requests." + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations, not one bug." }, { "left": "issue:43671", @@ -19878,73 +19825,73 @@ "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are both audio-related but not the same model or change." + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are different speech/audio models." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support and EUPE are unrelated model requests." + "reason": "Qwen3-ASR and EUPE are unrelated model additions." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support and weight-sparse transformer support are different problems." - }, - { - "left": "issue:43837", - "right": "issue:45027", - "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are separate speech-model support requests." - }, - { - "left": "issue:43837", - "right": "issue:45177", - "accept": false, - "reason": "Qwen3-ASR support and DeepSeek-OCR2 are different model additions." + "reason": "Qwen3-ASR support and circuitgpt are different features." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are unrelated model-support requests." + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model domains." }, { "left": "issue:43909", "right": "issue:46645", "accept": false, - "reason": "LFM2.5 Audio 1.5B and Granite SWA+sink option are different changes in different model families." + "reason": "LFM2.5 Audio support and SWA+sink for Granite are not the same change." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer support and EUPE model addition are not the same concrete change." + "reason": "OpenAI sparse transformer support and EUPE are unrelated model requests." }, { - "left": "issue:45027", - "right": "issue:46653", + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are distinct speech-model support requests." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated models." + }, + { + "left": "issue:43837", + "right": "issue:45177", "accept": false, - "reason": "Voxtral TTS support and the unrelated name-only issue are clearly not duplicates." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different tasks and code paths." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are both model requests, but they are distinct models and not the same underlying issue." + "reason": "EUPE and TIPSv2 are separate vision-language model requests." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT models and the unrelated name-only issue are not duplicates." + "reason": "ModernVBERT models and the unrelated person issue are not duplicates." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support and the unrelated name-only issue are not duplicates." + "reason": "Voxtral TTS support and the unrelated person issue are not duplicates." } ] }, @@ -19974,7 +19921,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "d4cfe72b759eef73e28b88485c5e3d51de6a1ba6a07e49b6c846f95be68314ce", + "prepared_review_unit_hash": "fce130c229484c5968d6420ea9c16206fce5ab6b7b8eda936c3615c24d88aa30", "component_budget": { "node_count": 2, "item_count": 2, @@ -19996,18 +19943,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is a docs request to add `pruna` as an optimization framework, and the other is a Vision RoPE refactor. They describe different areas and different kinds of work.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "The two issues are unrelated: one is a docs request about adding Pruna, and the other is a Vision RoPE refactor. The soft similarity is superficial only.", + "confidence": 0.99, + "canonical_issue_reason": "Issue 46443 is the better canonical representative because it is the substantive technical item; issue 38740 is a separate docs request and not part of the same change/problem.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 46443 is the clearest standalone issue in the cluster. Issue 38740 is unrelated and should not be used as the representative problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "No shared underlying bug or change: docs/framework listing request vs a vision-model refactor." + "reason": "Different domains and intents: documentation/vendor listing vs. a model architecture refactor. No shared code path or underlying bug/change." } ] }, @@ -20037,12 +19984,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "8168fa76fdb0da56039109666e5d3d2186f809e038ec74976c96fec52afbd845", + "prepared_review_unit_hash": "9fbe694375bc10af878582b80d7ec087fd4301715ac7e221203919df32a8921f", "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 }, @@ -20050,7 +19997,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -20059,18 +20006,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one is about block decoding, the other is a feature/support request for DeepSeek v3.2. No canonical cluster representative is evident from this pair.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue; the titles describe different underlying topics and do not point to the same bug or feature.", + "summary": "These two issues are not duplicates: one is a vague request about block decoding, while the other is a specific feature/support request for DeepSeek v3.2. They do not appear to describe the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If forced to pick a single issue as the clearer standalone report, issue:43037 is more specific and actionable, but it is not a duplicate of issue:41084.", + "best_issue_reason": "#43037 is the more specific and actionable issue, but it is not a duplicate of #41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different subjects: \"Set Block Decoding\" concerns a decoding behavior/setting, while \"DeepSeek v3.2 support\" is a model support request. They do not appear to describe the same underlying issue or change." + "reason": "The titles and contexts are unrelated: 'Set Block Decoding' is a generic decoding-control request, while 'DeepSeek v3.2 support' is a model/version support request. No clear shared bug or change." } ] }, @@ -20100,40 +20047,40 @@ "soft_pairs": [ "issue:41306|issue:41826" ], - "prepared_review_unit_hash": "fb553439431f8f0b13363706f2fadbc78d1b1b2aa14f8a8d5a12268fb0ad5d0c", + "prepared_review_unit_hash": "ac8c3f5871fdad618f17c5bd20c449c6b67fbee5c68d78320684656029bced2f", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1585, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1585, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two open issues are unrelated feature requests: one asks for a `from_pretrained()` flag for TiledFusedLogitsLoss / FLCE, while the other requests integrating TiledMLP to reduce memory footprint. They do not describe the same bug or change.", + "summary": "These two issues are not duplicates: one asks for a `from_pretrained()` flag for `TiledFusedLogitsLoss`/FLCE, while the other requests integrating `TiledMLP` for lower memory use. They share only the broad \u201cTiled\u201d naming and target different components and changes.", "confidence": 0.97, - "canonical_issue_reason": "Issue #41306 is the more concrete and implementation-specific request, centered on a single API flag and a clearly scoped component (`from_pretrained()` / FLCE).", + "canonical_issue_reason": "No true canonical duplicate here; the issues describe different features in different code paths, so neither should be collapsed into the other.", "canonical_pr_reason": null, - "best_issue_reason": "#41306 is the better canonical issue because it has a narrower, more actionable scope and is not just a broad architectural integration request.", + "best_issue_reason": "If forced to pick a representative issue for the cluster, #41306 is slightly more concrete and implementation-specific, but it should still remain separate from #41826.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41306", "right": "issue:41826", "accept": false, - "reason": "These are different feature requests in different areas: FLCE loading behavior vs adding TiledMLP for memory savings. No shared code path or same underlying change is evident." + "reason": "Reject: the titles and problem statements concern different modules and different requested changes (`FLCE` flag in `from_pretrained()` vs. integrating `TiledMLP` for memory reduction). Similar naming is insufficient to treat them as the same underlying issue." } ] }, @@ -20163,12 +20110,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "27adf8f618f9633d8542023a81c54fc8cb5f0e665ee32ec9ecca0558065ec7aa", + "prepared_review_unit_hash": "843a994792c93835f354b09d35fd506e2568a93d599e915d4f5debc3ef3dfbdd", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -20176,7 +20123,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -20185,18 +20132,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are only loosely related through custom Trainer behavior, but they describe different problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They should not be treated as duplicates.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 43598 is the broader, more directly Trainer-focused request and is the better representative of this cluster\u2019s theme.", + "summary": "These two issues are related to Trainer customization, but they describe different problems: one is about `remove_unused_column`, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "Issue #43598 is the more concrete and narrowly scoped report, centered on a specific Trainer behavior (`remove_unused_column`) and its customizability.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the best issue candidate because it centers on a concrete Trainer customization point, whereas 43600 is a narrower metric-estimation concern.", + "best_issue_reason": "#43598 is the best issue anchor because it names the exact Trainer mechanism involved and is easier to map to a concrete code change or discussion.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying problems: one concerns `remove_unused_column` handling, the other concerns FLOPS estimation for custom trainers. Shared Trainer/customization context is too broad to treat as the same bug." + "reason": "They share a broad Trainer/customization theme, but one concerns unused-column removal and the other concerns FLOPS estimation for custom trainers; these are different underlying bugs." } ] }, @@ -20226,7 +20173,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "4486a818be34a29d329e6d0939658b2cfbb58856e9368988aa5880458ddc2df5", + "prepared_review_unit_hash": "5df20f030f06c2a9755d026fd9248bdfdb69b8b743cec48cb04c99c4a137249b", "component_budget": { "node_count": 2, "item_count": 2, @@ -20248,18 +20195,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are related only at the high-level theme of Qwen model support, but they request different capabilities for different model families and different AutoModel heads. They should not be treated as duplicates.", + "summary": "The two issues are related only at a very broad level (Qwen model support in Auto classes) but describe different requests: one asks for registration of Qwen3-Omni in AutoModel/AutoModelForConditionalGeneration, while the other asks for AutoModelForSequenceClassification support for Qwen3.5. They are not the same underlying bug or change.", "confidence": 0.95, - "canonical_issue_reason": "No true canonical duplicate: #43684 is about registering Qwen3-Omni in AutoModel/AutoModelForConditionalGeneration, while #44405 asks for AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). Different models and different code paths.", + "canonical_issue_reason": "Issue #43684 is the broader and more general request, focused on AutoModel registration for a specific Qwen3-Omni model, so it is the better canonical representative of this small cluster.", "canonical_pr_reason": null, - "best_issue_reason": "#43684 is the broader and more central API-registration request, so it is the better representative issue if one must be chosen, but it does not subsume #44405.", + "best_issue_reason": "Issue #43684 is the best issue candidate because it is more general and closer to an AutoModel registration gap, whereas #44405 is a separate task for sequence-classification support on a different Qwen variant.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Reject: the requests target different model families (Qwen3-Omni vs Qwen3.5) and different auto-model heads (conditional generation vs sequence classification), so they are not the same underlying bug/change." + "reason": "Different model families/variants and different functionality: model registration for Qwen3-Omni vs sequence-classification support for Qwen3.5. They share the same ecosystem but not the same concrete fix." } ] }, @@ -20296,80 +20243,77 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "71b5ad079bc8517f9c11e6263891e5f8c71882eb9eaabbd775c15c387728ab46", + "prepared_review_unit_hash": "7b27d96e4216afe1eed2b8f25fb2d64999c8d02ba91745032d9c86942ef4cc5d", "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": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three near-identical low-information issues titled \"huggingface\" appear to be duplicates; issue #43835 titled \"Fett\" looks unrelated.", - "confidence": 0.78, - "canonical_issue_reason": "#43845 is the earliest of the repeated \"huggingface\" reports, so it is the best stand-in for the duplicate group.", + "summary": "Four closed issues with very little content; one has a distinct title (\u201cFett\u201d) and three share the same generic title (\u201chuggingface \u201d). There isn\u2019t enough concrete information to confidently merge any pair as the same underlying bug.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 43845 is the best representative of the repeated \u2018huggingface \u2019 titles and is the most central item in the apparent mini-cluster, but the available text is too sparse to prove it is a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "#43845 is the clearest canonical issue among the apparent duplicates because it is earliest and matches the repeated title exactly.", + "best_issue_reason": "43845 is the best issue to anchor the cluster because it matches the repeated title pattern and is one of the earliest of the similar issues, but the evidence is too thin for a stronger claim.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles with no shared details; no evidence they describe the same issue." + "reason": "Only weak title similarity; \u2018Fett\u2019 and \u2018huggingface \u2019 do not indicate the same underlying bug." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no body/target overlap; likely unrelated." + "reason": "Only weak title similarity; no shared details or explicit linkage show the same issue." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "\"Fett\" does not match the repeated \"huggingface\" reports and there is no shared evidence." + "reason": "Only weak title similarity; the distinct title makes a duplicate relationship unlikely." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same title and very close creation times; they look like duplicate reports of the same issue." + "accept": false, + "reason": "The titles match, but there is no body detail or linkage confirming the same concrete problem, so this is too thin to accept." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same title as the other duplicate \"huggingface\" issues and no distinguishing content is present." + "accept": false, + "reason": "Same generic title, but no substantive description or references to establish the same underlying bug." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Identical title and similarly sparse content; likely the same underlying issue." + "accept": false, + "reason": "Same generic title, but the available evidence is insufficient to confirm a duplicate relationship." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: #43845/#43846/#43859 share the same low-information title, while #43835 has a different title and no other linking evidence. The soft-edge verdicts are conservative and supported by the packet data (identical titles, close timestamps, and soft_similarity only)." - }, - "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 }, @@ -20403,88 +20347,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "2c3c7a53ff008adf35dd138e99fec7127d4378765e1ee7814c75c0449108f93c", + "prepared_review_unit_hash": "4af827ef5cc19a9b6c2b637ccca4a5d639648bec3654a0f67931e6f6cffff697", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are low-information, likely unrelated spam or placeholder issues with no concrete bug/report details. There isn\u2019t enough evidence to collapse them into a single duplicate cluster.", - "confidence": 0.9, - "canonical_issue_reason": "Issue #44371 is the earliest generic placeholder entry and is as representative as any other item in this set, but the cluster is too sparse to support a stronger canonical choice.", + "summary": "These are low-information, mostly unrelated issue reports: three identical spam placeholders, one documentation suggestion, and one branded phrase. There is no evidence they describe the same underlying bug or change, so none of the soft pairs should be merged as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44371 is the earliest of the identical \"\" entries, so it is the most reasonable canonical representative for the spam-like subset.", "canonical_pr_reason": null, - "best_issue_reason": "#44371 is the safest issue to keep as the representative item because it is earliest and has the most generic placeholder title, but none of the issues provide substantive content.", + "best_issue_reason": "issue:44371 is the best overall representative because it matches the repeated placeholder spam pattern and is the earliest of the identical titles.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are titled '', but that only indicates generic spam labeling, not the same underlying issue." + "reason": "Both are placeholder spam issues, but there is no concrete shared bug or change description to justify treating them as duplicates." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are generic spam placeholders; there is no concrete evidence they report the same problem." + "reason": "Same generic spam title, but no substantive content indicates the same underlying issue." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "'' and 'Documentation Enhancement Suggestion' are different in kind, and there is no shared report detail." + "reason": "One is spam placeholder text and the other is a documentation enhancement request; they are clearly different artifacts." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and 'Check out \"Google LLC\"' do not indicate the same underlying bug or change." + "reason": "A spam placeholder and a distinct phrase about \"Google LLC\" are not the same issue." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Documentation enhancement feedback is not the same as a spam placeholder issue." + "reason": "Documentation enhancement versus spam placeholder; no common concrete problem." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Documentation enhancement feedback is not the same as a spam placeholder issue." + "reason": "Documentation enhancement versus spam placeholder; unrelated." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but the metadata does not show a concrete shared issue to merge." + "reason": "Both are spam placeholders, but the available information is too generic to confirm the same underlying report beyond superficial similarity." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Generic spam placeholder and a company-name promo-style title are different low-information reports." + "reason": "Spam placeholder and a distinct branded phrase; no indication of the same bug or request." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Generic spam placeholder and a company-name promo-style title are different low-information reports." + "reason": "Spam placeholder and a distinct branded phrase; unrelated." } ] }, @@ -20514,12 +20458,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "78f3d5f0e1cb92b11fe4b2d0a7cf8303ee9244aba6e8594ea83b8aedda7da968", + "prepared_review_unit_hash": "a31fd907d278bd37b5386c7b5044a15e1528d1e31f63215411c5e9027bd0920d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -20527,7 +20471,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -20536,24 +20480,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues have identical titles and very similar body sizes/content, with no distinguishing links or filenames, so they are very likely duplicates of the same reported request.", + "summary": "These two closed issues appear to be duplicates of the same report, with identical titles and near-identical body size/structure. Issue #45145 is the better canonical choice because it has an inbound reference and was created earlier.", "confidence": 0.98, - "canonical_issue_reason": "Issue #45145 is the better canonical because it has an inbound reference and otherwise matches #45151 almost exactly, suggesting it is the earlier/linked copy.", + "canonical_issue_reason": "#45145 is slightly better as the canonical issue: it was created earlier and already has an inbound reference, suggesting it is the more established thread.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is a strong representative issue: same title as #45151, similar body length, and slightly more evidence of being the referenced thread.", + "best_issue_reason": "#45145 is the best representative issue for deduplication for the same reasons: earlier timestamp and more linkage/activity than #45151.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Identical title and highly similar issue text indicate the same underlying issue; no evidence suggests distinct reports." + "reason": "Accept: the issues have the same title and highly similar body characteristics, indicating the same underlying report rather than merely related topics." } ] }, "evaluator_result": { "accept": true, - "feedback": "The conclusion is grounded in the packet: the issues have identical titles, close body lengths, and a soft similarity signal, with issue #45145 also having the earlier timestamp and one inbound reference. The duplicate/representative judgment is appropriately conservative." + "feedback": "Grounded overall: the issues share the same title, have close body lengths, and the earlier issue (#45145) also has an inbound reference, so treating #45145 as canonical is reasonable. The wording is mostly conservative; the only slight caution is that \u201cbody size/structure\u201d is stronger than the packet explicitly proves, but the duplicate assessment itself is supported by the available evidence." }, "evaluator_used": true, "retried": false, @@ -20634,7 +20578,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "209e23c26fc13b7ae9385c70a537e76f034e4db0506f0adf33ac1a0b961ffd3a", + "prepared_review_unit_hash": "2fd8dd3a4ce2bf6f160770247a357d9a52e52cb6d27e557c3e31971a8fac2c11", "component_budget": { "node_count": 2, "item_count": 2, @@ -20656,18 +20600,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues appear unrelated: they are just distinct person-name/date strings with no shared technical context or description, so this is not a duplicate cluster.", - "confidence": 0.98, + "summary": "These two closed issues do not look like duplicates of the same bug or change; they appear to be separate person-name/date entries with only a shared surname.", + "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good canonical candidate; they do not describe the same bug or request and have no supporting context.", + "best_issue_reason": "Neither issue is a clear canonical duplicate target; if forced, 45762 is only a marginal representative, not because it matches 45763.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names and dates only; no evidence they report the same underlying problem or request." + "reason": "Different names and different dates; no evidence of the same underlying issue." } ] }, @@ -20697,7 +20641,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "d761bd67cd4aded20888b2315404e73eb4add6dc9baff4f4702439f84c82a94d", + "prepared_review_unit_hash": "c6ae58f6e3917949458a07d8459aa5d44251b817f5794deb0b81f0d51a46c684", "component_budget": { "node_count": 2, "item_count": 2, @@ -20719,18 +20663,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are just sparse name/date entries and do not look like the same underlying problem; the soft-similarity signal is too weak to treat them as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: there is no substantive issue content or shared problem statement, so neither item can be justified as the duplicate target.", + "summary": "The two issues are bare title-only stubs with different personal-name/date strings and no shared technical content, so they do not look like duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a strong representative, but if one must be picked, 45764 is no better or worse than 45765; both lack enough context to serve as a canonical issue.", + "best_issue_reason": "Neither issue is a good canonical target; there is no evidence of the same underlying bug or request, only unrelated name/date titles with no body.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Reject: these are just two unrelated-looking closed issues with different personal-name/date titles and no body, links, or shared technical context indicating the same bug or request." + "reason": "Reject: the titles refer to different people and dates, with no body, references, or technical overlap indicating the same issue." } ] }, @@ -20760,7 +20704,7 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "94cefa95ce53f45648aa48569ac655b5f8517bac5db31f8880b460529a0fd8d2", + "prepared_review_unit_hash": "ced237c45ae21303ebba322476c1b09c3f88d19b7eca17b51c2cfa469b025691", "component_budget": { "node_count": 2, "item_count": 2, @@ -20782,18 +20726,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated: one requests a new Qwen3_5 token-classification/value-model head, while the other asks for FP8 training support. This is not a duplicate cluster.", + "summary": "The two issues are unrelated: one requests a new Qwen3_5 token-classification/value-model head, while the other asks for FP8 training support. They do not describe the same bug or change.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No canonical issue for this pair; they target different functionality and code paths.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "45810 is the more self-contained and specific request, but it is not a true duplicate of 46295.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different underlying requests and code paths: model head support vs FP8 training support. Shared similarity is too broad to treat as the same issue." + "reason": "Different feature areas: model-head addition for token classification versus FP8 training support. They are not the same underlying change." } ] }, @@ -20833,12 +20777,12 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "121058cb8aa9676fa33fc30affcda48be1ba289adbaf6ec8b8cddc692e41ab29", + "prepared_review_unit_hash": "2334846e7b90faee21adaf922a7ade3774e74930ed9eaf0abf610e435e80a9ac", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5727, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -20846,7 +20790,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5727, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -20855,69 +20799,69 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster appears to split into two likely duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is thematically related, but it looks like a distinct dataset-related request rather than a duplicate of either pair.", - "confidence": 0.91, - "canonical_issue_reason": "No single canonical issue covers the whole cluster; the evidence supports two likely duplicate pairs, while 45869 is related but distinct.", + "summary": "The items split into two separate exact-duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is about a financial dataset bundle and is not the same underlying request as the example/service issues, so it should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "45872 is a clean exact-duplicate representative of the more specific DataForge financial time series + scraping service request; it is the clearest standalone issue among the items.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the best representative issue overall because it is the older copy in one likely duplicate pair and has the only discussion activity.", + "best_issue_reason": "45872 is the strongest issue to keep as the representative because it is an exact duplicate pair with 45873 and has the most concrete, differentiated title.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Related financial/DataForge theme, but the requests are different: dataset listing versus example/prediction use case." + "reason": "Dataset bundle vs example application; related domain, but not the same issue or change." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Same as 45869/45870: topical overlap only; not the same underlying request." + "reason": "Dataset bundle vs example application; similar financial theme, but different artifact type and intent." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Likely duplicate pair: identical title and matching body size make these look like the same request." + "reason": "Same title and same body length; looks like an exact duplicate issue." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both are financial/DataForge-related, but the concrete asks differ enough that they do not look like duplicates." + "reason": "Both are financial DataForge ideas, but one is an example and the other adds a scraping service; not the same concrete request." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Broad thematic similarity only; the request appears different from the example/prediction issue." + "reason": "Example issue vs scraping-service request; same broad area, different underlying ask." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Same reason as 45870/45872: similar domain, but different concrete request." + "reason": "Same broad financial topic, but one is an example and the other is a different service-oriented request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Related topic, but not the same underlying issue." + "reason": "Same broad topic, but not the same underlying issue; one is an example, the other is a scraping-service request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Likely duplicate pair: identical title and matching body size strongly suggest the same request." + "reason": "Same title and same body length; looks like an exact duplicate issue." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster split is plausible, but the writeup is a bit too strong for the packet. We only have titles, body lengths, and soft-similarity signals\u2014no body text or explicit links\u2014so calling 45870/45871 and 45872/45873 \u201cexact duplicates\u201d is stronger than the evidence supports. Also, 45869 is related thematically, so calling it \u201cunrelated\u201d overstates the separation. A more conservative phrasing would be \u201clikely duplicate pairs\u201d with 45869 as a related but distinct request." + "accept": true, + "feedback": "Grounded in the packet. The two accepted soft edges are supported by identical titles, identical body lengths, and jaccard 1.0, while the false edges stay conservative by distinguishing the dataset-bundle item from the example/service requests. The representative choice of 45872 is also reasonable." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -20941,7 +20885,7 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "e16667e91153091bb5a374bbe299b8bb38d95017b61268502af8312f2c00a2da", + "prepared_review_unit_hash": "9337c6ed8b3a6109cab9d1876ea3548c95c075e8a112826b467d6b885707c50d", "component_budget": { "node_count": 2, "item_count": 2, @@ -20963,18 +20907,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues both mention static cache and linear attention, but they describe different failure modes: one is a shape corruption/runtime error during early initialization, the other is a KeyError for `linear_attention` in Qwen3.5. They are related in subsystem, but not clearly the same underlying bug.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 46439 is the better canonical issue because it states a concrete mechanism (`early_initialization`) and a clearer failure mode (wrong shapes leading to `RuntimeError`) rather than a model-specific symptom.", + "summary": "Both issues concern static-cache behavior around linear attention, but they describe different failure modes: one is wrong-shape corruption during early initialization, the other is a KeyError for `linear_attention` in Qwen3.5. I would keep them separate absent stronger evidence they share the exact same root cause.", + "confidence": 0.73, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it is the more general description of the static-cache/linear-attention corruption, while 46441 is a specific model-specific crash report.", "canonical_pr_reason": null, - "best_issue_reason": "46439 is the strongest representative for this cluster since it describes the broader static-cache/linear-attention breakage in a way that could encompass multiple downstream crashes.", + "best_issue_reason": "46439 is the best issue candidate because it names the underlying mechanism and failure class more directly than the narrower Qwen3.5-specific report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Both involve static cache and linear attention, but 46439 reports shape corruption during early initialization while 46441 reports a `KeyError: 'linear_attention'` in Qwen3.5. That looks like different root causes, so I would not merge them." + "reason": "Reject: both mention static cache and linear attention, but one reports shape corruption in early initialization and the other reports a `KeyError` on `linear_attention`; these look like related symptoms, not clearly the same underlying bug." } ] }, @@ -21004,12 +20948,12 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "c960a2c36cdb79c8945534619c7227e62c1d9446fb666fc629a3e0134bb7e07f", + "prepared_review_unit_hash": "2d912bbda9b8672e316bec841fbb20eb470403f079b49ed4ec7281ef14dcda02", "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 }, @@ -21017,7 +20961,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1706, + "serialized_chars": 1707, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -21026,24 +20970,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are exact duplicates: identical titles, identical body length, and the same reported Wav2Vec2CTCTokenizer v5 special-token leak behavior.", + "summary": "The two issues are near-exact duplicates: same title, same body length, and the same tokenizer bug description. The open issue is the better canonical target.", "confidence": 0.99, - "canonical_issue_reason": "Issue #46552 is the better canonical target because it is still open and already has inbound references, making it the active discussion thread.", + "canonical_issue_reason": "Issue #46552 is the better canonical issue because it is still open and appears to be the active report for the same exact bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46552 is the best issue to keep because it is the live report for the same bug and has existing references from other items.", + "best_issue_reason": "Issue #46552 is the best issue to keep as the primary tracker: it has the same problem statement as #46551 and is open with inbound references.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Same title and same described tokenizer bug; these are the same underlying issue, so the duplicate edge should be accepted." + "reason": "Same title and same underlying Wav2Vec2CTCTokenizer bug report; these are duplicate issues for the same defect." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analyst\u2019s conclusion is well grounded in the packet: the two issues have identical titles, identical body lengths, and perfect similarity (jaccard 1.0), so treating them as duplicates is conservative and supported. The choice of issue #46552 as canonical is also justified by its open state and inbound references." + "feedback": "The packet supports the core conclusion: both issues have the same title and identical body length, and the soft-similarity signal is maximal. Treating #46552 as the canonical open tracker is reasonable, and the duplicate-pair verdict is appropriately conservative." }, "evaluator_used": true, "retried": false, @@ -21070,7 +21014,7 @@ "soft_pairs": [ "issue:46657|issue:46692" ], - "prepared_review_unit_hash": "a8e99994ad76b8c4a37973f61aff4dd55025b2160eb0a75a66304d944912da7e", + "prepared_review_unit_hash": "ff738f307664558bac3f47a38ea47d2b5cdb5ea31c4a1b9dd9d6939d9832305b", "component_budget": { "node_count": 2, "item_count": 2, @@ -21092,18 +21036,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are only loosely related by training object detection models, but they report different failure modes and likely different root causes: one is a tensor shape mismatch, the other is an out-of-memory problem occurring after several epochs in RT-DETR v2 training.", - "confidence": 0.89, - "canonical_issue_reason": "Issue #46657 is the cleaner representative of the tensor-shape mismatch bug report, but it should not be treated as a duplicate of #46692.", + "summary": "The two issues describe different training failures: one is a tensor-shape mismatch during object detection training, the other is an out-of-memory condition after several epochs in RT-DETR v2. They are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "issue:46657 is the better anchor because it reports a concrete tensor size mismatch bug; issue:46692 is a separate memory-exhaustion report with a different symptom and likely different root cause.", "canonical_pr_reason": null, - "best_issue_reason": "#46657 is the better standalone issue to anchor this cluster because it describes a concrete, reproducible error symptom more directly than the memory-growth report in #46692.", + "best_issue_reason": "issue:46657 is the more specific and self-contained bug report, while issue:46692 is about OOM during training and does not match the tensor mismatch failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46657", "right": "issue:46692", "accept": false, - "reason": "Different symptoms and likely different code paths: tensor size mismatch during object-detection training versus OOM after some epochs in RT-DETR v2. Shared domain is too broad to treat as the same bug." + "reason": "Different underlying problems: tensor size mismatch vs. out-of-memory after epochs. Similar training context, but not the same bug." } ] }, @@ -21133,7 +21077,7 @@ "soft_pairs": [ "issue:46822|issue:46823" ], - "prepared_review_unit_hash": "10c3b165dc38753e4874d2ae62d1aebbb6b3ab35d04538555ef1a3ef7b321735", + "prepared_review_unit_hash": "c4f8af730f86dc338d589c865b46ad095648a1b135b8dc0f923812fa3fde022f", "component_budget": { "node_count": 2, "item_count": 2, @@ -21155,18 +21099,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues describe different model-loading failures in different pathways: ZeRO-3 zero.Init partitioning during multi-GPU load versus 4-bit auto device mapping with max_memory dispatching to CPU. They are not the same underlying bug, so the soft edge should be rejected.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 46822 is more specific and actionable, with a concrete failure mode (ZeRO-3 zero.Init not partitioning a composite submodule leading to OOM on multi-GPU load).", + "summary": "These issues are about different failures in different loading paths: ZeRO-3 zero.Init partitioning of a composite submodule vs 4-bit auto device mapping wrongly spilling to CPU. They should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "If a single representative is needed, issue 46822 is the better anchor because it describes a specific ZeRO-3 partitioning bug with a clear failure mode and reproduction context; issue 46823 is a separate 4-bit dispatch problem.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46822 is the better canonical issue because it isolates a single code-path problem and includes a clear reproduction symptom tied to partitioning behavior.", + "best_issue_reason": "Issue 46822 is the best overall issue candidate for this cluster because it is more self-contained and points to a concrete code path (zero.Init / composite submodule partitioning).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46822", "right": "issue:46823", "accept": false, - "reason": "Different loading mechanisms and failure modes: one is ZeRO-3 partitioning/OOM, the other is 4-bit device-map/max_memory CPU dispatch with ValueError. They are not plausibly the same bug." + "reason": "Different underlying bugs and code paths: ZeRO-3 partitioning/OOM during model load versus 4-bit device_map/max_memory CPU dispatch ValueError. Shared theme is model loading, but not the same defect." } ] }, @@ -21188,85 +21132,114 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46906-4", + "cluster_id": "cluster-46906-5", "nodes": [ "issue:46906", "issue:46940", "issue:46948", - "issue:46969" + "issue:46969", + "issue:46985" ], "soft_pairs": [ "issue:46906|issue:46940", "issue:46906|issue:46948", "issue:46906|issue:46969", + "issue:46906|issue:46985", "issue:46940|issue:46948", "issue:46940|issue:46969", - "issue:46948|issue:46969" + "issue:46940|issue:46985", + "issue:46948|issue:46969", + "issue:46948|issue:46985", + "issue:46969|issue:46985" ], - "prepared_review_unit_hash": "cee8b404959bc90b004c12240eab93f2f276d35415befed70f6f31097eda804b", + "prepared_review_unit_hash": "fd17e0271905c4434cfcda0effc42b67b7654f1d9ebb1bda03a1a6f44b24ada8", "component_budget": { - "node_count": 4, - "item_count": 4, - "soft_pair_count": 6, - "serialized_chars": 4441, - "estimated_input_tokens": 1111, - "estimated_eval_tokens": 2478 + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6439, + "estimated_input_tokens": 1610, + "estimated_eval_tokens": 3476 }, "budget": { - "node_count": 4, - "item_count": 4, - "soft_pair_count": 6, - "serialized_chars": 4441, - "estimated_input_tokens": 1111, - "estimated_eval_tokens": 2478 + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6439, + "estimated_input_tokens": 1610, + "estimated_eval_tokens": 3476 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These four issues look like recurring daily triage reports for the same area (\u201c[serge] integration failure triage\u201d) rather than duplicate reports of one underlying bug. They should not be merged as duplicates.", - "confidence": 0.89, - "canonical_issue_reason": "issue:46906 is the earliest item and has the strongest inbound reference signal, so it is the best representative if one must be chosen for the cluster.", + "summary": "These are highly similar daily triage issues, but they appear to be separate reports for different dates (2026-06-26 through 2026-06-30), not duplicates of one underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate exists here; the items are date-stamped triage reports for different days. If one must be chosen as an anchor, issue 46906 is the earliest and most referenced.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46906 is the most suitable issue to keep as the cluster representative because it is earliest and most referenced, but it still appears to be one instance of a recurring triage thread, not a unique bug report.", + "best_issue_reason": "Issue 46906 is the best representative only as an anchor because it is the earliest and has the most inbound references/discussion, but it should not be treated as a duplicate of the later daily triage issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46906", "right": "issue:46940", "accept": false, - "reason": "Same triage-series title, but this looks like a different day\u2019s report rather than the same underlying bug or change; no shared explicit target or code artifact evidence." + "reason": "Same template and subsystem, but the titles/body indicate different daily triage reports on different dates, so they are separate issues rather than duplicates." }, { "left": "issue:46906", "right": "issue:46948", "accept": false, - "reason": "These appear to be separate daily integration triage entries, not duplicate descriptions of one concrete failure." + "reason": "Date-separated triage entries; similarity is format-driven, not evidence of the same underlying bug." }, { "left": "issue:46906", "right": "issue:46969", "accept": false, - "reason": "Similar recurring title only; insufficient evidence that both issues track the same exact bug instance." + "reason": "These are distinct daily triage reports for different days, not one shared issue." + }, + { + "left": "issue:46906", + "right": "issue:46985", + "accept": false, + "reason": "Different report date and separate triage context; not the same underlying issue." }, { "left": "issue:46940", "right": "issue:46948", "accept": false, - "reason": "Different dates in the same triage series, with no explicit bug identity or shared implementation target." + "reason": "Both are daily triage reports, but for different dates, so they should not be merged." }, { "left": "issue:46940", "right": "issue:46969", "accept": false, - "reason": "Likely separate triage notes for successive days, not duplicates." + "reason": "The overlap is only in wording and subsystem; the reports are for different days." + }, + { + "left": "issue:46940", + "right": "issue:46985", + "accept": false, + "reason": "Separate date-stamped triage issues, not duplicate bug reports." }, { "left": "issue:46948", "right": "issue:46969", "accept": false, - "reason": "Same subsystem and template, but not the same underlying issue; these are separate daily triage reports." + "reason": "Different daily triage entries; no shared concrete bug/change is indicated." + }, + { + "left": "issue:46948", + "right": "issue:46985", + "accept": false, + "reason": "These are consecutive-day triage reports, but the evidence points to distinct issues." + }, + { + "left": "issue:46969", + "right": "issue:46985", + "accept": false, + "reason": "Same format and area, but different dates and no concrete indication of the same underlying problem." } ] }, @@ -21296,46 +21269,46 @@ "soft_pairs": [ "issue:46983|issue:46984" ], - "prepared_review_unit_hash": "a170229f340a3f8b7a566ded45dc09754c85fcb7fd30b0749e2917db400b4273", + "prepared_review_unit_hash": "ada375511251ec3ea7bdb00d4eb660249dacd71fbc19edde65e47ffcac916be6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1488, - "estimated_input_tokens": 372, - "estimated_eval_tokens": 1000 + "serialized_chars": 1490, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1488, - "estimated_input_tokens": 372, - "estimated_eval_tokens": 1000 + "serialized_chars": 1490, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues have the same title and appear to request the same feature addition (RWKV7 model support), with no evidence of distinct scopes. They are strong duplicates, and the open issue is the better canonical tracker.", + "summary": "These are the same feature request: identical titles, very similar bodies, and no evidence of distinct scope. The open issue is the better canonical tracking issue; the closed one is a duplicate or older copy.", "confidence": 0.98, - "canonical_issue_reason": "Issue #46984 is open and has the same feature request title as #46983, making it the best place to track the ongoing work.", + "canonical_issue_reason": "Issue #46984 is open, so it is the better long-lived tracker for the same RWKV7 support request. #46983 appears to be a duplicate with no distinguishing details.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46984 is the best overall issue because it is still open and appears to represent the same underlying request as the closed issue.", + "best_issue_reason": "#46984 is the best issue overall because it is open and can continue collecting progress and discussion for the same request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46983", "right": "issue:46984", "accept": true, - "reason": "Same title and same requested feature; both appear to be the same underlying model-support request, so this is a clear duplicate pair." + "reason": "Same title and near-identical request for RWKV7 model support; no signs of different scope, so this is a duplicate pair." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded in the packet: both items have the same title, are both issues, and the only pairwise evidence is strong soft similarity. The conclusion is conservative enough, and the open issue (#46984) is a reasonable canonical tracker for the same request." + "feedback": "The conclusion is well grounded in the packet: both items have the same title, the bodies are reported as highly similar via soft similarity, and there is no evidence of different scope or separate targets. Choosing the open issue as the canonical tracker is a conservative and reasonable call." }, "evaluator_used": true, "retried": false, @@ -21354,15 +21327,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-36", + "cluster_id": "cluster-40546-39", "nodes": [ "pull_request:40546", "pull_request:43451", + "pull_request:44517", "pull_request:45862", "pull_request:46180", "pull_request:46196", + "pull_request:46266", "pull_request:46425", - "pull_request:46509", "pull_request:46658", "pull_request:46707", "pull_request:46836", @@ -21373,7 +21347,6 @@ "pull_request:46907", "pull_request:46908", "pull_request:46912", - "pull_request:46917", "pull_request:46920", "pull_request:46926", "pull_request:46927", @@ -21390,7 +21363,8 @@ "pull_request:46970", "pull_request:46971", "pull_request:46974", - "pull_request:46977" + "pull_request:46977", + "pull_request:46986" ], "soft_pairs": [ "pull_request:46895|pull_request:46958", @@ -21399,38 +21373,41 @@ "pull_request:46950|pull_request:46970", "pull_request:46927|pull_request:46950", "pull_request:46927|pull_request:46951", + "pull_request:46970|pull_request:46986", "pull_request:46970|pull_request:46971", "pull_request:46951|pull_request:46971", "pull_request:46949|pull_request:46971", "pull_request:46951|pull_request:46970", + "pull_request:46927|pull_request:46986", "pull_request:46908|pull_request:46928", "pull_request:46907|pull_request:46926", "pull_request:46902|pull_request:46965", "pull_request:46907|pull_request:46928", + "pull_request:46951|pull_request:46986", "pull_request:46926|pull_request:46928", + "pull_request:46941|pull_request:46986", "pull_request:46926|pull_request:46941", + "pull_request:46949|pull_request:46986", "pull_request:46941|pull_request:46970", "pull_request:46907|pull_request:46941", "pull_request:46928|pull_request:46941", "pull_request:46180|pull_request:46658", - "pull_request:46180|pull_request:46707", "pull_request:46658|pull_request:46903", - "pull_request:46954|pull_request:46961", - "pull_request:46954|pull_request:46977", "pull_request:46928|pull_request:46951", "pull_request:46907|pull_request:46927", + "pull_request:46954|pull_request:46961", + "pull_request:46954|pull_request:46977", "pull_request:46908|pull_request:46951", - "pull_request:46891|pull_request:46954", "pull_request:46926|pull_request:46927", - "pull_request:43451|pull_request:46954", "pull_request:46908|pull_request:46941", + "pull_request:46891|pull_request:46954", + "pull_request:43451|pull_request:46954", "pull_request:46954|pull_request:46957", "pull_request:46903|pull_request:46970", "pull_request:46958|pull_request:46961", "pull_request:46958|pull_request:46977", - "pull_request:46961|pull_request:46974", "pull_request:46658|pull_request:46970", - "pull_request:46707|pull_request:46920", + "pull_request:46961|pull_request:46974", "pull_request:46958|pull_request:46974", "pull_request:46974|pull_request:46977", "pull_request:43451|pull_request:46920", @@ -21438,390 +21415,357 @@ "pull_request:46836|pull_request:46920", "pull_request:46891|pull_request:46920", "pull_request:46957|pull_request:46958", + "pull_request:46658|pull_request:46920", "pull_request:46891|pull_request:46958", "pull_request:46891|pull_request:46974", - "pull_request:46957|pull_request:46974", + "pull_request:46970|pull_request:46974", "pull_request:45862|pull_request:46425", "pull_request:46957|pull_request:46965", - "pull_request:46196|pull_request:46965", + "pull_request:43451|pull_request:46707", + "pull_request:46707|pull_request:46836", + "pull_request:46266|pull_request:46707", + "pull_request:46707|pull_request:46961", + "pull_request:46707|pull_request:46957", + "pull_request:46707|pull_request:46891", "pull_request:40546|pull_request:43451", "pull_request:46912|pull_request:46977", + "pull_request:46196|pull_request:46965", "pull_request:46836|pull_request:46965", "pull_request:46425|pull_request:46965", + "pull_request:46196|pull_request:46977", + "pull_request:44517|pull_request:46425", "pull_request:46196|pull_request:46425", "pull_request:43451|pull_request:46425", "pull_request:46425|pull_request:46961", "pull_request:43451|pull_request:46912", - "pull_request:46425|pull_request:46836", - "pull_request:46425|pull_request:46891", - "pull_request:46912|pull_request:46957", - "pull_request:46891|pull_request:46912", - "pull_request:46836|pull_request:46912", - "pull_request:46425|pull_request:46509", - "pull_request:46425|pull_request:46917", - "pull_request:46950|pull_request:46971", - "pull_request:46949|pull_request:46950", - "pull_request:46949|pull_request:46951", - "pull_request:46927|pull_request:46949", - "pull_request:46927|pull_request:46971", - "pull_request:46949|pull_request:46970" + "pull_request:46425|pull_request:46836" ], - "prepared_review_unit_hash": "88a004799fca0c085fe25ba6769cc0a3934de923ff06ecc1df57e52c0a1afcfe", + "prepared_review_unit_hash": "f9325fad47da095af63226fcfa4ee6b620ff8889778852b4d3937e949f0a43b0", "component_budget": { - "node_count": 36, - "item_count": 36, - "soft_pair_count": 108, - "serialized_chars": 121126, - "estimated_input_tokens": 30282, - "estimated_eval_tokens": 60820 + "node_count": 39, + "item_count": 39, + "soft_pair_count": 113, + "serialized_chars": 131300, + "estimated_input_tokens": 32825, + "estimated_eval_tokens": 65906 }, "budget": { - "node_count": 35, - "item_count": 35, + "node_count": 36, + "item_count": 36, "soft_pair_count": 72, - "serialized_chars": 98386, - "estimated_input_tokens": 24597, - "estimated_eval_tokens": 49450 + "serialized_chars": 102500, + "estimated_input_tokens": 25625, + "estimated_eval_tokens": 51506 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated PRs; the only near-duplicate CohereASR pair is close but should be treated as alternative fixes, not a mergeable duplicate.", - "confidence": 0.91, + "summary": "This cluster is mostly heterogeneous: new-model PRs, docs-only updates, CI/workflow changes, and several isolated bugfixes for different models. The only clear duplicate-like pair is the two CohereASR loss-shift fixes; the rest are unrelated.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46895 is the clearest representative of the CohereASR double-shift bugfix because it has the explicit issue link, targets the concrete training-loss path, and includes the aligned test coverage.", + "canonical_pr_reason": "No single canonical PR for the whole cluster: the items span unrelated features and bugfixes across many subsystems, so there is no true duplicate center.", "best_issue_reason": null, - "best_pr_reason": "PR 46895 is the strongest standalone PR in the cluster because it directly fixes the concrete bug and is the most complete artifact in the only plausible duplicate group.", + "best_pr_reason": "PR 46895 is the cleanest narrowly scoped representative of the only real duplicate subcluster: it directly fixes the CohereASR training-loss double-shift with an explicit issue link.", "soft_edge_verdicts": [ { "left": "pull_request:46895", "right": "pull_request:46958", - "accept": false, - "reason": "Same CohereASR bug, but the implementations diverge: one threads shift handling through the existing loss path, the other replaces the loss computation with a direct CrossEntropyLoss. Those are alternative fixes, not one clearly mergeable PR." + "accept": true, + "reason": "Same underlying CohereASR loss double-shift bug on the same code path; the two PRs are alternative implementations of the same fix." }, { "left": "pull_request:46903", "right": "pull_request:46958", "accept": false, - "reason": "Same model area, but different underlying fix: PPFormulaNet vs CohereASR, so this is not the same bug or change." + "reason": "They overlap on CohereASR, but 46903 is a broader multi-fix PR (also PPFormulaNet) with a different implementation shape, so it is not a clean duplicate." }, { "left": "pull_request:46950", "right": "pull_request:46951", "accept": false, - "reason": "Both touch Mamba2-related failures, but one changes model execution while the other only adjusts test loading behavior; not the same fix." + "reason": "Different models and failures: Mamba2 OOM vs PEFT integration load_error." }, { "left": "pull_request:46950", "right": "pull_request:46970", "accept": false, - "reason": "Different models and different code paths; same general failure class is too weak." + "reason": "Unrelated models and bugs: Mamba2 OOM vs MusicGen-Melody generation mismatch." }, { "left": "pull_request:46927", "right": "pull_request:46950", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Different model families and failure modes; no shared bug or change." }, { "left": "pull_request:46927", "right": "pull_request:46951", "accept": false, - "reason": "Different models and unrelated code paths." + "reason": "GLM46V fix vs PEFT integration fix; unrelated code paths." + }, + { + "left": "pull_request:46970", + "right": "pull_request:46986", + "accept": false, + "reason": "Different models and issues: MusicGen-Melody generation logic vs GLM-OCR model construction." }, { "left": "pull_request:46970", "right": "pull_request:46971", "accept": false, - "reason": "Same model family and failure theme, but the fixes are distinct and not clearly the same bug." + "reason": "MusicGen-Melody output mismatch and Mamba2 OOM are unrelated." }, { "left": "pull_request:46951", "right": "pull_request:46971", "accept": false, - "reason": "Different models and different fixes." + "reason": "PEFT integration vs Mamba2 generation OOM; different bugs." }, { "left": "pull_request:46949", "right": "pull_request:46971", "accept": false, - "reason": "Different models and unrelated code paths." + "reason": "Whisper/AudioFlamingo cache fix vs Mamba2 OOM; unrelated." }, { "left": "pull_request:46951", "right": "pull_request:46970", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Different subsystems and failure classes." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46986", + "accept": false, + "reason": "Unrelated model fixes." }, { "left": "pull_request:46908", "right": "pull_request:46928", "accept": false, - "reason": "Different models and distinct test failures." + "reason": "DeepSeekV3.2 load_error test fix vs BLIP-2 tokenizer mapping; unrelated." }, { "left": "pull_request:46907", "right": "pull_request:46926", "accept": false, - "reason": "Different models and unrelated bugs." + "reason": "EdgeTam import/config test fix vs MiniMax-M3-VL load_error; unrelated." }, { "left": "pull_request:46902", "right": "pull_request:46965", "accept": false, - "reason": "Different bugs in different subsystems." + "reason": "Offloading/save-state bugfix vs DeepSeek V4 cache handling; unrelated." + }, + { + "left": "pull_request:46180", + "right": "pull_request:46658", + "accept": false, + "reason": "Two new-model additions in different domains; not the same bug or change." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46965", + "accept": false, + "reason": "Finegrained FP8 generation optimization vs DeepSeek V4 use_cache handling; unrelated." }, { "left": "pull_request:46907", "right": "pull_request:46928", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Same broad infra style of test fixes, but different models and failures." }, { - "left": "pull_request:46926", - "right": "pull_request:46941", + "left": "pull_request:46951", + "right": "pull_request:46986", "accept": false, - "reason": "Different models and distinct failures." + "reason": "PEFT integration vs GLM-OCR model init; unrelated." }, { - "left": "pull_request:46926", - "right": "pull_request:46927", + "left": "pull_request:46941", + "right": "pull_request:46986", "accept": false, - "reason": "Different models and unrelated code paths." + "reason": "DAC output mismatch vs GLM-OCR construction bug; unrelated." }, { - "left": "pull_request:46941", - "right": "pull_request:46970", + "left": "pull_request:46926", + "right": "pull_request:46941", "accept": false, - "reason": "Different models and bugs." + "reason": "MiniMax-M3-VL load_error vs DAC tensor mismatch; unrelated." }, { "left": "pull_request:46907", "right": "pull_request:46941", "accept": false, - "reason": "Different models and unrelated failures." + "reason": "Different model families and failure modes." }, { "left": "pull_request:46928", "right": "pull_request:46941", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "BLIP-2 tokenizer mapping vs DAC codec bug; unrelated." }, { "left": "pull_request:46180", "right": "pull_request:46658", "accept": false, - "reason": "Both are model-addition PRs, but for different models and unrelated code." + "reason": "Repeated pair: still unrelated new-model additions." }, { - "left": "pull_request:46180", - "right": "pull_request:46707", + "left": "pull_request:46658", + "right": "pull_request:46903", "accept": false, - "reason": "Different features in different areas; not the same change." + "reason": "Step-3.7-Flash new-model work vs CohereASR/PPFormulaNet loss fixes; unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:46903", + "left": "pull_request:46928", + "right": "pull_request:46951", + "accept": false, + "reason": "Tokenizer mapping tweak vs PEFT integration bugfix; unrelated." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46927", "accept": false, - "reason": "Different models and unrelated changes." + "reason": "Different model fixes with no shared code path." }, { "left": "pull_request:46954", "right": "pull_request:46961", "accept": false, - "reason": "Different concerns: model cleanup vs workflow automation." + "reason": "Pytest-v9 compatibility fix vs Serge normalization workflow; unrelated." }, { "left": "pull_request:46954", "right": "pull_request:46977", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Pytest-v9 fix vs Mllama argument deprecation; unrelated." }, { "left": "pull_request:46958", "right": "pull_request:46961", "accept": false, - "reason": "CohereASR bugfix vs workflow change; unrelated." + "reason": "CohereASR loss fix vs workflow normalization; unrelated." }, { "left": "pull_request:46958", "right": "pull_request:46977", "accept": false, - "reason": "Different models and unrelated code." + "reason": "CohereASR loss fix vs Mllama arg-name cleanup; unrelated." }, { - "left": "pull_request:46957", + "left": "pull_request:46961", "right": "pull_request:46974", "accept": false, - "reason": "Different features and unrelated code paths." + "reason": "Workflow cleanup vs MusicGen typo fix; unrelated." }, { - "left": "pull_request:46912", - "right": "pull_request:46957", + "left": "pull_request:46958", + "right": "pull_request:46974", "accept": false, - "reason": "Different subsystems; same broad math theme is not enough." + "reason": "CohereASR training-loss fix vs MusicGen generate typo; unrelated." }, { - "left": "pull_request:46920", - "right": "pull_request:46957", + "left": "pull_request:46974", + "right": "pull_request:46977", "accept": false, - "reason": "Different models and unrelated issues." + "reason": "MusicGen generate typo vs Mllama deprecation cleanup; unrelated." }, { - "left": "pull_request:46917", + "left": "pull_request:43451", "right": "pull_request:46920", "accept": false, - "reason": "Different subsystems; deprecation cleanup vs audio API update." + "reason": "Molmo2 model add vs Voxtral-realtime audio API adaptation; unrelated." }, { - "left": "pull_request:46957", - "right": "pull_request:46965", - "accept": false, - "reason": "Different features and unrelated code paths." - }, - { - "left": "pull_request:46196", - "right": "pull_request:46965", + "left": "pull_request:46836", + "right": "pull_request:46920", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Unlimited OCR model add vs Voxtral-realtime API adaptation; unrelated." }, { - "left": "pull_request:40546", - "right": "pull_request:43451", + "left": "pull_request:46891", + "right": "pull_request:46920", "accept": false, - "reason": "Different model additions; not duplicates." + "reason": "PEFT CI workflow test branch vs Voxtral-realtime API migration; unrelated." }, { - "left": "pull_request:40546", - "right": "pull_request:46954", + "left": "pull_request:46957", + "right": "pull_request:46958", "accept": false, - "reason": "Different models and unrelated code." + "reason": "Finegrained FP8 generation optimization vs CohereASR loss fix; unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:46954", + "left": "pull_request:46957", + "right": "pull_request:46965", "accept": false, - "reason": "Different models and unrelated changes." + "reason": "Finegrained FP8 generation optimization vs DeepSeek V4 cache handling; unrelated." }, { - "left": "pull_request:46196", - "right": "pull_request:46425", + "left": "pull_request:46961", + "right": "pull_request:46977", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Repeated pair: unrelated workflow cleanup vs Mllama cleanup." }, { - "left": "pull_request:46836", - "right": "pull_request:46965", + "left": "pull_request:40546", + "right": "pull_request:43451", "accept": false, - "reason": "Different models and failure modes." + "reason": "VibeVoice model add vs Molmo2 model add; different models." }, { "left": "pull_request:46912", - "right": "pull_request:46957", + "right": "pull_request:46977", "accept": false, - "reason": "Different subsystems; same broad math theme is not enough." + "reason": "RoPE tests/rotary logic work vs Mllama argument cleanup; unrelated." }, { "left": "pull_request:46196", "right": "pull_request:46965", "accept": false, - "reason": "Different models and unrelated code paths." + "reason": "Wav2vec2 SDPA dispatch vs DeepSeek V4 cache semantics; unrelated." }, { - "left": "pull_request:40546", - "right": "pull_request:43451", - "accept": false, - "reason": "Different model-addition PRs; not duplicates." - }, - { - "left": "pull_request:40546", - "right": "pull_request:46954", - "accept": false, - "reason": "Different models and unrelated changes." - }, - { - "left": "pull_request:43451", - "right": "pull_request:46954", + "left": "pull_request:46836", + "right": "pull_request:46965", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Unlimited OCR model add vs DeepSeek V4 cache fix; unrelated." }, { "left": "pull_request:46425", "right": "pull_request:46965", "accept": false, - "reason": "Different models and bugs." + "reason": "RADIO model support vs DeepSeek V4 cache fix; unrelated." }, { "left": "pull_request:46196", - "right": "pull_request:46425", - "accept": false, - "reason": "Different models and unrelated fixes." - }, - { - "left": "pull_request:43451", - "right": "pull_request:46912", - "accept": false, - "reason": "Different models and unrelated changes." - }, - { - "left": "pull_request:46425", - "right": "pull_request:46836", - "accept": false, - "reason": "Different models and unrelated fixes." - }, - { - "left": "pull_request:46425", - "right": "pull_request:46920", - "accept": false, - "reason": "Different models and unrelated code paths." - }, - { - "left": "pull_request:46891", - "right": "pull_request:46954", - "accept": false, - "reason": "CI/workflow change vs model cleanup; unrelated." - }, - { - "left": "pull_request:46957", - "right": "pull_request:46958", - "accept": false, - "reason": "Different models and unrelated fixes." - }, - { - "left": "pull_request:46891", - "right": "pull_request:46958", - "accept": false, - "reason": "Workflow automation vs CohereASR bugfix; unrelated." - }, - { - "left": "pull_request:46958", - "right": "pull_request:46974", + "right": "pull_request:46977", "accept": false, - "reason": "Different models and unrelated changes." + "reason": "Wav2vec2 SDPA dispatch vs Mllama deprecation cleanup; unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:46920", + "left": "pull_request:44517", + "right": "pull_request:46425", "accept": false, - "reason": "Different models and unrelated fixes." + "reason": "Qwen3-TTS model add vs RADIO model support; unrelated." }, { - "left": "pull_request:46891", - "right": "pull_request:46974", + "left": "pull_request:46196", + "right": "pull_request:46425", "accept": false, - "reason": "Workflow automation vs music model typo fix; unrelated." + "reason": "Wav2vec2 SDPA dispatch vs RADIO model support; unrelated." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Reject: the one accepted soft-edge pair is too strong for the packet. PR 46895 and PR 46958 both address the CohereASR double-shift bug, but they implement it in different ways: 46895 threads shift_labels / conditional shift handling through the existing loss path, while 46958 replaces the loss call with a direct CrossEntropyLoss over reshaped logits and labels. Those are alternative fixes rather than clearly mergeable changes for a single PR, so the accept verdict is not conservative enough. The rest of the verdicts are mostly conservative, but this one acceptance is enough to reject." + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is heterogeneous, and the only clearly duplicate-like pair is the CohereASR loss-double-shift fix pair (46895/46958). The conservative false verdicts for the other pairs are supported by the diffs and titles, with no obvious overstatement of evidence." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -21836,300 +21780,348 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-40546-36", + "cluster_id": "cluster-40546-39", "nodes": [ "pull_request:43451", "pull_request:46196", + "pull_request:46425", + "pull_request:46509", "pull_request:46658", - "pull_request:46707", "pull_request:46836", "pull_request:46891", "pull_request:46907", "pull_request:46908", + "pull_request:46912", + "pull_request:46917", "pull_request:46926", "pull_request:46927", + "pull_request:46949", + "pull_request:46950", + "pull_request:46951", "pull_request:46957", "pull_request:46961", "pull_request:46968", "pull_request:46970", + "pull_request:46971", "pull_request:46977" ], "soft_pairs": [ + "pull_request:46425|pull_request:46891", + "pull_request:46912|pull_request:46957", + "pull_request:43451|pull_request:46196", + "pull_request:46891|pull_request:46912", + "pull_request:46836|pull_request:46912", + "pull_request:46196|pull_request:46961", + "pull_request:46425|pull_request:46509", + "pull_request:46425|pull_request:46917", + "pull_request:46950|pull_request:46971", + "pull_request:46949|pull_request:46950", + "pull_request:46949|pull_request:46951", + "pull_request:46927|pull_request:46949", + "pull_request:46927|pull_request:46971", + "pull_request:46949|pull_request:46970", "pull_request:46927|pull_request:46970", "pull_request:46908|pull_request:46926", "pull_request:46907|pull_request:46908", "pull_request:43451|pull_request:46836", "pull_request:46658|pull_request:46836", - "pull_request:46891|pull_request:46957", "pull_request:43451|pull_request:46658", - "pull_request:46658|pull_request:46707", + "pull_request:46891|pull_request:46957", "pull_request:43451|pull_request:46891", - "pull_request:46196|pull_request:46836", - "pull_request:43451|pull_request:46707", - "pull_request:46707|pull_request:46891", "pull_request:46658|pull_request:46957", - "pull_request:46836|pull_request:46891", + "pull_request:46658|pull_request:46891", "pull_request:46836|pull_request:46957", + "pull_request:46836|pull_request:46891", "pull_request:46961|pull_request:46977", - "pull_request:46707|pull_request:46836", "pull_request:46891|pull_request:46968", - "pull_request:46707|pull_request:46957", + "pull_request:46891|pull_request:46961", "pull_request:46957|pull_request:46968", - "pull_request:43451|pull_request:46196", - "pull_request:46707|pull_request:46968", "pull_request:43451|pull_request:46957", "pull_request:43451|pull_request:46968", - "pull_request:46196|pull_request:46891", - "pull_request:46196|pull_request:46707", - "pull_request:46196|pull_request:46957", "pull_request:46836|pull_request:46968", "pull_request:46957|pull_request:46961", - "pull_request:46891|pull_request:46961", "pull_request:46957|pull_request:46977", "pull_request:46891|pull_request:46977", + "pull_request:46658|pull_request:46968", "pull_request:43451|pull_request:46961", "pull_request:43451|pull_request:46977", "pull_request:46836|pull_request:46961", "pull_request:46836|pull_request:46977" ], - "prepared_review_unit_hash": "bddd290b4b1113b39e76e74b778d6fac7bf412217de7602b3ea0ca82568c3e5c", + "prepared_review_unit_hash": "6d0dbf280bc56fa88add1f8f0cdec435984b153ede268dbcf5bebcf1dc9a04f0", "component_budget": { - "node_count": 36, - "item_count": 36, - "soft_pair_count": 108, - "serialized_chars": 121126, - "estimated_input_tokens": 30282, - "estimated_eval_tokens": 60820 + "node_count": 39, + "item_count": 39, + "soft_pair_count": 113, + "serialized_chars": 131300, + "estimated_input_tokens": 32825, + "estimated_eval_tokens": 65906 }, "budget": { - "node_count": 15, - "item_count": 15, - "soft_pair_count": 36, - "serialized_chars": 45249, - "estimated_input_tokens": 11313, - "estimated_eval_tokens": 22882 + "node_count": 22, + "item_count": 22, + "soft_pair_count": 41, + "serialized_chars": 61370, + "estimated_input_tokens": 15343, + "estimated_eval_tokens": 30942 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a broad set of unrelated PRs that only share superficial textual similarity (new model additions, small regression fixes, docs/CI/workflow edits, and infrastructure changes). None of the soft edges are true duplicates or mergeable as one concrete change.", - "confidence": 0.97, + "summary": "This is a heterogeneous cluster: mostly unrelated model-addition PRs plus separate one-off fixes, docs, CI, and test-only changes. I do not see any soft pair that is clearly the same concrete bug/change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the items cover distinct features and fixes across unrelated code paths, so any representative would be misleading.", + "canonical_pr_reason": "43451 is the broadest, earliest model-addition PR and the closest hub by similarity, but the cluster is too mixed for it to be a true duplicate anchor.", "best_issue_reason": null, - "best_pr_reason": "No global best PR; there isn't a single PR that meaningfully represents the others because they are not the same underlying change.", + "best_pr_reason": "43451 is the least-bad representative because it has the largest scope and highest discussion activity, but it still does not match the rest of the cluster as a duplicate.", "soft_edge_verdicts": [ { - "left": "pull_request:46927", - "right": "pull_request:46970", + "left": "pull_request:46425", + "right": "pull_request:46891", "accept": false, - "reason": "Different model regressions on different code paths (`glm46v` input casting vs `musicgen_melody` generation inputs)." + "reason": "Different features: RADIO model support vs PEFT CI workflow changes." }, { - "left": "pull_request:46908", - "right": "pull_request:46926", + "left": "pull_request:46912", + "right": "pull_request:46957", "accept": false, - "reason": "Both are load-error test fixes, but for different models and causes; not the same bug." + "reason": "Unrelated changes: RoPE test coverage vs FP8/generation optimization work." }, { - "left": "pull_request:46907", - "right": "pull_request:46908", + "left": "pull_request:43451", + "right": "pull_request:46196", "accept": false, - "reason": "Different models and fixes (`edgetam` kwarg handling vs `deepseek_v32` loading/offload issue)." + "reason": "Different model families and different changes; model addition vs wav2vec2 SDPA dispatch." }, { - "left": "pull_request:43451", - "right": "pull_request:46836", + "left": "pull_request:46891", + "right": "pull_request:46912", "accept": false, - "reason": "Both add model support, but for unrelated models; not one underlying change." + "reason": "CI/workflow change vs RoPE tests; not the same bug or code path." }, { - "left": "pull_request:46658", - "right": "pull_request:46836", + "left": "pull_request:46836", + "right": "pull_request:46912", "accept": false, - "reason": "Separate model additions with different architecture/codebases; too broad to merge as one PR." + "reason": "Unlimited OCR model addition is unrelated to RoPE tests." }, { - "left": "pull_request:46891", - "right": "pull_request:46957", + "left": "pull_request:46196", + "right": "pull_request:46961", "accept": false, - "reason": "Unrelated changes: CI/PEFT integration workflow versus generation/fp8 kernel optimization." + "reason": "Wav2vec2 attention dispatch vs Serge branch-normalization workflow; unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:46658", + "left": "pull_request:46425", + "right": "pull_request:46509", "accept": false, - "reason": "Distinct model integration PRs; same broad theme, not the same bug or feature." + "reason": "Different model additions (RADIO vs Nemotron Omni), no shared bug/change." }, { - "left": "pull_request:46658", - "right": "pull_request:46707", + "left": "pull_request:46425", + "right": "pull_request:46917", + "accept": false, + "reason": "Model support addition vs deprecation cleanup; unrelated." + }, + { + "left": "pull_request:46950", + "right": "pull_request:46971", + "accept": false, + "reason": "Same model family and symptom, but one changes model code and the other only test loading args; not mergeable as one PR." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46950", + "accept": false, + "reason": "Whisper/audioflamingo cache fix vs Mamba2 OOM fix; unrelated models and code paths." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46951", + "accept": false, + "reason": "Whisper/audioflamingo fix vs PEFT loading fix; unrelated." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46949", + "accept": false, + "reason": "GLM46V input-id fix vs Whisper/audioflamingo cache fix; unrelated." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46971", + "accept": false, + "reason": "GLM46V model fix vs Mamba2 test loading tweak; not the same issue." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46970", + "accept": false, + "reason": "Different models and failures: Whisper/audioflamingo vs MusicGen Melody output mismatch." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46970", + "accept": false, + "reason": "GLM46V vs MusicGen Melody; unrelated." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46926", "accept": false, - "reason": "A new model add versus distributed FSDP2 infrastructure; unrelated changes." + "reason": "Different models and different failure modes; DeepSeek V3.2 load error vs MiniMax-M3-VL OOM test tweak." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46908", + "accept": false, + "reason": "Separate model-specific test fixes for different architectures." }, { "left": "pull_request:43451", - "right": "pull_request:46891", + "right": "pull_request:46836", "accept": false, - "reason": "Model addition PR and CI workflow PR are not the same underlying change." + "reason": "Both are model additions, but for different models and independent code changes." }, { - "left": "pull_request:46196", + "left": "pull_request:46658", "right": "pull_request:46836", "accept": false, - "reason": "Docs/model SDPA support versus adding a new OCR model; unrelated." + "reason": "Step-3.7 Flash implementation vs Unlimited OCR addition; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46658", + "accept": false, + "reason": "Molmo2 addition vs Step-3.7 Flash build/conversion work; different model families." }, { - "left": "pull_request:43451", - "right": "pull_request:46707", + "left": "pull_request:46891", + "right": "pull_request:46957", "accept": false, - "reason": "Model addition and distributed runtime work are different code paths." + "reason": "PEFT CI workflow changes vs FP8/generation optimization; unrelated." }, { - "left": "pull_request:46707", + "left": "pull_request:43451", "right": "pull_request:46891", "accept": false, - "reason": "FSDP2 migration and PEFT CI workflow changes are unrelated." + "reason": "Model addition vs PEFT CI workflow; unrelated." }, { "left": "pull_request:46658", "right": "pull_request:46957", "accept": false, - "reason": "Different features in different subsystems; no shared concrete bug or patch." + "reason": "Step-3.7 Flash model-conversion work vs FP8/generation optimization; unrelated." }, { - "left": "pull_request:46836", + "left": "pull_request:46658", "right": "pull_request:46891", "accept": false, - "reason": "New model support versus CI plumbing; not mergeable as one fix." + "reason": "Step-3.7 Flash model work vs PEFT CI workflow; unrelated." }, { "left": "pull_request:46836", "right": "pull_request:46957", "accept": false, - "reason": "Unrelated model addition and generation kernel optimization." + "reason": "Unlimited OCR addition vs FP8/generation optimization; unrelated." }, { - "left": "pull_request:46961", - "right": "pull_request:46977", + "left": "pull_request:46836", + "right": "pull_request:46891", "accept": false, - "reason": "Workflow normalization and Mllama argument cleanup are unrelated changes." + "reason": "Unlimited OCR model addition vs PEFT CI workflow; unrelated." }, { - "left": "pull_request:46707", - "right": "pull_request:46836", + "left": "pull_request:46961", + "right": "pull_request:46977", "accept": false, - "reason": "Distributed FSDP2 work and OCR model addition are unrelated." + "reason": "Branch-normalization workflow vs Mllama argument deprecation; unrelated." }, { "left": "pull_request:46891", "right": "pull_request:46968", "accept": false, - "reason": "CI test workflow update and docs autolink cleanup are unrelated." - }, - { - "left": "pull_request:46707", - "right": "pull_request:46957", - "accept": false, - "reason": "Different code paths: distributed training support versus generation/fp8 optimization." + "reason": "PEFT CI workflow vs docs autolink cleanup; unrelated." }, { - "left": "pull_request:46196", - "right": "pull_request:43451", + "left": "pull_request:46891", + "right": "pull_request:46961", "accept": false, - "reason": "Both are large PRs, but one is SDPA dispatch/docs and the other is a new model addition." + "reason": "PEFT CI workflow vs branch-normalization workflow; different automation tasks." }, { - "left": "pull_request:46707", + "left": "pull_request:46957", "right": "pull_request:46968", "accept": false, - "reason": "Distributed FSDP2 code versus docs link fixes; not the same change." + "reason": "FP8/generation code vs docs autolink cleanup; unrelated." }, { "left": "pull_request:43451", "right": "pull_request:46957", "accept": false, - "reason": "A model addition PR and a generation performance PR are unrelated." + "reason": "Molmo2 model addition vs FP8/generation optimization; unrelated." }, { "left": "pull_request:43451", "right": "pull_request:46968", "accept": false, - "reason": "New model support and docs autolink fixes do not share the same underlying bug." - }, - { - "left": "pull_request:46196", - "right": "pull_request:46891", - "accept": false, - "reason": "Model attention dispatch changes versus PEFT CI changes are unrelated." - }, - { - "left": "pull_request:46196", - "right": "pull_request:46707", - "accept": false, - "reason": "Different subsystems: wav2vec2/audio attention dispatch vs distributed FSDP2 support." - }, - { - "left": "pull_request:46196", - "right": "pull_request:46957", - "accept": false, - "reason": "Audio model SDPA work and generation/fp8 optimization are unrelated." + "reason": "Molmo2 model addition vs docs cleanup; unrelated." }, { "left": "pull_request:46836", "right": "pull_request:46968", "accept": false, - "reason": "OCR model addition versus docs autolink cleanup; different scopes entirely." + "reason": "Unlimited OCR model addition vs docs cleanup; unrelated." }, { "left": "pull_request:46957", "right": "pull_request:46961", "accept": false, - "reason": "Generation optimization and Serge workflow normalization are unrelated." - }, - { - "left": "pull_request:46891", - "right": "pull_request:46961", - "accept": false, - "reason": "CI pipeline changes and branch-normalization workflow are different automation concerns." + "reason": "FP8/generation optimization vs branch-normalization workflow; unrelated." }, { "left": "pull_request:46957", "right": "pull_request:46977", "accept": false, - "reason": "Generation/fp8 optimization and Mllama API cleanup are different code paths." + "reason": "FP8/generation optimization vs Mllama arg deprecation; unrelated." }, { "left": "pull_request:46891", "right": "pull_request:46977", "accept": false, - "reason": "CI integration work and Mllama argument deprecation are unrelated." + "reason": "PEFT CI workflow vs Mllama arg-name cleanup; unrelated." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46968", + "accept": false, + "reason": "Step-3.7 Flash model work vs docs cleanup; unrelated." }, { "left": "pull_request:43451", "right": "pull_request:46961", "accept": false, - "reason": "A new model PR and a workflow-normalization PR are not the same change." + "reason": "Molmo2 model addition vs branch-normalization workflow; unrelated." }, { "left": "pull_request:43451", "right": "pull_request:46977", "accept": false, - "reason": "Model addition and Mllama argument renaming are unrelated fixes." + "reason": "Molmo2 model addition vs Mllama arg cleanup; unrelated." }, { "left": "pull_request:46836", "right": "pull_request:46961", "accept": false, - "reason": "New OCR model support versus repo workflow normalization; not duplicate." + "reason": "Unlimited OCR model addition vs branch-normalization workflow; unrelated." }, { "left": "pull_request:46836", "right": "pull_request:46977", "accept": false, - "reason": "OCR model integration and Mllama API cleanup are unrelated." + "reason": "Unlimited OCR model addition vs Mllama arg cleanup; unrelated." } ] }, @@ -22159,7 +22151,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "f3958475e33dd1b4c742643a9978c797b03bdb203501aaabab760de0ce33fc88", + "prepared_review_unit_hash": "997a7c81311f5231255a739db0b60380965b84a69f284a089ec11f4bdd3dbb81", "component_budget": { "node_count": 2, "item_count": 2, @@ -22181,18 +22173,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one adds TimesFM covariate forecasting support, while the other adds DINOv3 image classification support. They do not fix the same code path or represent the same change, so the soft duplicate edge should be rejected.", - "confidence": 0.98, + "summary": "The two PRs are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They are not duplicates and should not be clustered together for merge triage.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR for this cluster; the PRs target different models and different functionality, so neither can represent the same underlying change.", + "canonical_pr_reason": "No canonical PR: the items address different code paths, different models, and different user-facing capabilities.", "best_issue_reason": null, - "best_pr_reason": "No single best PR as a duplicate representative because these are unrelated feature additions rather than variants of the same fix.", + "best_pr_reason": "pull_request:41224 is the better representative only in the sense that it is a tighter, more self-contained feature addition; however, it is not a duplicate of the TimesFM PR.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different model families and different features: TimesFM covariate forecasting vs DINOv3 image classification. Not the same underlying change." + "reason": "Different models and different functionality. PR 40755 adds TimesFM covariate forecasting across TimesFM docs/code/tests, while PR 41224 adds DINOv3 image-classification support. They do not fix the same bug or implement the same change." } ] }, @@ -22313,12 +22305,12 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "9d3b6b417781ce3b595925e033ad292bbe1571e1ef972cc2e1373cbe857e44eb", + "prepared_review_unit_hash": "323bd23d01e6cc2ca125db73e1e51b687abaf813b90a512018005243863911c7", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34784, + "serialized_chars": 34782, "estimated_input_tokens": 8696, "estimated_eval_tokens": 17648 }, @@ -22326,7 +22318,7 @@ "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34784, + "serialized_chars": 34782, "estimated_input_tokens": 8696, "estimated_eval_tokens": 17648 }, @@ -22335,129 +22327,129 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is mostly unrelated across distinct subsystems. The clearest near-duplicate is 43251/43254: both change `loss_utils.fixed_cross_entropy` to pass additional `cross_entropy` kwargs for the same issue target, so they look like the same underlying fix. The other pairs are weak soft-similarity matches only.", - "confidence": 0.94, + "summary": "Only one pair looks like a true duplicate-style overlap: the two `fixed_cross_entropy` PRs that both fix ignored cross-entropy kwargs. The rest are broad-similarity matches across unrelated models, docs, or infrastructure changes and should stay separate.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:43251 is the best canonical PR anchor because it is the more explicit and complete `fixed_cross_entropy` kwargs fix in `src/transformers/loss/loss_utils.py`, and it cleanly subsumes the close variant in 43254.", + "canonical_pr_reason": "PR 43251 is the clearest representative of the only real overlap: it directly fixes `fixed_cross_entropy` to forward supported kwargs and explicitly tracks issue 43240.", "best_issue_reason": null, - "best_pr_reason": "pull_request:43251 is the best overall representative PR because it most clearly captures the concrete `cross_entropy` kwargs passthrough change in `loss_utils`, while 43254 is a near-identical variant of that same fix.", + "best_pr_reason": "PR 43251 is the best canonical PR because it is the more explicit, issue-linked version of the same loss-function fix and is the only duplicate-worthy change in this set.", "soft_edge_verdicts": [ { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Tokenizer method additions vs a new MobileLLM model; different goals and code paths." + "reason": "Different bugs/features: Mistral tokenizer missing methods vs adding MobileLLM model support." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer API work vs ExecuTorch export test additions; unrelated changes." + "reason": "Unrelated domains: tokenizer methods vs executorch export dynamic-shape testing." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor auto-loading robustness vs model addition; different subsystems." + "reason": "Both are processor/model-loading related only at a very broad level; they fix different models and different issues." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss helper kwargs fix vs new model addition; no shared bug or implementation path." + "reason": "Loss-function kwargs fix is unrelated to adding a new model package." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "`cross_entropy` kwargs fix vs Mistral4 model/test updates; unrelated." + "reason": "Different code paths: cross-entropy kwargs vs Mistral4 test/model changes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch dynamic-shape test work vs Mistral4 changes; not the same fix." + "reason": "Executorch export tests are unrelated to Mistral4 model/test fixes." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "ExecuTorch testing vs MobileLLM model addition; different code paths." + "reason": "Dynamic-shape export testing is unrelated to MobileLLM model addition." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "Model addition vs config-resolution ordering fix; different bugs." + "reason": "New model addition and config-resolution bug fix are distinct changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "Model addition vs Falcon output-interface refactor; not the same underlying problem." + "reason": "Different models and different problems: MobileLLM support vs Falcon output-interface refactor." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor vs model-creation skill/tooling content; no shared code-path bug." + "reason": "Falcon refactor is unrelated to a new-model scaffolding skill PR." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon interface refactor vs Mistral4 tests/model changes; unrelated." + "reason": "Falcon output-collection refactor and Mistral4 test fixes do not target the same bug." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model changes vs CLIP `layrnorm` renaming; different bugs." + "reason": "Mistral4 test fixes are unrelated to CLIP-like `layrnorm` conversion mapping." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 test fixes vs chat parsing/serve overhaul; unrelated." + "reason": "Mistral4 testing and chat response parsing updates are different subsystems and issues." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes vs TDT loss kernel addition; different code paths." + "reason": "Mistral4 tests are unrelated to adding the TDT loss kernel." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 changes vs config-resolution ordering fix; unrelated." + "reason": "Mistral4 test/model work does not match the `cls.config` resolution-order fix." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 model/test work vs model-creation skill/tooling; not a duplicate change." + "reason": "A model test PR and a model-creation skill PR are not the same underlying change." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition vs model-creation skill/tooling; unrelated." + "reason": "New model support and the skill/documentation scaffolding PR are not duplicates." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Same underlying `cross_entropy` kwargs bug in `loss_utils`; same file/function and plausibly mergeable into one fix." + "reason": "Both PRs fix the same concrete bug in `fixed_cross_entropy`: forwarding supported kwargs to `nn.functional.cross_entropy` for the same loss path." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The soft-edge verdicts are mostly conservative and well-grounded, and the 43251/43254 match is supported by the shared issue target and same file. However, the summary overstates the evidence by calling 43251/43254 an \"accepted loss fix\" and the \"only real duplicate pair\" without packet support. The packet only justifies that they are the clearest near-duplicate in this set, not that one is accepted or that no other duplicate exists beyond the reviewed similarities." + "accept": true, + "feedback": "The summary is grounded in the packet. The only accepted overlap, PR 43251 vs PR 43254, is supported by the shared file, shared issue target 43240, and the near-identical `fixed_cross_entropy` kwargs-forwarding fix. The remaining pairwise verdicts stay conservative: they are rejected based on clearly different subsystems/models or only broad soft similarity, without overstating the evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -22481,7 +22473,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "b282d8d4a533e6422c654510b8aa407d1899738cbbdcbc2f10f4fde1256f2aba", + "prepared_review_unit_hash": "e50f5e32747a82db1c53ad55c057f2cfd0483ce6092c6016d2a5a41a5848d921", "component_budget": { "node_count": 2, "item_count": 2, @@ -22503,18 +22495,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both perform CI workflow dependency bumps, but they target different actions and different upgrade goals: one is a broad Node 24 compatibility update for core GitHub Actions, the other is a Docker build action version bump. They are related maintenance work, not duplicates.", - "confidence": 0.92, + "summary": "These are related GitHub Actions maintenance PRs, but they are not the same underlying change: one is a Node 24 compatibility bump for core GitHub Actions, while the other upgrades docker/build-push-action and other workflows to newer releases. Similar area, different concrete fixes.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42975 is the more representative workflow-maintenance PR: it upgrades several core GitHub Actions used across many workflows and has the clearer, repo-wide compatibility goal.", + "canonical_pr_reason": "PR #42976 is the broader GitHub Actions version-bump PR and better represents the shared workflow-maintenance theme, while #42975 is a narrower compatibility-specific upgrade.", "best_issue_reason": null, - "best_pr_reason": "PR #42975 is the better standalone representative because it is broader and more central to GitHub Actions compatibility; #42976 is narrower and focused mainly on docker/build-push-action updates.", + "best_pr_reason": "PR #42976 is the better cluster anchor because it spans more workflows and is framed as the general latest-version upgrade, making it the most representative single PR here.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are workflow dependency bumps, but they change different action families and solve different maintenance issues. They do not appear to fix the same concrete bug/change, so they should not be merged as duplicates." + "reason": "Both are workflow dependency upgrades, but they target different action families and different problems: Node 24 compatibility for actions/checkout/cache/upload-artifact vs. docker/build-push-action and other latest-version bumps. They are not the same concrete change and would not naturally merge into one PR." } ] }, @@ -22549,12 +22541,12 @@ "pull_request:43448|pull_request:45490", "pull_request:45490|pull_request:46562" ], - "prepared_review_unit_hash": "c9a73943c470b72e1e4f73d0d9aeae0acc100ea9d55d02cfbfbf83a1d2599f39", + "prepared_review_unit_hash": "f9bf30b2a4a9854a7a5f296b04092acbded7cb28d06778fd050d023acdd55271", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13044, + "serialized_chars": 13041, "estimated_input_tokens": 3261, "estimated_eval_tokens": 6778 }, @@ -22562,7 +22554,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13044, + "serialized_chars": 13041, "estimated_input_tokens": 3261, "estimated_eval_tokens": 6778 }, @@ -22571,36 +22563,36 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are four unrelated model-addition PRs that happen to share boilerplate areas like docs, auto registries, and test scaffolding. They do not appear to fix the same bug or implement the same change, so none of the soft duplicate edges should be accepted.", - "confidence": 0.96, + "summary": "These PRs are separate model-addition changes, not duplicates. The shared files are mostly boilerplate docs/auto-registration plumbing, not the same underlying implementation or bug fix.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No PR here is a canonical duplicate target because each PR introduces a different model family and its own files, configs, tests, and docs. The strongest standalone PR is Step 3.7 Flash, mainly because it has the broadest implementation surface and the highest review activity, but it is not a duplicate of the others.", + "canonical_pr_reason": "45490 is the most central representative if one PR must be chosen: it sits between the others on shared infra/docs touchpoints, but the set is still unrelated overall.", "best_issue_reason": null, - "best_pr_reason": "pull_request:46562 is the best representative only in the sense that it appears the most mature and extensively reviewed among the unrelated model-addition PRs. It is not a duplicate match for the others; it simply looks like the most complete individual contribution in this cluster.", + "best_pr_reason": "45490 is the best cluster anchor only in the sense of centrality and shared plumbing; it is not a true duplicate of the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto files are boilerplate for registering new models, not evidence of the same underlying change." + "reason": "ViT NEPA and Molmo are different model additions; overlap is limited to generic docs/auto-registry scaffolding." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model additions: ViT NEPA vs CTSM. Overlap is limited to common integration points like auto mappings and docs, which is expected for separate models." + "reason": "Different model families and code paths; shared model-registration/docs files are boilerplate, not a common change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Different model additions: Molmo vs CTSM. They touch the same broad subsystem but implement distinct model families with different configs, processors, and tests." + "reason": "Molmo vs CTSM are unrelated architectures; shared files are generic model-addition plumbing." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "Different model additions: CTSM vs Step3p7. Shared registry and docs updates are generic new-model plumbing, not the same concrete feature or bug fix." + "reason": "CTSM and Step3p7 target different model stacks and modalities; shared auto/docs edits do not indicate the same fix." } ] }, @@ -22649,130 +22641,130 @@ "pull_request:46539|pull_request:46922", "pull_request:46886|pull_request:46922" ], - "prepared_review_unit_hash": "16baf7270bd8dbb6ccdabba689c7442120796199a86c1ebf1fc8433431824825", + "prepared_review_unit_hash": "9a569dd0b2396bfb6108857074bb5b55949dc1f822fcd76cf1821cef8e9815e2", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19589, - "estimated_input_tokens": 4898, - "estimated_eval_tokens": 10052 + "serialized_chars": 19601, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 }, "budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19589, - "estimated_input_tokens": 4898, - "estimated_eval_tokens": 10052 + "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": "Most items are unrelated PRs touching different bugs or test/CI infrastructure. The only strong duplicate-like pair is the two YOLOS fixes targeting the same device-mismatch issue in the same model file.", - "confidence": 0.97, + "summary": "This cluster is mostly a set of unrelated PRs that all happen to have generic or similar titles. The only strong duplicate-like pair is the two YOLOS fixes, which address the same device-mismatch issue in the same file and target the same tracking issue.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46922 is the best canonical representative for the YOLOS bug: it fixes the concrete device mismatch in the forward path and also preserves the no-split-module change, while sharing the same explicit issue target as 46886.", + "canonical_pr_reason": "PR 46886 is the best canonical PR because it has the clearest, specific title, explicit issue target, and a focused fix for the YOLOS device-mismatch bug.", "best_issue_reason": null, - "best_pr_reason": "PR 46922 is the strongest standalone fix for the actual YOLOS device_map=\"auto\" problem; 46886 looks like an earlier/partial variant of the same change.", + "best_pr_reason": "PR 46886 is the strongest representative of the cluster\u2019s one real duplicate pair; the others are either test-only changes or different subsystems and should not be grouped with it.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Different areas and bugs: Whisper tokenizer decode behavior vs CLIP/ViT test edits." + "reason": "Both are generic test-fix PRs, but they touch different model test files and appear to fix unrelated test issues." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decoding fix is unrelated to FP8 dequant/model-loading changes." + "reason": "Whisper test cleanup and FP8 dequant changes are unrelated bugs and code paths." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Different subsystems: Whisper decoding vs generation watermarking validation." + "reason": "Whisper test changes are unrelated to generation-config watermarking and generation logic." }, { "left": "pull_request:43291", "right": "pull_request:46922", "accept": false, - "reason": "Whisper tokenizer fix and YOLOS device-mismatch fix are unrelated." + "reason": "Whisper tokenizer/test changes are unrelated to the YOLOS device-mismatch fix." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Test assertion churn in CLIP/ViT is unrelated to Phimoe CI memory/runtime setup." + "reason": "Model test fixes do not indicate the same underlying bug; one is CLIP/ViT tests, the other is CI/integration setup." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "These target different code paths: model tests vs FP8 dequantization/loading." + "reason": "Different subsystems and different failure modes: vision model tests vs FP8 loading/dequant logic." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "Unrelated: model test edits vs generation config/watermarking bug fix." + "reason": "Test PR for CLIP/ViT is unrelated to generation config and Whisper code." }, { "left": "pull_request:43665", "right": "pull_request:46922", "accept": false, - "reason": "Test-only CLIP/ViT changes are not the same bug as the YOLOS runtime fix." + "reason": "The CLIP/ViT test edits do not match the YOLOS runtime/device-mismatch change." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "FP8 dequantization/model-loading logic is unrelated to Phimoe CI integration test setup." + "reason": "FP8 quantization/loading changes are unrelated to Phimoe integration-test CI fixes." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Different concrete issues: FP8 kernel/loading fix vs generation watermarking validation." + "reason": "These address different code paths: FP8 dequantization vs generation config validation." }, { "left": "pull_request:44815", "right": "pull_request:46922", "accept": false, - "reason": "FP8 dequantization and YOLOS device placement are distinct bugs in different models/code paths." + "reason": "No common bug or code path; FP8 padding/loading changes are unrelated to YOLOS." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation config validation is unrelated to CI environment fixes for Phimoe tests." + "reason": "Generation watermarking validation is unrelated to Phimoe CI/device-memory test setup." }, { "left": "pull_request:46380", "right": "pull_request:46922", "accept": false, - "reason": "Watermarking-config validation and YOLOS device mismatch are unrelated changes." + "reason": "Generation config validation fixes and YOLOS device handling are separate changes." }, { "left": "pull_request:46539", "right": "pull_request:46922", "accept": false, - "reason": "CI memory/device-map test infrastructure is unrelated to the YOLOS model bug." + "reason": "CI memory/device-map test harness changes are not the same bug as the YOLOS model forward/device fix." }, { "left": "pull_request:46886", "right": "pull_request:46922", "accept": true, - "reason": "Same underlying YOLOS device_map=\"auto\" device-mismatch bug, same file, same explicit issue target, and the later PR contains the concrete forward-path fix plus the split-module change." + "reason": "Both PRs target the same YOLOS device-mismatch issue, share the same file and issue target, and the second is a concrete refinement of the same fix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge verdicts are appropriately conservative. The only accepted pair, 46886/46922, is a mergeable same-bug YOLOS fix in the same file with the same issue target, and the other pairs are reasonably judged unrelated based on the diff previews." + "feedback": "The assessment is grounded in the packet: the non-YOLOS PRs are plausibly unrelated despite superficial title similarity, and the only clearly mergeable duplicate-like pair is the two YOLOS fixes sharing the same file and issue target. The soft-edge rejections are conservative, and the one accepted pair looks legitimately combinable into a single bugfix PR." }, "evaluator_used": true, "retried": false, @@ -22802,12 +22794,12 @@ "pull_request:43488|pull_request:43532", "pull_request:43532|pull_request:46747" ], - "prepared_review_unit_hash": "e42e5d2945faef1013c5e5f3433fad114956188f10ad5832815509fb342542c0", + "prepared_review_unit_hash": "7f95f34d31d74addebacc909990c023192cd474ed9a0efe626f8314b5dc4106c", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6348, + "serialized_chars": 6347, "estimated_input_tokens": 1587, "estimated_eval_tokens": 3430 }, @@ -22815,7 +22807,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6348, + "serialized_chars": 6347, "estimated_input_tokens": 1587, "estimated_eval_tokens": 3430 }, @@ -22824,30 +22816,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three pull requests are unrelated '[don't merge]' test/demo changes; they overlap only superficially in being workflow/test infrastructure edits. None should be clustered as duplicates.", - "confidence": 0.93, + "summary": "These are three unrelated synthetic \u201c[don\u2019t merge]\u201d PRs touching different areas (setup/deps, CI/workflows, and a security-gate workflow). They share some broad repository/testing context, but not the same underlying bug or change, so none should be clustered as duplicates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "No true canonical PR here because the changes target different code paths. If one representative must be chosen, 43532 is the most substantive and descriptive of the set because it spans multiple workflow and test-support files.", + "canonical_pr_reason": "#43532 is the broadest/most central artifact in the set by file spread and activity, but it still targets a different change set than the other PRs.", "best_issue_reason": null, - "best_pr_reason": "43532 is the strongest representative for the cluster only in the sense that it has the richest, most varied diff; it is still not a duplicate of the others.", + "best_pr_reason": "#43532 is the best representative only in the sense of being the most central and active PR here; it is not a true duplicate of #43488 or #46747.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:46747", "accept": false, - "reason": "43488 changes setup/version/dependency metadata and a Vit test; 46747 only changes the PR security-gate workflow trigger. Different underlying changes, not the same bug or fix." + "reason": "Different changes: setup/dependency/test-bot checks versus a PR security-gate workflow tweak. Shared repo/testing context is too broad to treat as the same fix." }, { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "43488 is a packaging/version tweak plus a test change, while 43532 edits CI workflows, testing utilities, and notification code. Shared test file name is incidental, not the same concrete change." + "reason": "#43488 is a setup/dependency/version test-bot PR, while #43532 changes CI workflows and helper utilities. They overlap only on a Vit test file, not the same code-path problem." }, { "left": "pull_request:43532", "right": "pull_request:46747", "accept": false, - "reason": "Both touch workflows, but they address different systems: 43532 adjusts self-comment CI and related helpers, while 46747 changes the security-gate workflow trigger. Not mergeable as one fix." + "reason": "Both touch workflows, but they modify different workflows for different purposes. They would not plausibly merge into one PR as the same underlying change." } ] }, @@ -22877,12 +22869,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "4fd726eb74f634e7111e97573dee9fcdd2398ac6f80c31768d936040156393be", + "prepared_review_unit_hash": "354c5f34f26384c3695ced6baa09e572fcbef94b108ad4947c6474295c9aa2c9", "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 }, @@ -22890,7 +22882,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -22899,18 +22891,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add different segmentation pipelines with overlapping scaffolding, but they implement distinct features: promptable concept segmentation vs promptable visual segmentation. The shared doc/autoload/test metadata changes are not enough to treat them as duplicates.", - "confidence": 0.96, + "summary": "These PRs are related in theme and scaffolding, but they add different pipelines for different segmentation tasks, so they should not be treated as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43613 is the broader, more complete implementation of a specific pipeline feature, with multiple model integrations and dedicated tests/doc updates; if one PR must represent the cluster, it has the strongest standalone scope.", + "canonical_pr_reason": "PR #43613 is the stronger representative: it adds the Promptable Visual Segmentation pipeline with broader model coverage and test support, making it the more complete change.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the better representative because it covers a fully fleshed-out change set for one concrete code path, while 43612 is a narrower variant of a different segmentation task.", + "best_pr_reason": "PR #43613 is the best standalone PR in this cluster because it is more comprehensive and better covers the intended code path and integrations.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Both touch pipeline/documentation scaffolding, but they target different features and code paths: concept segmentation versus visual segmentation. The model support, task docs, and pipeline modules differ enough that they are not the same underlying change and would not naturally merge into one PR." + "reason": "Different underlying features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They share pipeline/docs scaffolding, but target different tasks and backends, so they are not the same change." } ] }, @@ -22940,7 +22932,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "885d50507a5c8bc7ffad4efbbc349f5c3be4878e1489a8b60c25d9717c210eb9", + "prepared_review_unit_hash": "2a1ef1769f78e69bec1e3e8a40b004b9ca6b86907bc0427f1149f9289b7216c2", "component_budget": { "node_count": 2, "item_count": 2, @@ -22962,18 +22954,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two different trainer extensibility PRs in the same file, not the same underlying change: one adjusts gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT models.", - "confidence": 0.98, + "summary": "These two PRs both touch `trainer.py`, but they address different trainer extension points: one fixes gradient-accumulation loss scaling, the other refactors signature-column inspection for PEFT/custom column filtering. They are related to customization in Trainer, but not the same underlying change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:43651 is the better representative of the cluster because it has the clearer, more substantial behavioral change and the more explicit trainer-extension API addition, while 43663 is a separate refactor with a different purpose.", + "canonical_pr_reason": "PR #43651 is the more concrete representative of the cluster because it fixes a specific training-step bug in loss scaling and introduces the corresponding hook/property in one focused change.", "best_issue_reason": null, - "best_pr_reason": "pull_request:43651 is the strongest standalone PR: it introduces a focused new property and changes loss scaling behavior in a concrete code path. 43663 is unrelated enough that it should not be merged into the same duplicate bucket.", + "best_pr_reason": "PR #43651 is the best overall representative here: it is a self-contained bugfix with a clear behavioral change, whereas PR #43663 is a separate refactor for column-signature handling.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different trainer code paths and different goals: 43651 handles gradient accumulation loss normalization, while 43663 adds an overrideable helper for signature-column filtering. They are not the same bug/change and would not plausibly be merged as one PR." + "reason": "Both modify `trainer.py`, but they change different code paths and solve different problems: GA loss scaling vs. signature column filtering. They would not plausibly be merged as one duplicate PR." } ] }, @@ -23003,12 +22995,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "689efca2809432e984624f5f0b57f4040b5e3173526ef1e2c63b0bfe9628f322", + "prepared_review_unit_hash": "1596ff57e1eedcccf97d0096e418874cb2e49231d721e2f8a131f6ce6c4dab7a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -23016,7 +23008,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -23025,18 +23017,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs touch the same CLI file and both mention NameError, but they target different annotation-related failures in `src/transformers/cli/serve.py`: one is a broader typing/forward-reference cleanup, the other specifically fixes `TypeAdapter` when `pydantic` is absent.", + "summary": "Two open PRs touch `src/transformers/cli/serve.py`, but they fix different `TypeAdapter` annotation/runtime issues: one is a broader typing/postponed-annotations cleanup, the other specifically guards `TypeAdapter` for missing `pydantic`.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "`pull_request:43842` is the more specific and concrete fix: it guards the `TypeAdapter` import under `TYPE_CHECKING` and stringizes the annotation, directly addressing the runtime NameError when `pydantic` is not installed.", + "canonical_pr_reason": "PR #43656 is the better cluster representative because it has more activity and an explicit linked issue, and its fix is the broader CLI `TypeAdapter` NameError report.", "best_issue_reason": null, - "best_pr_reason": "`pull_request:43842` is the best representative because it fixes a narrowly defined crash path with a clear cause and minimal scope; `pull_request:43656` looks like a broader type-hint modernization rather than the same bug.", + "best_pr_reason": "PR #43842 is the strongest standalone fix because it targets a concrete runtime failure when `pydantic` is absent, with a minimal, focused change.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Shared file and similar symptom wording, but they fix different underlying problems: broader annotation/forward-reference issues vs. `TypeAdapter` import failure without `pydantic`. They are not merge-equivalent." + "reason": "Reject: these are not the same underlying bug. #43656 adds postponed-annotation typing changes to address a CLI NameError, while #43842 specifically makes `TypeAdapter` a TYPE_CHECKING-only import to avoid failures when `pydantic` is not installed." } ] }, @@ -23174,8 +23166,8 @@ "pull_request:44019|pull_request:44071", "pull_request:44068|pull_request:44722", "pull_request:44018|pull_request:44116", - "pull_request:44068|pull_request:44116", "pull_request:44071|pull_request:44722", + "pull_request:44068|pull_request:44116", "pull_request:44018|pull_request:44085", "pull_request:44018|pull_request:44086", "pull_request:44068|pull_request:44086", @@ -23183,8 +23175,8 @@ "pull_request:44085|pull_request:44086", "pull_request:44059|pull_request:44086", "pull_request:44098|pull_request:44101", - "pull_request:44059|pull_request:44722", "pull_request:44018|pull_request:44071", + "pull_request:44059|pull_request:44722", "pull_request:44017|pull_request:44161", "pull_request:44017|pull_request:44722", "pull_request:44017|pull_request:44073", @@ -23228,12 +23220,12 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "6ba7b061efabdf40fac14c850f7f32f09c08701dd8d8ec0661937c2be944ab43", + "prepared_review_unit_hash": "1685ef7a2234eb5ede85f1681c1bc275d6f44928442fb8ad74b4e4c4a8c69868", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84901, + "serialized_chars": 84902, "estimated_input_tokens": 21226, "estimated_eval_tokens": 42708 }, @@ -23241,369 +23233,357 @@ "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75097, - "estimated_input_tokens": 18775, - "estimated_eval_tokens": 37806 + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated model-specific refactors for standardized output tracing. The only strong duplicate groups are the GPT-Neo pair and the GPT-J/CodeGen variants; the rest are just same-theme PRs touching different model implementations.", - "confidence": 0.82, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the best representative anchor: it covers the GPT-J/CodeGen output-tracing refactor, shares the same tracking issue, and overlaps the earlier GPT-J-specific variants.", + "summary": "This cluster is mostly a batch of model-specific output-tracing refactors under the shared tracker #43979, not one duplicate bug. Only PRs that touch the same model/code path and the same tracing refactor are duplicates; cross-model pairs should stay separate.", + "confidence": 0.9, + "canonical_issue_reason": "No standalone issue artifact is present in the cluster; the common tracker referenced by many PRs is #43979.", + "canonical_pr_reason": "PR #44066 is the strongest canonical PR for the GPT-J/CodeGen subset: it has the explicit tracker, broader scope, and the clearest overlap with the other GPT-J tracing PRs.", "best_issue_reason": null, - "best_pr_reason": "PR 44722 is the strongest single representative of the cluster's main duplicate subthread because it is the broadest and most complete GPT-J-family standardized output-tracing change, with direct overlap to 44066 and 44085.", + "best_pr_reason": "PR #44066 is the best representative PR in this cluster because it is explicit, substantive, and closest to the other true duplicates in the GPT-J family.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; same output-tracing theme is not enough." + "reason": "Different models and code paths; same issue target is just a shared tracking issue." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different model families and code paths." + "reason": "Different model families; only the output-tracing theme is shared." }, { "left": "pull_request:44066", "right": "pull_request:44085", - "accept": false, - "reason": "Related GPT-J family work, but 44066 is broader and includes CodeGen; not the same concrete PR change." + "accept": true, + "reason": "Same GPT-J output-tracing refactor on the same file; these are plausibly mergeable into one PR." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet and EfficientNet are separate model implementations; only the refactor pattern matches." + "reason": "Different backbones and implementations; same tracker does not make them duplicates." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "Different model families and code paths." + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen versus EfficientNet; same theme only." + "reason": "GPT-J/CodeGen and EfficientNet are different model refactors." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated implementations." + "reason": "MobileNetV2 and DeBERTa V2 are unrelated model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "Different model families and code paths." + "reason": "GPT-J vs MGP-STR; same general refactor theme only." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "Different model families and code paths." + "reason": "GPT-J/CodeGen and MPT are different code paths." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Same GPT-Neo output-tracing refactor in the same file/test area; these look like the same underlying change." + "reason": "Both are GPT-Neo output-tracing refactors in the same file and look mergeable together." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen versus GPT-Neo; shared issue theme only." + "reason": "GPT-J/CodeGen and GPT-Neo are different models." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "Different model families and code paths." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific refactors." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J are unrelated code paths." + "reason": "ResNet and GPT-J/CodeGen are unrelated despite the shared tracing pattern." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "Different model families and code paths." + "reason": "MGP-STR and GPT-J/CodeGen are different model code paths." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are unrelated code paths." + "reason": "ResNet and GPT-J are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are unrelated model implementations." + "reason": "GPT-Neo and ResNet are different models; same tracker is not enough." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are unrelated code paths." + "reason": "ResNet and MPT are different code paths." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo versus GPT-J/CodeGen; same refactor pattern only." + "reason": "GPT-Neo and GPT-J/CodeGen are different model-specific refactors." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are separate code paths." - }, - { - "left": "pull_request:44068", - "right": "pull_request:44116", - "accept": false, - "reason": "GPT-Neo and Flaubert are separate code paths." + "reason": "GPT-Neo and Flaubert are unrelated model implementations." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "Different model families and code paths." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44085", - "accept": false, - "reason": "Different model families and code paths." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44086", - "accept": false, - "reason": "Different model families and code paths." + "reason": "MPT and GPT-J/CodeGen are different code paths." }, { "left": "pull_request:44068", - "right": "pull_request:44086", - "accept": false, - "reason": "Different model families and code paths." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44085", + "right": "pull_request:44116", "accept": false, - "reason": "Different model families and code paths." + "reason": "GPT-Neo and Flaubert are not the same underlying change." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "Different model families and code paths." + "reason": "GPT-J and MGP-STR are different models." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated code paths." + "reason": "GPT-2 and MGP-STR are unrelated implementations." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and XLM/Flaubert are unrelated model implementations." + "reason": "ViLT and Flaubert/XLM are different model families." }, { - "left": "pull_request:44059", - "right": "pull_request:44071", + "left": "pull_request:44017", + "right": "pull_request:44161", "accept": false, - "reason": "GPT2 and MPT are different model families." + "reason": "SegFormer and LongT5 are unrelated code paths." }, { - "left": "pull_request:44056", - "right": "pull_request:44101", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "MPNet and XLM/Flaubert are different code paths." + "reason": "SegFormer and GPT-J/CodeGen are different model refactors." }, { - "left": "pull_request:43997", - "right": "pull_request:44017", + "left": "pull_request:44017", + "right": "pull_request:44073", "accept": false, - "reason": "RegNet and SegFormer are different models; only the refactor pattern matches." + "reason": "SegFormer and VisualBERT are unrelated models." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:44066", + "right": "pull_request:44086", "accept": false, - "reason": "RegNet and MPNet are unrelated implementations." + "reason": "GPT-J and MGP-STR are different model-specific changes." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:44066", + "right": "pull_request:44071", "accept": false, - "reason": "MPT and ImageGPT are different model families." + "reason": "GPT-J/CodeGen and MPT are different code paths." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44018", + "right": "pull_request:44086", "accept": false, - "reason": "TextNet and XLM/Flaubert are unrelated code paths." + "reason": "GPT-Neo and MGP-STR are unrelated model implementations." }, { - "left": "pull_request:43997", - "right": "pull_request:44018", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated model implementations." + "reason": "GPT-Neo and MPT are separate model refactors." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "MPNet and LongT5 are unrelated code paths." + "reason": "GPT-J and MGP-STR are different code paths." }, { - "left": "pull_request:43997", + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-2 and GPT-Neo are different models." + }, + { + "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are unrelated code paths." + "reason": "MPNet and LongT5 are unrelated implementations." }, { - "left": "pull_request:44074", + "left": "pull_request:44059", "right": "pull_request:44071", "accept": false, - "reason": "TextNet and MPT are unrelated model implementations." + "reason": "GPT-2 and MPT are different model families." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "SqueezeBert and MPT are different code paths." + "reason": "GPT-2 and GPT-J/CodeGen are unrelated changes." }, { - "left": "pull_request:44018", - "right": "pull_request:44129", + "left": "pull_request:44017", + "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo and SpeechT5 are unrelated model implementations." + "reason": "SegFormer and LongT5 are different model-specific refactors." }, { - "left": "pull_request:44010", + "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are different code paths." + "reason": "SegFormer and VisualBERT are not the same underlying change." }, { - "left": "pull_request:43997", - "right": "pull_request:44074", + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "RegNet and TextNet are unrelated model implementations." + "reason": "TextNet and Flaubert/XLM are unrelated code paths." }, { - "left": "pull_request:44017", + "left": "pull_request:44074", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are unrelated code paths." + "reason": "TextNet and LongT5 are different model implementations." }, { - "left": "pull_request:44017", + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated model refactors." + }, + { + "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SegFormer and VisualBert are unrelated model implementations." + "reason": "SqueezeBert and VisualBERT are different models." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:44010", + "right": "pull_request:44018", "accept": false, - "reason": "VisualBert and Flaubert are separate code paths." + "reason": "SqueezeBert and GPT-Neo are separate code paths." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:44010", + "right": "pull_request:44044", "accept": false, - "reason": "MPNet and LongT5 are unrelated model implementations." + "reason": "SqueezeBert and DeBERTa V2 are unrelated model implementations." }, { - "left": "pull_request:43997", - "right": "pull_request:44017", + "left": "pull_request:44010", + "right": "pull_request:44074", "accept": false, - "reason": "RegNet and SegFormer are different code paths." + "reason": "SqueezeBert and TextNet are different model-specific changes." }, { - "left": "pull_request:43997", - "right": "44074", + "left": "pull_request:44044", + "right": "pull_request:44074", "accept": false, - "reason": "RegNet and TextNet are unrelated model implementations." + "reason": "DeBERTa V2 and TextNet are unrelated code paths." }, { - "left": "pull_request:44010", - "right": "pull_request:44044", + "left": "pull_request:44044", + "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and DeBERTa are unrelated code paths." + "reason": "DeBERTa V2 and SpeechT5 are different model families." }, { "left": "pull_request:44044", - "right": "pull_request:44074", + "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and TextNet are different model families." + "reason": "DeBERTa V2 and VITS are unrelated implementations." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J-family standardized output-tracing change with overlapping files and the same tracking issue; 44722 looks like the later, broader version of the same refactor." + "reason": "Same GPT-J output-tracing refactor family; overlapping file and change intent make them mergeable." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both target GPT-J output-tracing in the same file family and same issue thread; they look like the same underlying refactor." + "reason": "Both are GPT-J tracing refactors on the same code path, so they look like duplicates." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated code paths." + "reason": "GPT-J and SpeechT5 are different model code paths." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "FNet/CVT and RegNet are different model implementations." + "reason": "CVT/FNet and RegNet/ResNet are different model refactors." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "FNet/CVT and SpeechT5 are unrelated code paths." + "reason": "CVT/FNet and SpeechT5 are unrelated model implementations." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are unrelated model implementations." + "reason": "DeBERTa V2 and SpeechT5 are different underlying changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and Vits are different code paths." + "reason": "DeBERTa V2 and VITS are unrelated model-specific refactors." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays grounded in the packet: the only clearly mergeable duplicate subthreads are the GPT-Neo pair and the GPT-J/CodeGen group, while the rest are same-theme but model-specific changes. The soft-edge accepts are conservative and supported by shared files/issue targets; the false verdicts also look appropriately cautious." + "feedback": "The summary is grounded in the packet and stays conservative: it correctly frames the cluster as mostly model-specific output-tracing refactors under shared tracker #43979, and the accepted soft-edge pairs are all within the same model/code path with plausible single-PR mergeability. No pair is overclaimed across unrelated model families." }, "evaluator_used": true, "retried": false, @@ -23641,12 +23621,12 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "02a059ac6999a1119685a552ae036410693cca4a6c17308fb965e6f1d89fc5ca", + "prepared_review_unit_hash": "6296e5f7064674c4ad2e4ba1580d861792b7fae146fdc87f5ea3f44981e73edc", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84901, + "serialized_chars": 84902, "estimated_input_tokens": 21226, "estimated_eval_tokens": 42708 }, @@ -23654,7 +23634,7 @@ "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17125, + "serialized_chars": 17127, "estimated_input_tokens": 4282, "estimated_eval_tokens": 8820 }, @@ -23663,48 +23643,48 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all variations of a standardized output-tracing refactor, but they target different model implementations and are not duplicates of one another. They look like a family of related maintenance changes rather than the same underlying bug fix or feature branch.", - "confidence": 0.89, + "summary": "These PRs are thematically related by a shared refactor pattern around standardized output tracing/capture, but they touch different model implementations and code paths, so they are not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the best anchor: it has an explicit issue link, a concrete SpeechT5 output-tracing refactor, and the clearest standalone scope among the set.", + "canonical_pr_reason": "pull_request:44129 is the best single anchor for the cluster because it has the clearest issue linkage and a concrete end-to-end output-tracing refactor, but it still represents only one model-specific instance of the broader pattern.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the strongest representative because it is the most explicitly scoped and least ambiguous about the intended change.", + "best_pr_reason": "pull_request:44129 is the strongest representative PR: it is issue-linked, has a focused scope, and shows the standardized output-capture change most explicitly.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both modernize output tracing, but they touch different models (FocalNet vs ImageGPT) and different code paths, so they are not the same change and would not plausibly merge into one PR." + "reason": "Both are standardized-output refactors, but they target different models (FocalNet vs ImageGPT) and different internal execution paths." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are separate model-specific refactors; same migration pattern, but not the same concrete bug or code path." + "reason": "Similar decorator-based refactor, but these are separate model implementations (VisualBert vs TextNet) with distinct forward logic." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert have similar output-capture plumbing changes, but they affect different architectures and cannot be treated as duplicates." + "reason": "Same general output-tracing theme, but MPNet and VisualBert are unrelated code paths and not plausibly one merged fix." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These are parallel refactors in different models, not one underlying fix." + "reason": "Both are refactors for output capture, but they modify different models and should not be merged as one duplicate change." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert each migrate output collection independently; similarity is thematic, not a shared code-path problem." + "reason": "RegNet and SqueezeBert are unrelated model families; the similarity is only at the refactor pattern level." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 both use the new output-capture utilities, but they modify different model internals and solve different model-specific implementation updates." + "reason": "Both adopt standardized output tracing, but they affect different model stacks (SqueezeBert vs SpeechT5) and different concrete layers." } ] }, @@ -23784,280 +23764,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "3a9b31a5af36de62d5d5bf51b39b42673fbde9577f67f5ba92288a464b6ca34e", + "prepared_review_unit_hash": "757e63c33ca9ba911bf249c8631ec4e466c640d541f707d81ec1624f400bc842", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40162, - "estimated_input_tokens": 10041, - "estimated_eval_tokens": 20338 + "serialized_chars": 40178, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40162, - "estimated_input_tokens": 10041, - "estimated_eval_tokens": 20338 + "serialized_chars": 40178, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are all separate model-specific output-tracing refactors, not true duplicates. The depth_anything PR is the best representative of the cluster, but none of the soft-edge pairs should be merged as duplicates.", - "confidence": 0.92, + "summary": "This cluster is a set of related-but-distinct output-tracing refactors across many different Transformer models. They share a naming pattern and some helper plumbing, but they are not duplicates of the same bug or change.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44025 is the strongest representative of the cluster: it spans two closely related files (`depth_anything` and `prompt_depth_anything`) and cleanly applies the output-tracing/can_return_tuple refactor pattern.", + "canonical_pr_reason": "PR 44003 is the best representative because it is the broadest of the group, touching two closely related Mamba-family files and adding the new output-capture plumbing; it best captures the cluster\u2019s theme, though it is not a duplicate of the others.", "best_issue_reason": null, - "best_pr_reason": "PR 44025 is the most central and complete-looking candidate in the set, so it is the best single PR to anchor this cluster even though the others are separate model-specific changes.", + "best_pr_reason": "PR 44003 is the strongest cluster representative for the same reason: it has the widest scope and most central refactor shape among these output-tracing PRs.", "soft_edge_verdicts": [ { "left": "pull_request:43998", - "right": "pull_request:44000", + "right": "pull_request:44002", "accept": false, - "reason": "Different model families (`timm_backbone` vs `vision_text_dual_encoder`); same refactor theme only." + "reason": "Different model families and code paths (timm_backbone vs upernet); same refactor style, not the same change." }, { "left": "pull_request:43998", - "right": "pull_request:44001", + "right": "pull_request:44000", "accept": false, - "reason": "Different code paths (`timm_backbone` vs `univnet`); not the same bug/change." + "reason": "Unrelated architectures (timm_backbone vs vision_text_dual_encoder); too broad a similarity to treat as one fix." }, { "left": "pull_request:43998", - "right": "pull_request:44002", + "right": "pull_request:44001", "accept": false, - "reason": "Separate model implementations (`timm_backbone` vs `upernet`); only broadly similar output-refactor work." + "reason": "Different models and behavior surfaces (timm_backbone vs univnet); not one concrete bug or mergeable PR." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "`timm_backbone` and `mamba/falcon_mamba` are unrelated code paths." - }, - { - "left": "pull_request:43998", - "right": "pull_request:44004", - "accept": false, - "reason": "Different subsystems (`timm_backbone` vs `codegen`); not a duplicate PR." + "reason": "Both are tracing refactors, but they target different model stacks and code paths; not the same underlying change." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Unrelated model families (`timm_backbone` vs `rwkv`); same broad output-tracing idea only." + "reason": "timm_backbone and rwkv are unrelated implementations; similarity is only the output-tracing pattern." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "`mobilenet_v1` and `univnet` are distinct models with separate forward paths." + "reason": "MobileNetV1 and UnivNet are unrelated models; shared helper imports do not make this one duplicate." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Separate model code paths (`mobilenet_v1` vs `upernet`); not the same fix." + "reason": "Different model code paths (MobileNetV1 classification vs UperNet segmentation); not mergeable as one PR." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "`mobilenet_v1` and `depth_anything` are unrelated model families." + "reason": "MobileNetV1 and Depth Anything fix different architectures and output APIs; only the refactor theme matches." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "`mobilenet_v1` and `vision_encoder_decoder` are different implementations; no shared concrete bug." + "reason": "MobileNetV1 and VisionEncoderDecoder are different code paths; not the same concrete bug/change." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Different model families (`mobilenet_v1` vs `speech_encoder_decoder`)." + "reason": "Different wrappers and modalities (image classifier vs speech encoder-decoder); too broad to merge together." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "`vision_text_dual_encoder` and `univnet` touch unrelated forward/output code paths." + "reason": "Vision-text dual encoder and UnivNet are unrelated models; same helper style, different fixes." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Separate models (`vision_text_dual_encoder` vs `upernet`); only same high-level refactor pattern." + "reason": "Different model families and forward paths; no shared concrete bug." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "`vision_text_dual_encoder` and `mamba/falcon_mamba` are not the same underlying change." + "reason": "Vision-text dual encoder and Mamba-family refactors are separate model stacks, so this is not a duplicate." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (`vision_text_dual_encoder` vs `depth_anything`)." + "reason": "Different tasks and architectures (vision-text vs depth estimation); not the same underlying change." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "`univnet` and `upernet` are separate model-specific refactors." + "reason": "Unrelated model code paths (audio generation vs segmentation); same refactor motif only." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "`univnet` and `mamba/falcon_mamba` do not share the same code path." + "reason": "Different model families and implementations; too broad to consider one fix." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different models (`univnet` vs `depth_anything`); not a duplicate." - }, - { - "left": "pull_request:44001", - "right": "pull_request:44004", - "accept": false, - "reason": "`univnet` and `codegen` are unrelated changes." + "reason": "Unrelated models and output contracts; not a single bug or patch." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "`upernet` and `rwkv` are separate model implementations with different output plumbing." + "reason": "UperNet segmentation and RWKV language modeling are unrelated code paths; not a duplicate." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "`upernet` and `mamba/falcon_mamba` are distinct code paths." + "reason": "Different model stacks (segmentation vs Mamba language models); same theme, not same change." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "`mamba/falcon_mamba` and `superpoint` are unrelated model families." + "reason": "Mamba-family output tracing and SuperPoint feature extraction are different implementations; not one concrete fix." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "`mamba/falcon_mamba` and `codegen` are different concrete fixes." + "reason": "Different code paths and model classes (Mamba vs CodeGen); the overlap is only in refactor style." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "`timm_backbone` and `codegen` are unrelated model paths." + "reason": "timm_backbone and CodeGen are unrelated models; same output-tracing pattern is not enough for duplicate status." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "`univnet` and `codegen` are separate model-specific changes." + "reason": "UnivNet and CodeGen are different model families with distinct forward paths." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "`upernet` and `codegen` do not share the same underlying issue." + "reason": "UperNet segmentation and CodeGen generation are unrelated; not the same underlying bug/change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder and CodeGen have different code paths and outputs; only the refactor theme matches." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "`depth_anything` and `vision_encoder_decoder` are different wrappers and code paths." + "reason": "Depth Anything and VisionEncoderDecoder are different model families; not mergeable as one PR." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "`depth_anything` and `speech_encoder_decoder` are separate model-specific refactors." + "reason": "Depth estimation and speech encoder-decoder are unrelated changes." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "`depth_anything` and `superpoint` are unrelated models." + "reason": "Different models and output semantics (depth estimation vs SuperPoint); not the same fix." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "`depth_anything` and `rwkv` are not the same fix, only similarly themed." + "reason": "Depth Anything and RWKV are unrelated code paths; only the output-tracing refactor pattern overlaps." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "`depth_anything` and `dpr` are different model families." + "reason": "Depth estimation and DPR retrieval models are different implementations; not a duplicate." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44029", "accept": false, - "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are analogous but distinct wrapper implementations." + "reason": "VisionEncoderDecoder and RWKV are unrelated model stacks; same helper refactor, different changes." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44027", "accept": false, - "reason": "`vision_encoder_decoder` and `superpoint` are unrelated code paths." + "reason": "VisionEncoderDecoder and SpeechEncoderDecoder are similar wrapper names but different modalities and code paths." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44028", "accept": false, - "reason": "`vision_encoder_decoder` and `rwkv` are different model families." + "reason": "Different model types and outputs; not the same concrete bug or mergeable change." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "`vision_encoder_decoder` and `dpr` are separate model-specific refactors." + "reason": "VisionEncoderDecoder and DPR address different components and APIs; not duplicates." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "`speech_encoder_decoder` and `superpoint` are unrelated models." + "reason": "SpeechEncoderDecoder and SuperPoint are unrelated models; similarity is only superficial." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "`speech_encoder_decoder` and `rwkv` do not share the same concrete code path." + "reason": "Speech encoder-decoder and RWKV are different forward paths and output contracts." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "`speech_encoder_decoder` and `dpr` are separate fixes." + "reason": "Different architectures and behaviors (speech seq2seq vs retrieval encoder); not one bug fix." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "`superpoint` and `rwkv` are unrelated model implementations." + "reason": "SuperPoint feature extraction and RWKV language modeling are unrelated code paths." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "`superpoint` and `dpr` are different code paths." + "reason": "SuperPoint and DPR are different model families; not the same underlying change." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "`rwkv` and `dpr` are separate model-specific output changes." + "reason": "RWKV and DPR are unrelated implementations; same refactor motif does not imply duplicate." } ] }, @@ -24087,7 +24061,7 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "20d013b24060ef8230c93be54ae9932adc1b688dc1d3c98ccc634e516651470a", + "prepared_review_unit_hash": "2d3fbe7891b78329a12f27bdeba09b30b5bd479c29904fe286ccfa42649bff43", "component_budget": { "node_count": 2, "item_count": 2, @@ -24109,18 +24083,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are both new model additions, but they target different models and code paths: PenguinVL vs Voxtral TTS. The overlap is mostly shared Transformers boilerplate, not the same underlying change.", - "confidence": 0.98, + "summary": "These are two unrelated model-addition PRs that just share the same Transformers integration pattern and some auto-registration files. PenguinVL and Voxtral TTS implement different models, so they should not be deduplicated.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR for a duplicate cluster here; if one representative must be chosen, 45401 is the better anchor because it is explicitly tied to issue 45027 and is a self-contained Voxtral TTS implementation, while 44662 is a separate PenguinVL add.", + "canonical_pr_reason": "PR #45401 is the more canonical standalone change: it has an explicit issue target, a narrower and well-scoped Voxtral TTS implementation, and clear intent tied to a specific model release.", "best_issue_reason": null, - "best_pr_reason": "45401 is the stronger representative PR because it has an explicit issue target and a narrower, clearly scoped implementation. 44662 is a distinct PenguinVL model PR and should not be treated as a duplicate.", + "best_pr_reason": "PR #45401 is the better representative PR for this cluster because it is tightly scoped, linked to a tracking issue, and clearly centered on one concrete model integration.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Reject: they implement different models (PenguinVL vs Voxtral TTS). Shared files are generic registry/docs updates, not evidence of the same bug or mergeable change." + "reason": "Different underlying features: PenguinVL and Voxtral TTS are separate model implementations with different codepaths, tests, and docs. Shared auto-registration and docs scaffolding are not enough to treat them as the same change." } ] }, @@ -24150,7 +24124,7 @@ "soft_pairs": [ "pull_request:44965|pull_request:46845" ], - "prepared_review_unit_hash": "a50468292278721e9f8a7aa15e5776924faa7253223347bbd04981926c94f3ce", + "prepared_review_unit_hash": "c56d0c6abf629bed27687fc470d8fa944f394d4978541e9085f46eef3b61b973", "component_budget": { "node_count": 2, "item_count": 2, @@ -24172,18 +24146,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one rewrites a PR-documentation workflow into a draft-conversion workflow, while the other changes the PR CI caller and a model utility file for a smaller CI group. They target different code paths and different intents.", - "confidence": 0.93, + "summary": "Two open PRs, but they do not look like the same underlying change. One is a draft-conversion/documentation workflow experiment; the other is a CI branch/workflow tweak with an unrelated placeholder code edit.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "Choose pull_request:46845 as the canonical PR only as the more representative item in this cluster: it has a coherent CI-related change with explicit workflow and code edits, whereas pull_request:44965 is a separate workflow tweak with a different purpose.", + "canonical_pr_reason": "No issue is present. Of the two PRs, #46845 is the better representative only because it has a clearer CI-oriented scope and a more coherent diff; #44965 looks like a separate workflow experiment with no clear problem statement.", "best_issue_reason": null, - "best_pr_reason": "pull_request:46845 is the better representative PR because its diff has a clearer, broader implementation story and touches the actual CI invocation path; pull_request:44965 appears to be an unrelated workflow experiment.", + "best_pr_reason": "PR #46845 is the strongest standalone candidate in the cluster, but only marginally; it is more explicit about a CI/runner experiment than #44965. Still, it should not be grouped with #44965 as a duplicate.", "soft_edge_verdicts": [ { "left": "pull_request:44965", "right": "pull_request:46845", "accept": false, - "reason": "Different changes: 44965 modifies a documentation/draft workflow, while 46845 adjusts the PR CI caller and a utility file for a smaller runner setup. They do not look like the same underlying bug or fix." + "reason": "Different changes: #44965 rewrites a docs workflow into a draft-conversion workflow, while #46845 changes CI branch routing and includes an unrelated placeholder edit in model code. They do not fix the same concrete problem and would not plausibly merge into one PR." } ] }, @@ -24213,12 +24187,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "d5f0b8f845839d4de8aabd8de3c540cf196cc6f237f9bc9981757423c4106d6c", + "prepared_review_unit_hash": "e0a5d2287d07b32937993cbf5eaaec6c047bdc596335b0cf55214a479c2b928d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -24226,7 +24200,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -24235,18 +24209,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are related to fusion, but they operate at different layers: PR 44979 introduces a generic Module Fusion API, while PR 45363 adds a specific n-to-1 kernel fusion path via KernelConfig and integration plumbing. They are complementary rather than duplicates.", - "confidence": 0.91, + "summary": "These PRs are related by the broad theme of fusion, but they target different changes: #44979 adds a new Module Fusion API and tests, while #45363 integrates n-to-1 kernel fusion through KernelConfig across different integration points. They are not the same underlying change and should not be deduplicated.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45363 is the more concrete, user-facing feature implementation with KernelConfig integration and example usage, so it is the better cluster representative.", + "canonical_pr_reason": "#44979 is the more foundational, self-contained API addition (new `module_fusion.py` plus tests), so it is the better anchor for this cluster, even though #45363 is a separate integration PR.", "best_issue_reason": null, - "best_pr_reason": "PR 45363 best represents the cluster because it ties the fusion idea into actual model-loading and integration code, not just the foundational API scaffold.", + "best_pr_reason": "#44979 is the clearest representative because it introduces the core module-fusion abstraction rather than a specific downstream KernelConfig use case.", "soft_edge_verdicts": [ { - "left": "PR 44979: Module Fusion API", - "right": "PR 45363: n-to-1 kernel fusion via `KernelConfig`", + "left": "pull_request:44979", + "right": "pull_request:45363", "accept": false, - "reason": "They share the general fusion theme, but 44979 adds a generic module-fusion abstraction while 45363 implements a specific kernel-fusion workflow using KernelConfig. Different scope and different code path, so not a duplicate." + "reason": "Both are fusion-related, but they implement different concrete features in different code paths: one adds a Module Fusion API, the other wires n-to-1 fusion through KernelConfig. They are not plausibly mergeable as one PR." } ] }, @@ -24276,7 +24250,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "8e879d1cde6900b6c1afb5261a89381e71f1fcbcdcc4fc1796e21eac71ae6cf7", + "prepared_review_unit_hash": "e2d48929dd036853cd6f6f7d729cce6d3048a78363fb4a142d9b0b4ce3881d27", "component_budget": { "node_count": 2, "item_count": 2, @@ -24298,18 +24272,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same AttributeError in `_patch_mistral_regex` by switching from `tokenizer.backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; PR 45317 is the more complete version because it adds tests and keeps the same fix.", - "confidence": 0.97, + "summary": "Both PRs patch the same AttributeError in `_patch_mistral_regex` with the same code change; the later PR adds tests and is the stronger canonical candidate.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same core fix as 45086 plus test coverage, making it the more complete and merge-ready implementation of the bug fix.", + "canonical_pr_reason": "PR #45317 is the better canonical PR because it contains the same fix plus test coverage, making it the more complete implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best PR overall: it fixes the same concrete code path, is more complete, and includes validation/tests beyond the minimal patch in 45086.", + "best_pr_reason": "PR #45317 is the best overall PR candidate: same underlying bug, same affected file and issue target, and it includes tests verifying the fix.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug and same code change in the same file; 45317 is essentially a follow-up/expanded version of 45086 with tests, so they are merge-compatible variants of one fix." + "reason": "Same underlying AttributeError fix in the same function and file, with effectively identical patch intent; the second PR appears to be an expanded version with tests." } ] }, @@ -24339,40 +24313,40 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "851c6d965bf7eda367c0cfd2d5997de4ce4c6857f76dc67ab3bcbb547c917880", + "prepared_review_unit_hash": "bdab1c16b4c19e920a55fe430161fe22235ee3ec9509194363ff43bdf5e4b18f", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6064, - "estimated_input_tokens": 1516, - "estimated_eval_tokens": 3288 + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6064, - "estimated_input_tokens": 1516, - "estimated_eval_tokens": 3288 + "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 related NemotronH pull requests, but they target different scopes: one is a small architecture/conversion refactor, the other is a broader model split into dense/sparse variants with docs/tests. They are not duplicates.", - "confidence": 0.83, + "summary": "These are related NemotronH PRs but not duplicates: one is a refactor to make NemotronH inherit GraniteMoeHybrid and adjust conversion mappings, while the other introduces the proper NemotronH plus dense/sparse variants, docs, and tests. They overlap in files, but they target different changes and would not plausibly be merged as one fix.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the better representative if one must be chosen: it covers the larger end-state model refactor and includes the follow-on file moves, auto-mapping, docs, and tests. PR 45115 is narrower and mainly adjusts inheritance/conversion mapping.", + "canonical_pr_reason": "PR 45569 is the more complete and canonical model-architecture change: it adds the new NemotronH dense/sparse model split, documentation, auto-mappings, and tests, whereas PR 45115 is a narrower refactor/conversion-mapping adjustment.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the strongest standalone candidate because it appears more complete and user-facing, while 45115 is a partial implementation focused on internal conversion and class wiring.", + "best_pr_reason": "PR 45569 is the best representative PR for this cluster because it describes the larger intended feature work and touches the full set of user-facing and model-definition changes.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "They share NemotronH files, but the changes are not the same concrete fix. 45115 is a narrow GranitemoeHybrid-style refactor and conversion mapping update; 45569 is a broader redesign that introduces nemotron_h_dense and nemotron_h_sparse plus docs/tests. They would not plausibly be merged as one duplicate PR." + "reason": "Shared NemotronH files alone are not enough: 45115 is a refactor/conversion-mapping patch, while 45569 is a broader model-split implementation with docs and tests. Different concrete changes, so not a duplicate pair." } ] }, @@ -24402,40 +24376,40 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "6a670855b72d558b3ff8e50284ba9c44cabaa06c1e5029da0f73da06ebe84e7f", + "prepared_review_unit_hash": "ca3931663cc2f3b28f07668d44e4fac69c5c66e132259b67f21b108a5bdfb2ce", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5954, - "estimated_input_tokens": 1489, - "estimated_eval_tokens": 3234 + "serialized_chars": 6091, + "estimated_input_tokens": 1523, + "estimated_eval_tokens": 3302 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5954, - "estimated_input_tokens": 1489, - "estimated_eval_tokens": 3234 + "serialized_chars": 6091, + "estimated_input_tokens": 1523, + "estimated_eval_tokens": 3302 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are closely related and share substantial overlap, but they are not the same concrete change: #45333 is a narrower per-layer configuration feature, while #45332 includes that plus additional heterogeneous modeling/runtime support. I would keep them separate and treat #45332 as the broader canonical PR for this cluster.", - "confidence": 0.9, + "summary": "These PRs are related but not duplicates: both introduce heterogeneous configuration docs and config utilities, but 45332 also adds the full modeling/runtime pieces (cache, masking, skip, supported models, mxfp4), while 45333 is configuration-only.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "#45332 is the more comprehensive PR: it covers the shared heterogeneous configuration work and also adds modeling/runtime integration, masking/cache changes, and broader tests/docs, so it is the better canonical representative of the feature area.", + "canonical_pr_reason": "45332 is the broader umbrella change: it includes the heterogeneous config work plus the modeling/runtime integration that 45333 does not.", "best_issue_reason": null, - "best_pr_reason": "#45332 is the best overall PR because it contains the superset of the shared work and is the most complete implementation of the heterogeneous model support feature.", + "best_pr_reason": "45332 is the best representative for the cluster because it is the more complete end-to-end heterogeneous model support PR.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "They overlap heavily on docs and config files, but #45333 appears to be only the per-layer configuration slice, while #45332 extends into modeling/runtime support and other code paths. That is related, not the same underlying change." + "reason": "Shared docs and config helpers are not enough here; 45332 adds substantial modeling, cache, and masking changes that make it a different, larger feature rather than the same mergeable PR." } ] }, @@ -24465,7 +24439,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "24528ad76273fa1e370e59d0d676570ae71c7d1b635cef4bca6a6e8838941e79", + "prepared_review_unit_hash": "99c7b366702fd8e2159db3abcd512512241a134ce4c3107c6f06a768ca7f22b5", "component_budget": { "node_count": 2, "item_count": 2, @@ -24487,18 +24461,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are typing-focused pull requests in transformers, but they address different scopes: one is a broad repo-wide type-checking pass, the other adds shared typing helpers and targeted typing fixes. They are related, not duplicates.", - "confidence": 0.91, + "summary": "Both items are typing-focused PRs, but they address different changes: PR 45415 is a broad type-checking sweep/configuration update across many files, while PR 45425 adds shared typing protocols and typing support around modeling_utils/integrations. They are related in theme but not duplicates of the same code change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45425 is the better representative because it centers on a coherent typing-infrastructure change and targeted follow-up edits, rather than a broad sweep across many files.", + "canonical_pr_reason": "PR 45425 is the better canonical representative because it introduces reusable typing abstractions in core typing helpers and then applies them to specific integration/modeling code, which is a more coherent single change than the broad repo-wide type-checking sweep in 45415.", "best_issue_reason": null, - "best_pr_reason": "PR 45425 is the strongest standalone PR in this cluster: it has a focused purpose, touches a smaller coherent set of typing-related code paths, and includes substantial review activity.", + "best_pr_reason": "PR 45425 is the best overall fit for the cluster because it is the more focused, self-contained typing change. PR 45415 is broader and mostly a repository-wide static typing pass, making it less representative of a single underlying fix/change.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-related, but 45415 is a broad type-checking pass across many modules while 45425 adds typing helpers and targeted fixes in different files. They do not appear to fix the same concrete bug/change and would not plausibly be merged as one PR." + "reason": "Both are about typing, but 45415 is a large type-checking pass/config update across many files, while 45425 adds specific typing protocols and targeted modeling/integration support. They do not appear to fix the same concrete problem or be mergeable as one PR." } ] }, @@ -24528,7 +24502,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "04c8dc216ee41f6235d615aaf45129a188550882dde4d7b26d15aa7f4ab8f2db", + "prepared_review_unit_hash": "6a75c4b755f4192839f7fdcdd46ad53ac4f96a591f75e34ce5130a7b6330d4d7", "component_budget": { "node_count": 2, "item_count": 2, @@ -24550,18 +24524,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a high level (expert parallelism), but they fix different concrete failure modes: one is about EP/DeepSpeed ZeRO-3 loading and config serialization, the other about FSDP2/rank-0 broadcast overwriting experts during parameter loading. They should not be merged as duplicates.", + "summary": "Both pull requests concern expert-parallel/distributed model loading, but they fix different failure modes: PR 45548 targets accelerate launch with DeepSpeed ZeRO-3 and config serialization, while PR 45662 targets FSDP2/rank-0 broadcast overwriting experts during load.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue is present. If one PR must represent the cluster, 45548 is the cleaner canonical example because it targets EP loading/config serialization directly and is narrower than the broader FSDP2 overwrite fix.", + "canonical_pr_reason": "No true duplicate PR exists here; if one must be chosen as the cluster representative, PR 45662 is the stronger canonical candidate because it addresses a specific concrete load-path overwrite bug with a larger, more focused fix.", "best_issue_reason": null, - "best_pr_reason": "45548 is the better global representative for the cluster because it is a self-contained EP loading bug fix with explicit config/modeling changes, whereas 45662 is a separate, more specialized distributed-loading defect.", + "best_pr_reason": "PR 45662 is the better representative for the cluster because its bug description is more concrete and its code changes are centered on the actual parameter-loading path, whereas PR 45548 is a separate accelerate/ZeRO-3 compatibility fix.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Reject: both touch expert parallelism and share modeling_utils.py, but they fix different code paths and bugs. 45548 handles EP/DeepSpeed ZeRO-3 loading via accelerate launch and config serialization; 45662 prevents experts from being overwritten by rank-0 broadcast during FSDP2 loading." + "reason": "Do not merge: they address different distributed-loading bugs and different code paths. 45548 is about EP + DeepSpeed ZeRO-3 loading via accelerate launch and config serialization; 45662 is about FSDP2 expert tensors being overwritten by rank-0 broadcast during parameter loading." } ] }, @@ -24593,7 +24567,7 @@ "pull_request:45630|pull_request:46978", "pull_request:46831|pull_request:46978" ], - "prepared_review_unit_hash": "8335b0fa993822fbaa32eff94816867d1b7936b5dfbcd893ce0b74783cd17764", + "prepared_review_unit_hash": "7a8b888ba53cd4ea88654c8119d139b531928582c56b5fdeeaa19ab6ddb39f93", "component_budget": { "node_count": 3, "item_count": 3, @@ -24615,31 +24589,34 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is not a duplicate set: one PR adds the Kimi2-6 model, one fixes a Qwen3.5/Qwen3-Next CPU-offload bug in model forwards, and one adds a broader accelerate-hook mechanism across many linear-attention models. The only overlap is the Qwen hook path, but the changes target different scopes and abstractions.", + "summary": "PR 46978 is the broad upstream fix for accelerate hook firing across linear-attention models; PR 46831 is a narrower Qwen-specific application of the same underlying hook problem, while PR 45630 is unrelated model addition work.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46978 is the best representative because it addresses the shared accelerate-hook code path in the broadest, most central way and covers the overlapping Qwen models plus other affected architectures.", + "canonical_pr_reason": "PR 46978 is the best canonical PR because it generalizes the accelerate-hook fix across affected models and subsumes the Qwen-specific hook workaround.", "best_issue_reason": null, - "best_pr_reason": "PR 46978 is the strongest standalone candidate: it proposes the most general fix for the shared hook/offload problem and has the widest impact across models, unlike the model-specific workaround in 46831 or the unrelated Kimi2-6 addition in 45630.", + "best_pr_reason": "PR 46978 is the strongest representative of the cluster since it targets the core device_map/offload hook bug directly and applies it broadly rather than only to one model family.", "soft_edge_verdicts": [ { "left": "pull_request:45630", "right": "pull_request:46978", "accept": false, - "reason": "Unrelated changes: 45630 adds Kimi2-6 docs/model wiring, while 46978 changes accelerate hook behavior for linear-attention models." + "reason": "Kimi2-6 model integration is unrelated to accelerate hook firing or linear-attention offload behavior." }, { "left": "pull_request:46831", "right": "pull_request:46978", - "accept": false, - "reason": "Related Qwen files, but not the same change: 46831 is a model-specific CPU-offload workaround, while 46978 introduces a broader hook decorator across many models and a different abstraction level." + "accept": true, + "reason": "Both address the same underlying bug: accelerate hooks not firing when Qwen linear-attention code accesses child module weights directly under CPU offload/device_map." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45630 is clearly a model-addition PR with no evidence of the hook/offload bug, and 46978 is the broader accelerate-hook fix that also touches Qwen files. The 46831\u219446978 pairing is conservative enough because both patches address the same device_map/CPU-offload hook-firing issue, with 46978 broadening the fix. No evident overstatement." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -24662,40 +24639,40 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "c063cf64c106b26d9d3e866e7e318d81d5a40341357555a4e56971095b386675", + "prepared_review_unit_hash": "7781730a063d57a95443096d2217e47e2170e9b497a020c65a1e8b6fbd2c0e5e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4673, - "estimated_input_tokens": 1169, - "estimated_eval_tokens": 2594 + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4673, - "estimated_input_tokens": 1169, - "estimated_eval_tokens": 2594 + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist-related captured_info logging problem in transformers testing utilities; PR #45645 is the more complete fix because it also updates CI log collection to read multiple worker-specific files.", - "confidence": 0.82, + "summary": "Both PRs address the same xdist-related captured_info/logging collision problem around `testing_utils.py`; one is a narrower helper fix, the other includes the helper fix plus CI artifact aggregation to handle per-worker files.", + "confidence": 0.8, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45645 is the better canonical PR because it covers the core xdist-safe file split in testing_utils and the downstream CI workflow change needed to preserve/collect the new per-worker logs.", + "canonical_pr_reason": "PR 45645 is the broader, more complete fix: it makes the log output xdist-safe and also updates CI/reporting to collect multiple `captured_info` files, so it best represents the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #45645 is the strongest representative of the change set: it fixes the artifact collision and also updates the workflow that consumes those artifacts.", + "best_pr_reason": "PR 45645 is the best single PR to keep because it covers the core bug and the downstream CI log-preservation behavior, while 45639 is just the narrower helper-side portion.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Same underlying xdist captured_info collision fix; #45645 is effectively a superset that changes the producer and the CI consumer, so these could plausibly be merged into one PR." + "reason": "They target the same underlying xdist collision in captured_info handling, share the same issue target and files, and 45645 is effectively an expanded version of the same fix." } ] }, @@ -24725,7 +24702,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "a39c02ee5a23d7e42a80e93e7966f3c5e8054f23481db9a1226bca1befd260e3", + "prepared_review_unit_hash": "b9e4e68a2e2a13d8dc3ebc36225a54f32a1746c9ea49c16659aab820871e0f57", "component_budget": { "node_count": 2, "item_count": 2, @@ -24747,18 +24724,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated linter/chore PRs in the Transformers codebase: one implements TRF003 by adjusting model forward signatures/tuple-return handling, the other implements TRF018 by changing weight-initialization patterns across many model classes. They share broad modeling filenames but not the same underlying bug or change.", + "summary": "Two unrelated linter-driven PRs overlap on model files but fix different code paths: TRF003 adjusts forward/return_dict behavior, while TRF018 refactors weight initialization. They should not be deduplicated.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45720 is the better representative of the cluster because it is the earlier, more focused implementation of a specific modeling-rule fix (TRF003) with a clearly identifiable code-path change around forward/return handling.", + "canonical_pr_reason": "No canonical PR: the changes target different modeling concerns (forward-signature handling vs weight initialization) and are not the same underlying fix.", "best_issue_reason": null, - "best_pr_reason": "PR #45720 is a stronger cluster anchor than #46259 since it has a narrower, more concrete behavioral change in model forwards and is easier to recognize as the representative linter-rule implementation.", + "best_pr_reason": "Neither PR is a strong global duplicate representative because they address different rules and different code paths; the overlap is only broad subsystem/file scope.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different lint rules and different fixes: TRF003 changes forward signatures/tuple-return behavior, while TRF018 changes weight initialization across model classes. Shared model files alone are not enough to indicate the same underlying change, and these would not plausibly be merged into one PR." + "reason": "Reject: PR 45720 implements TRF003 forward/return_dict tuple handling across model forwards, while PR 46259 implements TRF018 weight-initialization cleanup. Shared filenames are incidental; these are not the same bug/change and would not be merged as one PR." } ] }, @@ -24854,40 +24831,40 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "59147d15f53894c316b8b56a3e26e3efc39c3c1bfb7370fec3ae8733ca95cced", + "prepared_review_unit_hash": "40be774674dcfd128e5189812d09598da14f0fac616f89fcc6ab8f27f3da5693", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5161, - "estimated_input_tokens": 1291, - "estimated_eval_tokens": 2838 + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5161, - "estimated_input_tokens": 1291, - "estimated_eval_tokens": 2838 + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are both test/infra-heavy, but they address different changes: one is a small CI/telemetry and model-tuning PR, while the other is a broader test migration adding new kernel/model tests. They do not look like the same underlying fix and should not be clustered as duplicates.", - "confidence": 0.95, + "summary": "Two open WIP PRs, but they target different changes: one is a CI/infra + model/config tweak bundle, the other is a test migration adding kernel/model tests. They are not the same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46100 is the more substantial and coherent change set: it introduces a new DDP config and multiple migrated test files, whereas PR 45932 is a narrow CI/OTel tweak plus scattered model file edits. As a cluster representative, 46100 is more central.", + "canonical_pr_reason": "PR #45932 is the broader, more substantive change set in the cluster, but it is not a duplicate of #46100; the two PRs address different code paths and purposes.", "best_issue_reason": null, - "best_pr_reason": "PR 46100 is the best standalone PR in this set because it has a clear, unified purpose (test migration) and materially larger scope. PR 45932 appears to be a separate CI/debugging change and is not a better canonical representative.", + "best_pr_reason": "PR #45932 is the better representative only because it is the more general PR, but it should not be merged with #46100 as a duplicate.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Different underlying work: 45932 is a CI/observability plus minor model/config adjustment PR, while 46100 migrates and adds tests. They do not fix the same code-path problem and would not plausibly be merged into one PR." + "reason": "Different goals and edits: #45932 changes CI secrets and multiple model/config files, while #46100 adds kernel tests and test migration files. Shared test-related context is not enough to treat them as the same fix/change." } ] }, @@ -24917,12 +24894,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "b0be90ac07b552cff0f2b1cf7d254aa591163c2dd6bfbb5c8a17270c5d5b1038", + "prepared_review_unit_hash": "fa0f4a8cd54fdefc3e1ac2c7cd88acc3e6893768312863b348eb884b4ab79642", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5526, + "serialized_chars": 5525, "estimated_input_tokens": 1382, "estimated_eval_tokens": 3020 }, @@ -24930,7 +24907,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5526, + "serialized_chars": 5525, "estimated_input_tokens": 1382, "estimated_eval_tokens": 3020 }, @@ -24939,18 +24916,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: PR #46126 is a focused EP/Trainer integration change, while PR #46269 is a broad distributed-branch base spanning many unrelated distributed subsystems. The shared files are not enough to indicate the same underlying change.", - "confidence": 0.92, + "summary": "These are not duplicates: one PR is a focused EP/trainer integration, the other is a broad distributed-branch base touching many unrelated areas. Shared filenames alone are not enough to treat them as the same change.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "#46126 is the more canonical PR because it has a single, specific objective and a tight file set centered on EP/Trainer integration.", + "canonical_pr_reason": "No canonical PR: 46126 is a targeted EP/trainer feature change, while 46269 is a much broader distributed base branch with additional FSDP, CI, modeling, and integration work.", "best_issue_reason": null, - "best_pr_reason": "#46126 is the best representative PR for this cluster: it is concrete, scoped, and directly describes one change path; #46269 is too broad and infrastructure-heavy to serve as a clean canonical example.", + "best_pr_reason": "46126 is the better representative if one PR must be chosen because it has a single concrete purpose and tighter scope; 46269 is too umbrella-like to serve as the same-change representative.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "They do not appear to fix the same concrete problem or implement the same change. #46126 targets EP + Trainer integration, while #46269 is a broad distributed branch base touching many areas and likely includes multiple unrelated refactors/features." + "reason": "Shared distributed files do not imply duplication here. 46126 is EP/trainer integration around tensor parallel routing; 46269 is a broad distributed branch base with many unrelated edits, so they are not the same underlying fix/change and should not be merged as one PR." } ] }, @@ -25018,220 +24995,220 @@ "pull_request:46504|pull_request:46716", "pull_request:46432|pull_request:46716" ], - "prepared_review_unit_hash": "f9e4e897a1d1795d8a698b3111efc5e6f710fa0c5666e94bb51d0b254f4630a6", + "prepared_review_unit_hash": "ef2ac52c548ecf85c89cf474aba7b941d4730b56db230e084cc88dc0b58d4abc", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26679, - "estimated_input_tokens": 6670, - "estimated_eval_tokens": 13596 + "serialized_chars": 26671, + "estimated_input_tokens": 6668, + "estimated_eval_tokens": 13592 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 31, - "serialized_chars": 26679, - "estimated_input_tokens": 6670, - "estimated_eval_tokens": 13596 + "serialized_chars": 26671, + "estimated_input_tokens": 6668, + "estimated_eval_tokens": 13592 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a bag of unrelated workflow-maintenance PRs. The only near-duplicate-looking subgroup is pr-ci-security-gate.yml, but those PRs change different security controls and should stay separate.", - "confidence": 0.98, + "summary": "These are mostly unrelated workflow-maintenance PRs. The only close-looking set is the pr-ci-security-gate.yml edits, but they change different behaviors (logic, permissions, comments, and secrets handling) rather than the same underlying fix, so I would not merge any of the soft edges.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46554 is the most substantive and code-path-relevant change in the cluster, but it is only a representative of the pr-ci-security-gate subgroup; it is not actually duplicate to the other PRs.", + "canonical_pr_reason": "PR 46554 is the best anchor for the only substantive mini-cluster: it makes the actual pr-ci-security-gate behavior change (checkout + workflow-file checks), while the other similar PRs are distinct follow-ups or metadata/comment tweaks.", "best_issue_reason": null, - "best_pr_reason": "PR 46554 is the best representative to keep because it introduces the concrete security-gate logic change; the rest are narrower or unrelated workflow tweaks.", + "best_pr_reason": "PR 46554 is the strongest overall representative in this set because it changes runtime security-gate logic, not just YAML permissions, comments, or unrelated workflow wiring.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflow files and different behavior; title similarity alone is not enough." + "reason": "Different workflows and different problems: self-comment CI stdin handling vs pr-ci-security-gate security logic." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Unrelated workflow updates in different files; not the same underlying fix." + "reason": "Different workflows; one updates comment-pipe handling, the other changes PR CI gate permissions/approval flow." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different CI workflow concerns; no shared code-path problem." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46138", "right": "pull_request:46716", "accept": false, - "reason": "Self-comment CI vs pr-ci-security-gate are distinct workflows and fixes." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow file and different shell-handling change; not duplicates." + "reason": "Different workflows: stdin piping for a script vs a grep/printf tweak in build-ci image logic." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "SSH runner hardening versus benchmark trigger logic; unrelated changes." + "reason": "SSH-runner hardening is unrelated to benchmark_v2 dispatch behavior." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "Different workflow files and purposes; no same bug/change." + "reason": "SSH-runner hardening is unrelated to anti-slop permission placement." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "SSH runner update and reviewer assignment change are unrelated." + "reason": "SSH-runner hardening is unrelated to assign-reviewers checkout ref changes." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different security issues; not mergeable as one PR." + "reason": "Different workflow files and different security concerns; not the same concrete change." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "SSH runner hardening vs security-gate permissions; distinct changes." + "reason": "Different workflows and different fixes; one hardens SSH runner secrets, the other adjusts PR CI gate permissions." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Different workflows: anti-slop permission scoping vs benchmark trigger expansion." + "reason": "Anti-slop permissions and benchmark_v2 workflow_dispatch are unrelated." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Same general CI area, but different workflow files and changes." + "reason": "Unrelated workflow edits: anti-slop permissions vs assign-reviewers checkout ref." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Anti-slop workflow permissions and build-image shell quoting are unrelated." + "reason": "Unrelated workflow edits: anti-slop permissions vs build-ci image tagging shell change." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different goals; not the same underlying fix." + "reason": "Different workflows and different purposes." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Reviewer assignment checkout ref change is unrelated to benchmark dispatch logic." + "reason": "Unrelated workflow edits: reviewer assignment vs benchmark trigger conditions." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow files and distinct operational fixes." + "reason": "Unrelated workflow edits: reviewer assignment vs build-ci image tagging shell change." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Assign-reviewers checkout behavior is not the same as security-gate logic." + "reason": "Different workflows and different purposes." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "Benchmark dispatch trigger change and build-image message parsing are unrelated." + "reason": "Benchmark trigger logic and build-ci image tagging are different code paths and different issues." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Benchmark workflow trigger expansion vs PR security gate checks; different code paths." + "reason": "Different workflow files and unrelated changes." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Build-image commit-message handling and PR security checks are separate fixes." + "reason": "Different workflows; build-ci shell safety tweak is unrelated to PR CI security-gate logic." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "Different workflows and distinct security/trigger behavior." + "reason": "Build-ci image tagging and PR CI gate permissions/approval are distinct changes." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same file, but one adds diff-based security checks while the other changes permissions/approval gating; different fixes." + "reason": "Same workflow file, but different fixes: 46554 adds checkout plus file-change checks; 46558 adjusts permissions and approval gating." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "One changes security checks, the other adds comments and a reusable-workflow call note; not the same change." + "reason": "46554 changes security-gate behavior; 46563 only adds explanatory notes about the reusable workflow." }, { "left": "pull_request:46554", "right": "pull_request:46716", "accept": false, - "reason": "One adds local security checks, the other changes secrets inheritance; distinct problems." + "reason": "46554 adds gate checks; 46716 only changes secrets passing to the reusable workflow." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Both touch the same workflow, but they address different security-gate details." + "reason": "Permissions/approval gating changes are not the same as comment-only notes about the reusable workflow." }, { "left": "pull_request:46558", "right": "pull_request:46716", "accept": false, - "reason": "Permissions/approval gating and secrets handling are separate changes." + "reason": "Different changes in the same file: permissions/approval gating vs secrets inheritance." }, { "left": "pull_request:46563", "right": "pull_request:46716", "accept": false, - "reason": "Reusable-workflow documentation/note change is not the same as secrets scoping." + "reason": "Comments about workflow safety are not the same change as altering secrets passing." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Different workflow files and unrelated shell/security adjustments." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46432", "right": "pull_request:46563", "accept": false, - "reason": "SSH runner hardening versus reusable-workflow annotation; unrelated." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46504", "right": "pull_request:46716", "accept": false, - "reason": "Build-image shell fix and pr-ci security secrets change are unrelated." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46432", "right": "pull_request:46716", "accept": false, - "reason": "Different workflows and different security contexts; not duplicates." + "reason": "Different workflows and unrelated changes." } ] }, @@ -25324,12 +25301,12 @@ "soft_pairs": [ "pull_request:46281|pull_request:46890" ], - "prepared_review_unit_hash": "af63ea4cce1075d23b768f77524cdceeb6609ba7f43654d9a886f842480577a8", + "prepared_review_unit_hash": "b7d582f59348856da747f45a18005d2bb49b426ba2df8a206f6d941b967d1498", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4279, + "serialized_chars": 4280, "estimated_input_tokens": 1070, "estimated_eval_tokens": 2396 }, @@ -25337,7 +25314,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4279, + "serialized_chars": 4280, "estimated_input_tokens": 1070, "estimated_eval_tokens": 2396 }, @@ -25346,18 +25323,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are security hardening changes against path traversal, but they address different inputs and code paths: one validates sharded checkpoint filenames from the index `weight_map`, the other validates the `transformers_weights` config field. They are related in theme but not the same underlying bug, so they should not be merged as duplicates.", + "summary": "The two PRs are both security hardening around path traversal, but they target different loaders and different inputs: one validates shard filenames from a checkpoint index (`weight_map`), the other validates `transformers_weights` in config. They are related but not the same change.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 46281 and 46890 fix different path-traversal bugs in different loaders (`hub.py` sharded checkpoints vs `modeling_utils.py` explicit weights resolution).", + "canonical_pr_reason": "No single canonical PR: these are distinct fixes in different code paths, so neither cleanly subsumes the other.", "best_issue_reason": null, - "best_pr_reason": "46890 is the better representative of the cluster only insofar as it is a clean, self-contained security fix with explicit containment checks, but it is not a duplicate of 46281.", + "best_pr_reason": "PR #46890 is the better cluster representative only in the sense that it is a clear, self-contained hardening change; however, it is still not a duplicate of #46281.", "soft_edge_verdicts": [ { "left": "pull_request:46281", "right": "pull_request:46890", "accept": false, - "reason": "Different bug surfaces and inputs: `weight_map`-driven shard file resolution vs `transformers_weights` config resolution. Same broad security theme, but not the same concrete code-path problem." + "reason": "Different attack surfaces and code paths: shard file resolution from an index vs explicit `transformers_weights` config handling. Same broad security theme, but not the same concrete bug or mergeable change." } ] }, @@ -25387,12 +25364,12 @@ "soft_pairs": [ "pull_request:46376|pull_request:46417" ], - "prepared_review_unit_hash": "e079d7bf73d297e49ecb029075e042976b0d40d1c6df43fd74abbca3032112ca", + "prepared_review_unit_hash": "85cc4b9ce142026519acbb0eb60a903715426b1b3fb6a543d08d3004d00b1e3e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5958, + "serialized_chars": 5959, "estimated_input_tokens": 1490, "estimated_eval_tokens": 3236 }, @@ -25400,7 +25377,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5958, + "serialized_chars": 5959, "estimated_input_tokens": 1490, "estimated_eval_tokens": 3236 }, @@ -25409,18 +25386,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated pull requests: one adds multimodal support to continuous batching, and the other adds the HunYuan VL model and its docs/tests. They should not be clustered as duplicates.", - "confidence": 0.98, + "summary": "The two PRs are unrelated: one adds multimodal continuous batching support, the other adds the HunYuan VL model and docs. No duplicate or mergeable overlap is evident.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46417 is the cleaner representative: it is a focused, self-contained model addition with docs, code, and tests, whereas PR 46376 is a broad draft work-in-progress.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46417 is the best standalone PR in this set because it has a clear single purpose and appears more complete; PR 46376 is broader and still draft.", + "best_pr_reason": "PR 46417 is the better standalone anchor: it is a coherent model-addition PR with docs/tests, while PR 46376 is a draft, broader infrastructure change and not related to the same code path.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different underlying changes: continuous batching multimodal support vs. adding a specific HunYuan VL model. They touch different code paths and would not plausibly be merged into one PR." + "reason": "Different features and code paths: continuous batching/multimodal runtime changes versus adding a new HunYuan VL model package and documentation. They are not the same underlying change." } ] }, @@ -25450,7 +25427,7 @@ "soft_pairs": [ "pull_request:46574|pull_request:46580" ], - "prepared_review_unit_hash": "e7d6212c000a940dcf557189d344f470eed06da452f26b130b29e7a3ac70e037", + "prepared_review_unit_hash": "030455be56c860aadaf2f2db78eef6d19271f87accb148fbc6e38489e7afff72", "component_budget": { "node_count": 2, "item_count": 2, @@ -25472,18 +25449,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated fixes: one corrects CodeLlama decoding behavior, the other corrects DeepSeek checkpoint-to-tokenizer-class resolution. They should not be clustered as duplicates.", + "summary": "These are two unrelated pull requests: one fixes CodeLlama decode stripping leading spaces, the other fixes DeepSeek-Coder/DeepSeek-LLM tokenizer class selection. They do not look like duplicates and should not be clustered together as the same change.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: these PRs fix different bugs in different code paths and files, so neither can serve as a duplicate representative for the other.", + "canonical_pr_reason": "PR #46574 is the better canonical representative only because it is a self-contained, concrete tokenizer behavior fix with a clear code-path and expected decode regression to verify.", "best_issue_reason": null, - "best_pr_reason": "No single best PR for the cluster; PR 46574 is a tokenizer decode fix for CodeLlama, while PR 46580 is an auto-tokenizer mapping fix for DeepSeek checkpoints.", + "best_pr_reason": "PR #46574 is the strongest standalone PR in this set: it fixes a specific user-visible decoding bug in one code path and is narrowly scoped with an explicit regression target.", "soft_edge_verdicts": [ { "left": "pull_request:46574", "right": "pull_request:46580", "accept": false, - "reason": "Different underlying problems and code paths: CodeLlama decode stripping vs. DeepSeek checkpoint tokenizer-class selection." + "reason": "Different bugs and different code paths: CodeLlama decode stripping vs. model-type/tokenizer-class misclassification for DeepSeek checkpoints. They are not mergeable as one PR." } ] }, @@ -25513,7 +25490,7 @@ "soft_pairs": [ "pull_request:46603|pull_request:46604" ], - "prepared_review_unit_hash": "73832ecb8bd86511a57ae65e33e5d123b8eea39bba6830076914f8c0056c20d5", + "prepared_review_unit_hash": "b99bf71d816db838d7c3581e7889318b7e2dc4827e51abf2352f00e010a3255a", "component_budget": { "node_count": 2, "item_count": 2, @@ -25535,18 +25512,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are both Mistral-related and share files, but they implement different changes: one is a refactor/move of MistralConverter, the other adds tekken tokenizer detection, conversion, and save utilities. They are not the same underlying change.", - "confidence": 0.94, + "summary": "The two open PRs are related to Mistral tokenization infrastructure, but they address different changes: #46603 refactors/moves the existing MistralConverter into an integrations package, while #46604 adds tekken tokenizer detection, conversion, and save utilities. They overlap in files because they touch the same subsystem, not because they fix the same underlying problem.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46604 is the broader, more central change: it adds tekken tokenizer detection/conversion/save support. PR #46603 is a narrower package refactor moving MistralConverter, which looks more like groundwork than the main feature.", + "canonical_pr_reason": "PR #46604 is the broader, user-facing feature change in the Mistral tokenizer stack; PR #46603 is a packaging/refactor move of MistralConverter. They are adjacent but not the same change.", "best_issue_reason": null, - "best_pr_reason": "PR #46604 best represents the cluster because it contains the substantive Mistral tokenizer functionality change, whereas PR #46603 is mostly a structural relocation of existing converter code.", + "best_pr_reason": "PR #46604 is the best representative of the cluster because it introduces the substantive new tokenizer behavior, whereas #46603 is a structural relocation of existing code.", "soft_edge_verdicts": [ { "left": "pull_request:46603", "right": "pull_request:46604", "accept": false, - "reason": "Shared Mistral files and soft similarity are not enough: #46603 moves MistralConverter into a package, while #46604 adds tekken tokenizer detection/conversion/save utilities. Different code-path changes, not one mergeable duplicate PR." + "reason": "Different concrete changes: one moves MistralConverter into an integrations package, the other adds tekken tokenizer detection/conversion/save utilities. Shared Mistral files do not imply the same underlying bug or mergeable PR." } ] }, @@ -25576,7 +25553,7 @@ "soft_pairs": [ "pull_request:46933|pull_request:46942" ], - "prepared_review_unit_hash": "4a8ba5252af8e2e7240cf4bfdce889e69a81b0610b9bcc040fc66e3a1671feb0", + "prepared_review_unit_hash": "a5ce36c911e14dcecb4138cf6ca14d429c21d974700b2c91acf9df99b7384b75", "component_budget": { "node_count": 2, "item_count": 2, @@ -25598,18 +25575,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate pull requests with a similar quantized-dtype casting fix pattern, but they target different model families and different code paths, so they are not duplicates.", - "confidence": 0.97, + "summary": "Both PRs fix the same general quantized-weight dtype-casting pattern, but they target different model implementations and code paths (Gemma4 vs TimesFM 2.5), so they are not duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46933 is the stronger representative because it has an explicit issue link, a narrowly defined Gemma4 fix, and clear diff scope across the relevant model files.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46933 is the best overall fit for the cluster since it is the more explicit and well-scoped example of the shared dtype-casting bug pattern.", + "best_pr_reason": "PR 46933 is the stronger representative of the cluster because it has an explicit issue link and a clear, narrowly scoped fix, but it is still not a duplicate of PR 46942.", "soft_edge_verdicts": [ { "left": "pull_request:46933", "right": "pull_request:46942", "accept": false, - "reason": "Both PRs fix quantized-weight input casting, but they apply to different architectures and code paths: Gemma4 vision/audio embedders vs TimesFM 2.5 ResidualBlock. They are conceptually similar, not the same underlying change, and would not plausibly merge into one PR." + "reason": "Same broad dtype-cast/quantization theme, but they patch different model families and different classes of modules; they do not appear to fix the same concrete bug or be mergeable into one PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index d5d97878c64b48fdbcc602531b578f23867467e4..cca6df62794d6fbb369aa760f23be18953e2e8c7 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260701t000029z", + "analysis_id": "hybrid-model-20260701t060031z", "archived_artifacts": { - "hybrid": "snapshots/20260701T000029Z/analysis-runs/hybrid-model-20260701t000029z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260701T000029Z/analysis-runs/hybrid-model-20260701t000029z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-07-01T00:11:03Z", + "published_at": "2026-07-01T06:12:29Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260701T000029Z", + "snapshot_id": "20260701T060031Z", "variant": "hybrid" } diff --git a/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..96e8cd1c1478825ec5705a81d9848043f77d58e9 --- /dev/null +++ b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json @@ -0,0 +1,2579 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260701T060031Z", + "generated_at": "2026-07-01T06:12:27Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46894-6", + "summary": "Cluster of 3 issues and 3 PRs centered on issue #46894.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46894, + "canonical_pr_number": 46903, + "issue_numbers": [ + 46894, + 46897, + 46901 + ], + "pr_numbers": [ + 46895, + 46903, + 46958 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46895, + "right_pr_number": 46903, + "code_similarity": 0.362, + "size_similarity": 0.443, + "file_overlap": 0.333, + "area_overlap": 0.419, + "patch_similarity": 0.178, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 615, + 627 + ], + [ + 639, + 654 + ] + ], + "right_ranges": [ + [ + 637, + 651 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 478, + 490 + ], + [ + 502, + 517 + ] + ], + "right_ranges": [ + [ + 500, + 514 + ] + ] + } + ] + }, + { + "left_pr_number": 46895, + "right_pr_number": 46958, + "code_similarity": 0.715, + "size_similarity": 0.833, + "file_overlap": 1.0, + "area_overlap": 0.366, + "patch_similarity": 0.802, + "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": [ + [ + 615, + 627 + ], + [ + 639, + 654 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 478, + 490 + ], + [ + 502, + 517 + ] + ], + "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": 46903, + "right_pr_number": 46958, + "code_similarity": 0.325, + "size_similarity": 0.369, + "file_overlap": 0.333, + "area_overlap": 0.364, + "patch_similarity": 0.156, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 637, + 651 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 500, + 514 + ] + ], + "right_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-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-46831-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46831, + "issue_numbers": [], + "pr_numbers": [ + 46831, + 46978 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46831, + "right_pr_number": 46978, + "code_similarity": 0.113, + "size_similarity": 0.229, + "file_overlap": 0.182, + "area_overlap": 0.0, + "patch_similarity": 0.083, + "shared_filenames": [ + "src/transformers/models/qwen3_5/modeling_qwen3_5.py", + "src/transformers/models/qwen3_5_moe/modeling_qwen3_5_moe.py", + "src/transformers/models/qwen3_next/modeling_qwen3_next.py", + "src/transformers/models/qwen3_next/modular_qwen3_next.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/qwen3_5/modeling_qwen3_5.py", + "left_ranges": [ + [ + 466, + 474 + ] + ], + "right_ranges": [ + [ + 33, + 39 + ], + [ + 437, + 443 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_5_moe/modeling_qwen3_5_moe.py", + "left_ranges": [ + [ + 463, + 471 + ] + ], + "right_ranges": [ + [ + 33, + 39 + ], + [ + 434, + 440 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_next/modeling_qwen3_next.py", + "left_ranges": [ + [ + 620, + 628 + ] + ], + "right_ranges": [ + [ + 30, + 36 + ], + [ + 592, + 598 + ] + ] + }, + { + "filename": "src/transformers/models/qwen3_next/modular_qwen3_next.py", + "left_ranges": [ + [ + 461, + 469 + ] + ], + "right_ranges": [ + [ + 23, + 29 + ], + [ + 433, + 439 + ] + ] + } + ] + } + ] + }, + { + "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-46894-6", + "canonical_issue_number": 46894, + "duplicate_issue_numbers": [ + 46897, + 46901 + ], + "reason": "Issues in cluster-46894-6 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-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-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-46831-2", + "canonical_pr_number": 46831, + "duplicate_pr_numbers": [ + 46978 + ], + "target_issue_number": null, + "reason": "PRs in cluster-46831-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-46869-3", + "canonical_pr_number": 46886, + "duplicate_pr_numbers": [ + 46922 + ], + "target_issue_number": 46869, + "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46894-6", + "canonical_pr_number": 46903, + "duplicate_pr_numbers": [ + 46895, + 46958 + ], + "target_issue_number": 46894, + "reason": "PRs in cluster-46894-6 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-46894-6", + "issue_number": 46894, + "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 6 artifacts, and carries 0 discussion comments plus 6 inbound references.", + "score": 166.0 + }, + "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/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..ebcf79bc1af6f3938eee2ef95556ff4b1107e696 --- /dev/null +++ b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,25601 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260701T060031Z", + "generated_at": "2026-07-01T06:12:27Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45685|issue:46869", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:44292|issue:45237", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:45800|issue:46515", + "issue:34928|issue:41720", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:46468", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:44918|issue:45161", + "issue:43716|issue:45685", + "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": "2f811f2e920d66a277e7a03e2286cf312b291b2d0cacb07e93d3335e6e4ca450", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 63, + "serialized_chars": 46580, + "estimated_input_tokens": 11645, + "estimated_eval_tokens": 23546 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous. I would only merge two soft edges as true duplicates: the Bert/DeepSpeed ZeRO-3 IndexError pair and the Qwen3.5 MoE rotary-embedding/offload pair. The remaining pairs are related by broad subsystem or wording, but not the same concrete bug/change.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43596 is the best canonical anchor for the ZeRO-3/Bert IndexError subcluster: it states the concrete failure mode clearly, and issue:43638 reads like the same bug with a slightly narrower phrasing.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 is the strongest representative of the only clearly duplicated issue pair and is the most actionable target for de-duplication.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures in the Qwen3.5 family: missing config attribute vs streaming chat/completions behavior." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Both mention hardware/backend problems, but one is GPT-OSS on AMD GPUs and the other is an MPS histogram op error; not the same bug." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different model-specific failures: Mistral image/text dtype mismatch vs GPT-OSS GPU backend issue." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Both are dtype/device related, but they affect different models and code paths." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "Unrelated model/example failures; no shared concrete failure mode." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different MoE/model problems: AMD GPU execution vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export error and FineGrainedFP8 failure are separate model/tooling issues." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "Export-time ONNX failure and auto device-map CUDA assert are different code paths." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash vs RAG documentation freshness; not the same bug." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export issue and GenerationMixin import error are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 doc cleanups, but they remove different references and would not be one concrete fix." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding unpacking failure is unrelated to resuming training with trust-remote-code." + }, + { + "left": "issue:46165", + "right": "issue:46869", + "accept": false, + "reason": "Resume/trust-remote-code bug is unrelated to YOLO multi-GPU device-map failure." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask-shape IndexError and Gemma4 PLE device mismatch are different forward-pass errors." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexErrors, but one is ZeRO-3/Bert initialization and the other is a mask/tensor shape mismatch." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "ZeRO-3 Bert initialization bug is unrelated to resume/trust-remote-code behavior." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different IndexError families: ZeRO-3 empty tensor initialization vs mask/tensor shape mismatch." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Different multimodal failures: mask indexing vs video token/feature mismatch in chat template." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask shape mismatch and GPT-OSS tensor-parallel failure are different problems." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Training resume issue and Nemotron example breakage do not share a concrete failure mode." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration in Qwen3-vl-embedding is unrelated to training resume with remote code." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice preset bug and ModernBERT ONNX export are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "DeepSpeed collective mismatch is not the same as ZeRO-3 Bert initialization IndexError." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying ZeRO-3/Bert out-of-bounds IndexError during model initialization/loading; the titles describe the same failure with minor wording differences." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross-entropy call issue is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different defects: DeepSeek detokenization in v5 vs LoRA merge/save/reload mismatch." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA merge/save inconsistency are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class/config_class attribute errors." + }, + { + "left": "issue:43716", + "right": "issue:46869", + "accept": false, + "reason": "Mistral dtype mismatch and YOLO multi-GPU device-map failure are different model/back-end bugs." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "FSDP activation-checkpointing tensor recompute error is unrelated to the mask indexing bug." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed training weight-shape error and DeepSpeed collective mismatch are different failures." + }, + { + "left": "issue:45685", + "right": "issue:46869", + "accept": false, + "reason": "MPS histogram op error is unrelated to YOLO multi-GPU device-map behavior." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": true, + "reason": "Same Qwen3.5 MoE rotary-embedding/device-materialization bug: one report says CPU offload incompatibility, the other identifies uninitialized inv_freq after meta\u2192to_empty(cuda)." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 Bert IndexError and GPT-OSS TP failure are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation-checkpointing/FSDP recomputation mismatch is unrelated to CustomTrainer.compute_loss signature breakage." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 Bert initialization bug is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific issues: Mistral dtype mismatch vs GraniteMoEHybrid invalid method." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Both are Gemma4/Gemma-related multimodal issues, but one is an expensive reverse-embedding lookup and the other is a video token/feature mismatch." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure and GPT-OSS AMD GPU failure are different." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config attribute missing for vLLM and Mistral tokenizer regex patch crash are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Different Gemma4 bugs: device mismatch in forward vs costly reverse embedding lookup." + }, + { + "left": "issue:45800", + "right": "issue:46515", + "accept": false, + "reason": "Both are torch 2.4.1 incompatibility reports, but they are not clearly the same concrete failure and may target different stacks/versions." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation-checkpointing issue and Qwen3 auto device-map CUDA assert are different problems." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "General torch incompatibility and Qwen-3-8B-NVFP4 runtime failure are not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Phi-tiny-MoE autocast dtype mismatch and GPT-OSS TP failure are different issues." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral torch-dynamo grouped_mm error and Mistral dtype mismatch are different code paths." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different model/runtime failures; only a broad 'inference issue' similarity." + }, + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Mistral dtype mismatch is unrelated to Nemotron example breakage." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 error and TRL SFT embedding unpacking are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and MPS histogram op error are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Mistral dtype mismatch and Qwen-3-8B-NVFP4 runtime error are distinct." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL SFT embedding unpacking failure is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral dtype mismatch and MPS histogram op error are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross-entropy issue is unrelated to TRL SFT embedding unpacking." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading IndexError and Bert ZeRO-3 loading IndexError are different model families." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions in v5, but one is about pipeline-task docs and the other about the run command; separate fixes." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Phi-tiny-MoE autocast dtype mismatch and MPS histogram op error are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Liger Kernel cross-entropy path and Phi-tiny-MoE autocast mismatch are different issues." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and trainer loss-signature breakage are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP activation-checkpointing mismatch and GenerationMixin import error are unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "GPT-OSS AMD GPU failure and RAG docs issue are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross-entropy call and mask shape mismatch are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and the soft-edge calls stay conservative. The two accepted pairs are well supported by the titles, and the remaining rejections avoid overclaiming duplicate status." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 2, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720", + "issue:43901|issue:44512", + "issue:44589|issue:46304", + "issue:45237|issue:45901", + "issue:43975|issue:46015", + "issue:43039|issue:44560", + "issue:45823|issue:46294", + "issue:44841|issue:45823", + "issue:43866|issue:46772", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:45812|issue:45920", + "issue:43866|issue:44863", + "issue:44292|issue:46015", + "issue:43872|issue:46304", + "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:45362|issue:45464", + "issue:43792|issue:45070", + "issue:43596|issue:45800", + "issue:43541|issue:46869", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:36010|issue:39290", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:43404|issue:43866", + "issue:45507|issue:46869", + "issue:41093|issue:41762", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823" + ], + "prepared_review_unit_hash": "8cf457eb55708644dd316cc1c5fc21b8c60d1dd6ff4e5336dfa29679d6b15cf5", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 40043, + "estimated_input_tokens": 10011, + "estimated_eval_tokens": 20278 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a noisy mix of unrelated issue reports. I only see two believable duplicate pairs: the v5 AutoTokenizer wrong-token-id regression, and the DeepSpeed ZeRO-3 index-out-of-bounds loading failure. The rest look like separate bugs, model-specific failures, or docs/version-compatibility issues rather than the same underlying change.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 45920 is the best canonical anchor for the clearest duplicate subcluster: it describes the v5 AutoTokenizer token-ID regression broadly across multiple model families, so it subsumes the narrower Granite-only report (45812) better than the reverse.", + "canonical_pr_reason": null, + "best_issue_reason": "45920 is the strongest representative issue in this set because it captures the widest shared failure mode with explicit regression framing and multiple affected models. The zero3 pair is separate and should not be merged into the same canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Both are model loading/weight issues, but one is a corrupted checkpoint claim and the other is missing text-encoder weights; different model families and likely different root causes." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Both mention version compatibility, but one is a torch 2.4.1 / transformers 5.8.0 incompatibility and the other is a transformers 5.9.0 distributed-build regression; not the same bug." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and USE_DISTRIBUTED=0 support regression are unrelated failure modes with different stacks and symptoms." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "These describe the same ZeRO-3 out-of-bounds index failure during model loading/init; the model names differ, but the underlying bug appears to be the shared zero-sized tensor/indexing path." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3TextConfig missing sliding_window_pattern in vLLM and Qwen3 auto-device-map cudaErrorAssert are different model/runtime failures." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "One is a pipeline docs mismatch and the other is outdated docs about a removed command; same documentation area, but not the same issue." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup failure and distributed support breakage are separate runtime incompatibilities." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "AMD GPU execution failure for GPT-OSS-20B is unrelated to the table-question-answering crash." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Incorrect detokenization and outdated RAG examples are different problem types with no shared code path." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross_entropy dispatch and Qwen3-vl video StopIteration are unrelated issues." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Device-mapping mismatch in Gemma4 PLE and Qwen3.5 video token/feature mismatch are different bugs." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral and Gemma4 PLE device mismatch are not the same underlying defect." + }, + { + "left": "issue:43866", + "right": "issue:46772", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and DiffusionGemma ModelOpt compatibility are unrelated." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading and outdated RAG docs are unrelated." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code breakage in recent transformers and distributed-build support regression are different compatibility problems." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "These are the same AutoTokenizer wrong-token-ID regression in v5; 45920 is just the broader multi-model report." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint versus inability to load checkpoints are superficially similar but different models and failure mechanisms." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and outdated RAG examples have no shared code path." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params constructor incompatibility is unrelated to the distributed support regression." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but one is a missing config attribute in vLLM and the other is a Qwen3-MoE model issue; not clearly the same bug." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are index errors under DeepSpeed ZeRO-3, but the reported failures are different (mask shape mismatch vs zero-sized Bert init path)." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper-large-v2 failing to run and table-question-answering crashing are unrelated issues." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading failure and Ovis2 corrupted checkpoint are different model-loading problems." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 compatibility and pydantic PretrainedConfig field breakage are unrelated versioning issues." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Incorrect detokenization and table-question-answering crash are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Qwen3.5 chat crash and streaming chat/completions API failure are different Qwen-related symptoms with different surfaces." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure and pydantic model-field breakage are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3 index failure and torch/transformers incompatibility are different bug classes." + }, + { + "left": "issue:43541", + "right": "issue:46869", + "accept": false, + "reason": "Mixtral grouped_mm tracing error and YOLO multi-GPU device-map failure are unrelated model/runtime issues." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Qwen NVFP4 execution error and table QA crash do not share an underlying code path." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 sliding_window_pattern missing under vLLM and Gemma4 device mismatch during forward are different failures." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import error and Gemma3 sliding_window_pattern error are unrelated API breakages." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming chat/completions failure and MPS histogram_not_implemented error are different runtime issues." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Mixtral grouped_mm tracing and MPS histogram failure are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different code paths and symptoms." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure and non-template-node compile error are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 weight-tying bug and Ovis2 checkpoint corruption are different model-specific loading problems." + }, + { + "left": "issue:45507", + "right": "issue:46869", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU device-map failure are separate issues." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are erroring under different loading scenarios, but the mask-shape mismatch and zero-sized index error are not the same defect." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "Flash-attn fallback gating and remote-code config_class AttributeError are different compatibility checks." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and pydantic PretrainedConfig field breakage are unrelated." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility and Gemma4 device mismatch are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The two accepted soft pairs are well supported by the issue titles and descriptions: the v5 AutoTokenizer token-ID regression pair and the DeepSpeed ZeRO-3 out-of-bounds loading failures. The other soft-edge rejections are cautious and do not overclaim duplication from superficial similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 3, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "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: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:44387|issue:45005", + "issue:43828|issue:45464", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:43531|issue:45070", + "issue:41093|issue:43596", + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43054|issue:43994", + "issue:44479|issue:46294", + "issue:46468|issue:46869", + "issue:45759|issue:45902", + "issue:43596|issue:46165", + "issue:46640|issue:46661", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:40990|issue:44960", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44530|issue:45005", + "issue:44492|issue:44512", + "issue:44863|issue:45005", + "issue:45183|issue:46463", + "issue:43975|issue:44292" + ], + "prepared_review_unit_hash": "e6e882216b63a48d342786db6f5852ef03b02bc125eaa543fd9c0a6e1d8f509d", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36792, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab bag of unrelated transformers issues, with a few real duplicate pockets around input validation, Qwen video handling, and SigLIP2 regressions.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45183 is the best umbrella representative: it describes the generic input_ids/inputs_embeds validation bug that the T5-specific report (46463) appears to instantiate.", + "canonical_pr_reason": null, + "best_issue_reason": "45183 is the broadest, most reusable issue in the set and captures a concrete validation-path bug rather than a model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different failures: one is a model load problem, the other is GGUF architecture support for gpt-oss." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both mention indexing problems, but the titles point to different models/code paths and the same root cause is not clear." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both are MoE-related, but one is a dynamo grouped_mm tracing failure and the other is a tensor-parallel support issue." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "One is an int4 quantization OOM regression; the other is a tied-weights issue on translation models." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Different modalities and symptoms: autocast dtype mismatch versus streaming chat/completions failure." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "No shared bug path; gpt-oss GGUF support and Qwen streaming inference are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "One is about init_empty_weights argument handling, the other about a pydantic PretrainedConfig field regression." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the pydantic model breakage." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors, but the specific failure modes and contexts differ too much to treat as the same bug." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Image-preprocessor dtype mismatch and video token/feature mismatch are different multimodal bugs." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Remote_code breakage and torch/version incompatibility are too broad and not the same concrete failure." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 regression: both reports describe broken/nonsensical outputs from the same model family and likely the same text-embedding path." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "46294 is a specific instance of the broader Qwen video-input regression described in 44479, with the same video feature/token mismatch symptom." + }, + { + "left": "issue:46468", + "right": "issue:46869", + "accept": false, + "reason": "Completely different areas: a Nemotron example issue versus YOLO multi-GPU device-map behavior." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "One is about unwrapping text_config for composite configs; the other is an initialization/memory-materialization bug in Qwen3_5MoeTextRotaryEmbedding." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "Different remote-code/deepspeed resume problems; not the same underlying bug." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 support and remote-code model_class/config_class handling are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different failures: dtype mismatch in a multimodal processor versus embedding unpacking in TRL SFT training." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight shape error and Qwen3.5 embedding unpacking are not the same code-path problem." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss and a GLM5 issue are unrelated model-specific reports." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback gating are different regressions." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both touch trust_remote_code caching, but one is a leaf-directory cache collision while the other is wrong custom-module loading after save_pretrained." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "A missing lm_head weight tie in Mistral3 is not the same bug as translation-model tied weights in v5." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FP8 loading failure and flash-attn fallback gating are separate issues." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior is not the same as tied-weights regressions." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer discussion is unrelated to a docs typo about the removed run command." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention failure and tied-weights regressions are different concrete bugs." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo does not match a documentation command removal issue." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading and translation-model tied weights are unrelated." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": true, + "reason": "46463 is a specific model-family manifestation of the broader input_ids/inputs_embeds XOR validation bug described in 45183." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Incorrect detokenization and an unexpected _is_hf_initialized argument error are different failure modes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the accepted soft edges are conservative. The 45183/46463 duplicate call is well supported by the titles, and the other pair rejections avoid overstating weak similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 4, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45520", + "issue:45663", + "issue:45701", + "issue:45859", + "issue:45902", + "issue:45920", + "issue:45987", + "issue:46489", + "issue:46491", + "issue:46858", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:44918|issue:45507", + "issue:46489|issue:46491", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:36331|issue:41762", + "issue:46899|issue:46932", + "issue:43872|issue:44589", + "issue:43828|issue:44928", + "issue:45920|issue:46489", + "issue:45987|issue:46858", + "issue:43054|issue:45701", + "issue:38175|issue:43054", + "issue:43950|issue:45902", + "issue:41553|issue:43883", + "issue:43054|issue:43493", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:45663|issue:45902", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:24643|issue:28530", + "issue:42947|issue:43856", + "issue:44403|issue:44488", + "issue:43643|issue:45020", + "issue:41553|issue:43881", + "issue:41762|issue:42915", + "issue:44991|issue:45399", + "issue:43782|issue:44960", + "issue:43824|issue:43872" + ], + "prepared_review_unit_hash": "714728aac67122dfa6f46e4a7dbc12c5b6d342ada7adf97e41098d58868bb8a3", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36857, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous issue reports, with only a few clear duplicate-style pairs. The strongest duplicate cluster is around the transformers 5.x non-persistent buffer regression; a Gemma4 quantized-embedder pair also looks identical. Most other pairs are related only at a broad subsystem level and should stay separate.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest, most concrete regression report in the set: a transformers 5.x non-persistent-buffer corruption bug with a very similar neighboring report (#44534) and another close buffer-related report (#43644).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the best overall representative because it is specific, actionable, and sits at the center of the only obvious duplicate-like issue cluster here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different failures and code paths: Qwen3.5/TRL embedding unpacking vs a GraniteMoEHybrid invalid method call." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but for different models and with different symptoms/root causes." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying transformers 5.x non-persistent-buffer regression; both describe junk/corruption in buffer materialization/loading." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Docs-example model/processor mismatch and quantization failure is unrelated to a general SigLIP2 text-embedding quality issue." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same underlying bug class: transformers 5.x mishandling of non-persistent buffers leading to junk/corruption." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "CustomTrainer API keyword breakage is unrelated to Gemma3 DeepSpeed ZeRO-3 loading failure." + }, + { + "left": "issue:46899", + "right": "issue:46932", + "accept": true, + "reason": "These appear to be the same Gemma4 quantized vision/audio embedder dtype-cast bug; #46932 explicitly points to the same fix." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated quantization/backend failures: bitsandbytes Int8Params kwargs vs missing Float8 storage lookup." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Both involve dtype/precision problems, but on different models and different execution paths." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": false, + "reason": "A broad multi-model tokenizer regression is not the same as a DeepSeek-Coder v1-specific detokenization bug." + }, + { + "left": "issue:45987", + "right": "issue:46858", + "accept": false, + "reason": "Both mention get_seq_length, but one is a return-type contract issue and the other is a backend-specific tensor/device arithmetic failure." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "Text-embedding quality regression is unrelated to a version-induced tokenization change." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Unexpected zero probabilities and worse text embeddings are different symptoms with different implied causes." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Both are buffer-related, but one is a from_pretrained regression while the other is a specific meta\u2192to_empty(cuda) uninitialized-memory issue." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad error message loading Voxtral is unrelated to MolmoForCausalLM missing all_tied_weights_keys." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Model quality degradation is not the same as an HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Flash-Attention illegal-memory crash from 3D position_ids is unrelated to CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "LoRA/gradient-checkpointing effectiveness and int4 quantization memory growth are different bugs." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states handling is unrelated to Qwen3_5MoeTextRotaryEmbedding buffer initialization." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise is not the same underlying bug as a specific ScandiBERT load failure." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Tokenizer detokenization regression and a flash_attn import KeyError are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "DeepSpeed training shape error and early-stopping metric configuration are separate concerns." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different memory/training issues in different models and setups; not the same concrete bug." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both involve loading, but one is generic noise and the other is a specific model load failure." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Remote_code breakage is too broad to treat as the same bug as missing fields in AutoConfig.from_pretrained without stronger evidence." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different models and different failures: Voxtral tokenizer error messaging vs glm-4v-9b loading failure." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 load failure and Qwen3Moe FP8 config failure are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure and flash-attn fallback gating are separate issues." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight-only loading error is unrelated to a GLM5 issue with no evidence of the same root cause." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Importing a missing VL class and a bitsandbytes Int8Params incompatibility are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only true soft-edge pairs are the ones with strong title/body alignment around the non-persistent-buffer regression and the Gemma4 quantized-embedder dtype-cast bug; the rest are appropriately rejected as only loosely related. No overclaiming 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": 25, + "review_unit_index": 5, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:45685", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45923", + "issue:46315", + "issue:46489", + "issue:46661", + "issue:46869", + "issue:46894", + "issue:46897" + ], + "soft_pairs": [ + "issue:45685|issue:45823", + "issue:45081|issue:45399", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43881|issue:43883", + "issue:42491|issue:44368", + "issue:43828|issue:45823", + "issue:45020|issue:45356", + "issue:45715|issue:46315", + "issue:45488|issue:46489", + "issue:43827|issue:44509", + "issue:43872|issue:46869", + "issue:45759|issue:46661", + "issue:46894|issue:46897", + "issue:43404|issue:43950", + "issue:43872|issue:44291", + "issue:45020|issue:46661", + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:36331|issue:41093", + "issue:41553|issue:44991", + "issue:43854|issue:44863", + "issue:43940|issue:45313", + "issue:44928|issue:45588", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:24643|issue:30064" + ], + "prepared_review_unit_hash": "c4c2ca8040072c029c9ce3aa63d5f63c1b35bcc641863424db22cba2a2c45fc8", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36753, + "estimated_input_tokens": 9189, + "estimated_eval_tokens": 18634 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most of these items are only loosely related by subsystem; only a few pairs are true duplicates with the same concrete failure mode. The clearest duplicate clusters are the v5 docs cleanup, the `_is_hf_initialized`/`init_empty_weights` incompatibility, the mask-shape mismatch, and the training-loss double-shift bug pattern.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43872 is the best canonical anchor for the `_is_hf_initialized` / `init_empty_weights` incompatibility cluster because it names the concrete bitsandbytes failure and has the clearest external compatibility context.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44291 is the most actionable issue in that same cluster because it states the exact trigger (`init_empty_weights`) and the version scope (`transformers >= 5.0.0rc0`).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "Both are runtime device/backend problems, but they describe different code paths and symptoms." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex crash vs flash-attn fallback gating are different bugs." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote-code breakage and device-map mismatch are unrelated failures." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Different problems in model/tokenizer registration and remote-code loading." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different model-loading failures with different missing attributes and code paths." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen/LoRA/tied embeddings, but one is a model usability regression and the other is a warning/flag issue, not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: autocast dtype mismatch vs device placement mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote-code regression and tokenizer codec handling regression are separate issues." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they affect different conversion directions and likely different hot paths." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but the titles point to different underlying bugs and model families." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs cleanup: both refer to removed v5 pipeline tasks still mentioned in documentation." + }, + { + "left": "issue:43872", + "right": "issue:46869", + "accept": false, + "reason": "Bitsandbytes/init_empty_weights incompatibility is unrelated to YOLO multi-GPU device-map behavior." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "Different auto-model loading issues: composite config handling vs remote-code config_class lookup." + }, + { + "left": "issue:46894", + "right": "issue:46897", + "accept": true, + "reason": "Same training-loss double-shift pattern reported for two models; a single fix in the shared loss path could plausibly cover both." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Weight tying bug and non-persistent buffer corruption are different regression classes." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same concrete incompatibility: loading with `init_empty_weights` passes unexpected `_is_hf_initialized` through the same code path." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Both involve remote-code models, but the failures are different and would not necessarily share one fix." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper model runtime failure and `_torch_extract_fbank_features()` issue are too broad to confirm as the same underlying bug from the titles alone." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Both mention DeepSpeed, but one is weight loading and the other is gradient averaging; different bugs." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both touch gradient checkpointing, but one is ineffective LoRA checkpointing and the other is a missing capability flag." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different errors and different training-time failure modes." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading issues, but for different models and different root causes." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and different checkpoint-loading failures." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 weight-loading problems, but the reported symptoms and model layouts differ too much to treat as one bug." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "RLHF NaN/SDPA collapse and flash-attn sink-less AttributeError are unrelated." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask/tensor shape mismatch error pattern; these look like duplicate reports of the same indexing bug." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both concern video input regressions, but the titles point to different model-specific failures and not one clearly shared bug." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Training with DeepSpeed and void segmentation map processing are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The accepted pairs are supported by the titles, especially the docs cleanup, `_is_hf_initialized`/`init_empty_weights`, and the mask-shape mismatch cluster. No soft-edge verdict seems to overstate identity beyond what the packet supports." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 6, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:41553", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43632", + "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:46489", + "issue:46515", + "issue:46858", + "issue:46899" + ], + "soft_pairs": [ + "issue:44530|issue:44863", + "issue:41553|issue:44488", + "issue:44863|issue:46402", + "issue:43366|issue:43828", + "issue:45325|issue:45902", + "issue:43366|issue:43531", + "issue:43596|issue:46515", + "issue:44155|issue:45584", + "issue:45663|issue:45823", + "issue:43638|issue:46515", + "issue:43366|issue:44589", + "issue:43950|issue:45834", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:44910|issue:46858", + "issue:43716|issue:44560", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:45685", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44560|issue:44918", + "issue:44683|issue:46515", + "issue:45812|issue:46489", + "issue:45823|issue:46515", + "issue:44661|issue:45823", + "issue:38175|issue:43994", + "issue:44560|issue:44805", + "issue:45672|issue:46899", + "issue:43632|issue:45834", + "issue:44877|issue:45030", + "issue:42491|issue:43720", + "issue:43638|issue:44661", + "issue:43596|issue:44661" + ], + "prepared_review_unit_hash": "dad94451779cb27e8ade854c5b9763296695f17633f25337dc8a4b95fad81d1f", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38061, + "estimated_input_tokens": 9516, + "estimated_eval_tokens": 19288 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of Transformers v5/model-loading regressions, but only one soft pair looks like the same underlying bug: custom model initialization broken in 5.0.0. The rest are different model-specific failures or unrelated backend/device issues.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the most general and representative report of the shared regression: Transformers 5.0.0 breaking custom model initialization. It subsumes the notebook-specific variant in 43645.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single issue to anchor the cluster because it states the core bug broadly and is most reusable for deduplication.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading-related at a high level, but one is a Qwen3.5 cache/runtime error and the other is a NemotronH checkpoint-loading failure. Different underlying bugs." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both mention loading problems, but they affect different models and likely different code paths; the titles do not indicate the same concrete bug." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Different models and failure modes: NemotronH checkpoint loading versus Sam3Model text encoder weights not being loaded." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Unrelated issues: GGUF architecture support versus an autocast dtype mismatch in Phi-tiny-MoE." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Both involve rotary/position embedding internals, but they are different models and different failure mechanisms." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "One is GGUF gpt-oss support, the other is a Qwen3-MoE sliding_window bug; not the same bug." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "One is a DeepSpeed ZeRO3 Bert index error; the other is a generic Torch compatibility issue. Too different to dedupe." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Both are incorrect-generation style bugs, but one is AudioFlamingo3 batching leakage and the other is Whisper empty-transcription handling." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states bug versus Gemma4 PLE device mismatch; same model family, different concrete failure paths." + }, + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "Both involve loading/runtime regressions, but one is Bert/ZeRO3 initialization and the other is a Torch version incompatibility. Not the same bug." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF gpt-oss support and missing Float8 storage are unrelated." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Persistent-buffer corruption in from_pretrained is a different class of loading bug from Kosmos2.5 long OCR index errors." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Both are loading regressions, but one is a missing class import and the other is meta-device loading breakage. Different concrete problems." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure is not the same as generic meta-device loading breakage." + }, + { + "left": "issue:44910", + "right": "issue:46858", + "accept": false, + "reason": "Flash-Attention 3D position_ids corruption on Qwen3.5 is different from static-cache seq-length tensor arithmetic on non-CUDA backends." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 image preprocessor dtype mismatch and Qwen3-vl StopIteration are different multimodal bugs." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Both describe the same Transformers 5.0.0 regression breaking custom model initialization; 43645 is just the Jupyter-specific manifestation." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Torch.compile/SDPA on Bamba is unrelated to Qwen3.5 Flash Attention position_ids handling." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-vl video StopIteration and MPS histogram on Int are unrelated failures." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime failure is not the same as GGUF gpt-oss support." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 failure are completely different issues." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Qwen3-vl video StopIteration and generic mask shape mismatch are different bugs." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "Compiled flex_attention on torch >= 2.9 is a different backend-specific issue from torch 2.4.1 incompatibility." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Granite AutoTokenizer token-ID regression and DeepSeek-Coder tokenizer regression are separate tokenizer bugs." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 PLE device mismatch is not the same as the torch 2.4.1 incompatibility report." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping failure in add-new-model-like and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": false, + "reason": "Both mention SigLIP2 outputs, but one is zero probabilities and the other is nonsensical AutoModel/pipeline results; not clearly the same concrete bug." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Qwen3-vl StopIteration and mask/index shape mismatch are different failures." + }, + { + "left": "issue:45672", + "right": "issue:46899", + "accept": false, + "reason": "Both concern Gemma4 and quantization, but one is uint8 torch.finfo in audio modules and the other is unsafe dtype casts in multimodal embedders; different code paths." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized flag regression and Kosmos2.5 long OCR index error are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Strict config rejection for granite_speech and glm4v tiny-random config validation are different config-validation issues." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility and BitNet packed-weight loading are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO3 Bert index error and tokenizer-mapping-name handling are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "A ZeRO3/Bert initialization index error is not the same as add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted soft pair, issue:43645 \u2194 issue:43646, is well-supported by the titles and cluster context, and the remaining rejections do not overstate similarity. The choice of issue:43646 as the canonical anchor is reasonable because it is the broader report of the same Transformers 5.0.0 custom model initialization regression." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 7, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "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:44910", + "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:45588", + "issue:45812", + "issue:46050", + "issue:46459", + "issue:46468", + "issue:46512", + "issue:46846", + "issue:46897", + "issue:46901" + ], + "soft_pairs": [ + "issue:43317|issue:44756", + "issue:44661|issue:44805", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:41720|issue:42915", + "issue:42491|issue:45276", + "issue:42915|issue:44756", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:44863|issue:46846", + "issue:46459|issue:46512", + "issue:45313|issue:45357", + "issue:45507|issue:46468", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:39692|issue:43994", + "issue:43257|issue:43854", + "issue:43856|issue:43873", + "issue:42491|issue:45292", + "issue:43854|issue:43931", + "issue:43856|issue:44387", + "issue:43708|issue:44075", + "issue:42491|issue:45335", + "issue:44960|issue:46050", + "issue:46897|issue:46901", + "issue:44164|issue:44291", + "issue:45362|issue:45406", + "issue:45488|issue:45812", + "issue:45276|issue:45335", + "issue:44910|issue:45588" + ], + "prepared_review_unit_hash": "6ea19ae15da147eb16df3e2e87d9f0df3163757cb3e82575f70735120c6f895a", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36654, + "estimated_input_tokens": 9164, + "estimated_eval_tokens": 18584 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but it contains tight duplicate families around tokenizer-loading regressions in v5, non-persistent buffer initialization, resize_token_embeddings propagation, and training-loss double-shift. The SigLIP2 items are only loosely related and should not be treated as a tight duplicate family.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 44991 is the clearest canonical representative because it is a specific, well-scoped transformers>=5.0.0 tokenizer-loading regression with a reproducible model impact and strong fit to the tokenizer-family duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44991 is the best global issue representative: it is concrete, narrowly framed, and anchors the strongest duplicate subgroup in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Both are offload/memory related, but they describe different failures and likely different fixes: dequantized model loading on gpu+cpu offload vs disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES lookup failure and mask/tensor shape mismatch are unrelated code paths." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is inference output corruption and the other is a documentation example/model-processor mismatch with quantization failure; not the same underlying bug." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Both report the same Transformers v5 non-persistent buffer regression; the titles are essentially the same bug phrased differently." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3Moe FineGrainedFP8Config failure are different concrete failures." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "LoRA hf4.x to hf5.x incompatibility and resize_token_embeddings not updating output embeddings are unrelated bugs." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "FineGrainedFP8Config on Qwen3Moe and Strix Halo mmap OOM are different problems." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention issues do not share the same underlying failure." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Both describe the same tokenizer-loading regression in v5, expressed as model-specific load failures." + }, + { + "left": "issue:44863", + "right": "issue:46846", + "accept": false, + "reason": "NemotronH checkpoint loading failure and missing linear_attn TP-plan entries are different bugs." + }, + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Ignoring dtype during multimodal checkpoint loading and stale config.dtype after from_config are related but distinct issues." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "ZeRO-3 language_model weight loading and incorrect visual-encoder keys in save_pretrained are separate serialization/loading bugs." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and broken Nemotron example are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": true, + "reason": "Both describe the same tokenizer-loading regression in v5, expressed as model-specific load failures." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Non-persistent buffer junk and Jupyter custom-model initialization breakage are different regressions." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "The first is a doc example/model-processor mismatch plus quantization failure; the second is inference corruption on AutoModel/pipeline. Similar model, different bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and GLM-4.7-Flash unit-test load failure are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency and quantization offload not working are different concrete issues." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA hf4.x/hf5.x compatibility is unrelated to resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "GLM-4.7-Flash test loading and Qwen3-VL weight-shape mismatch are different model-loading failures." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Training memory inefficiency and increased reserved memory under int4 quantization are not the same bug." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint-resume max_steps bug and SGD args ignored are unrelated training issues." + }, + { + "left": "issue:42491", + "right": "issue:45335", + "accept": false, + "reason": "LoRA model incompatibility and t5gemma resize_token_embeddings propagation are different problems." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 issue and sinq save/load breakage are unrelated." + }, + { + "left": "issue:46897", + "right": "issue:46901", + "accept": true, + "reason": "Both are training-loss double-shift bugs: labels are shifted one token too far in the loss path, just on different models." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "extra_state handling and _is_hf_initialized argument handling are different save/load regressions." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are different failures." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one overrides ByteLevel with Metaspace for DeepSeek V3/R1 and the other produces wrong token IDs for Granite models; not the same code-path bug." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "These are the same resize_token_embeddings propagation bug on different model variants: output embeddings are not updated." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Illegal memory access from misread 3D position_ids and AttributeError on s_aux=None are different flash-attention failures." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is slightly overstated. The tokenizer-loading, non-persistent buffer, resize_token_embeddings, and double-shift families are grounded, but the SigLIP2 portion is too strong: the packet supports related SigLIP2 reports, not a clearly established tight duplicate family, and the analyst explicitly rejected the relevant pairings. Please downgrade that to a looser similarity claim or omit it." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 8, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42796", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44079", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44756", + "issue:44863", + "issue:44877", + "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:44402|issue:44488", + "issue:39692|issue:44960", + "issue:41720|issue:44155", + "issue:45663|issue:45759", + "issue:43425|issue:43957", + "issue:43404|issue:43454", + "issue:42947|issue:45704", + "issue:44291|issue:44756", + "issue:43873|issue:44387", + "issue:44079|issue:46829", + "issue:45663|issue:45859", + "issue:44292|issue:44912", + "issue:36010|issue:39404", + "issue:44387|issue:44756", + "issue:44155|issue:45823", + "issue:42947|issue:43381", + "issue:43957|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:45588|issue:45663", + "issue:42796|issue:46326", + "issue:45684|issue:45698", + "issue:43854|issue:45216", + "issue:45292|issue:45335", + "issue:44402|issue:44451", + "issue:43646|issue:43708", + "issue:43645|issue:44164" + ], + "prepared_review_unit_hash": "e8f12c843f84a80ec57d5bdcda13b119039503eb2971a3b668964bceb8ed857a", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but there are two small duplicate-like subclusters: the resize_token_embeddings reports (45276/45292/45335) and the untied lm_head reports (43404/43454). I\u2019d use 45292 as the canonical issue because it is the most generic description in the strongest duplicate set.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 45292 is the most generic and central report in the clearest duplicate subcluster, covering the shared resize_token_embeddings propagation problem across model variants.", + "canonical_pr_reason": null, + "best_issue_reason": "Best representative of the cluster\u2019s strongest duplicate family: it states the shared symptom broadly, while the nearby issues are model-specific variants of the same resize_token_embeddings bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Related BERT loading complaints, but one is a tokenizer vocab-size mismatch and the other is a model-specific load failure; not enough to call the same bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Different topics entirely: SigLIP2 doc example errors vs a GLM5 model issue." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Unrelated models and failure modes (Qwen3 device-map crash vs AudioFlamingo3 batched inference token leak)." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Different code paths and symptoms: FSDP/shared_kv_states KeyError vs AutoModelForCausalLM text_config unwrapping." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "One is a Torch-version incompatibility, the other is a meta-device loading regression; not the same bug." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": true, + "reason": "Both report untied lm_head weights causing bad generation in related multimodal generation models; same underlying weight-tying bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA behavior is unrelated to the apex.FusedRMSNorm memory leak." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected init_empty_weights argument and Strix Halo mmap OOM are separate loading/runtime problems." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both mention quantization, but one is offloading behavior and the other is increased reserved memory/OOM; different root causes." + }, + { + "left": "issue:44079", + "right": "issue:46829", + "accept": false, + "reason": "ModelOutput key assignment bug is unrelated to VideoPrism output field mapping." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "FSDP2 shared_kv_states KeyError and CPU offload incompatibility in Qwen3_5Moe are different issues." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and symptoms: NVFP4 load error vs MXFP4 fallback to bf16." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Whisper return_language pipeline regression." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Quantization OOM and disabling mmap on Strix Halo are distinct problems." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 track leakage and Gemma4 device mismatch are not the same code-path problem." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode incompatibility." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading breakage is unrelated to the Strix Halo mmap OOM report." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and Gemma4 FSDP2 shared_kv_states are unrelated model bugs." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "Nemotron checkpoint loading failure and an official example being broken are too vague to treat as the same bug." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same resize_token_embeddings bug: 45276 is a Gemma4-specific manifestation, while 45292 states the generic output_embeddings failure." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression and strict config loading are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction is unrelated to RLHF NaN explosions from 3D position_ids." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sinkless-model crash is unrelated to the Gemma4 FSDP2 attention KeyError." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer BPE-marker handling and mistral_common continue_final_message are different tokenizer bugs." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Custom-model file permissions and loading the wrong custom module after save_pretrained are separate save/load issues." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading and Qwen3.5 save_pretrained regression are different model-specific problems." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Both describe the same resize_token_embeddings omission; 45335 is the T5Gemma-specific instance of the generic issue in 45292." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both concern model loading, but the tokenizer vocab-size mismatch and the ScandiBERT load failure are not clearly the same root cause." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization regression and Trainer resume_from_checkpoint max_steps calculation are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Jupyter custom-model initialization and save/from_pretrained extra_state handling are different bugs despite both touching model persistence/loading." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary stays conservative about the cluster being mostly heterogeneous and only calls out the two clearly supported duplicate families. The canonical choice of 45292 is reasonable because it is the most generic title in the resize_token_embeddings subcluster. The accepted soft-edge pairs (43404/43454, 45276/45292, 45292/45335) are defensible from the packet titles, and the rejected pairs do not overreach." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 9, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:43299", + "issue:43425", + "issue:43452", + "issue:43596", + "issue:43611", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "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:46897", + "issue:46899", + "issue:46901" + ], + "soft_pairs": [ + "issue:46894|issue:46901", + "issue:43828|issue:43957", + "issue:44589|issue:44756", + "issue:46315|issue:46491", + "issue:45812|issue:46315", + "issue:43425|issue:43828", + "issue:39404|issue:45701", + "issue:44530|issue:45084", + "issue:43611|issue:43645", + "issue:39404|issue:46665", + "issue:43425|issue:46869", + "issue:44683|issue:45823", + "issue:43425|issue:43716", + "issue:43425|issue:45823", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:43906|issue:44530", + "issue:43901|issue:44509", + "issue:43994|issue:44079", + "issue:36032|issue:43452", + "issue:44265|issue:44898", + "issue:39404|issue:44912", + "issue:46050|issue:46326", + "issue:44164|issue:45759", + "issue:43906|issue:46326", + "issue:44671|issue:44991", + "issue:45823|issue:46899", + "issue:44991|issue:45081", + "issue:43299|issue:43931", + "issue:45479|issue:46894", + "issue:43856|issue:44756", + "issue:44164|issue:46821", + "issue:45479|issue:46897", + "issue:43596|issue:44683" + ], + "prepared_review_unit_hash": "4946c1c6c06a3d52f69de3b146df984a7c71c9cc86be7098cdd67b92cb01d3f3", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38135, + "estimated_input_tokens": 9534, + "estimated_eval_tokens": 19324 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a collection of distinct Transformers v5 regression reports spanning tokenization, model loading, dtype/device mismatches, and quantization. The soft pairs are largely topical similarities rather than the same concrete bug.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43611 is the best representative because it is a broad, concrete v5 model-loading regression and sits near the center of the cluster\u2019s recurring loading/config failures.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the clearest anchor for triage: it is specific enough to be actionable, but broad enough to represent the dominant loading-regression theme in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46894", + "right": "issue:46901", + "accept": false, + "reason": "Same high-level pattern (training-loss double shift), but these are different model reports with no evidence of the same underlying code path." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Both mention dtype mismatches, but one is an autocast forward issue and the other is a meta-device loading issue for different models." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Unrelated failures: missing Float8 storage vs a Strix Halo mmap OOM workaround." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "One is a performance regression in convert_tokens_to_ids; the other is a round-trip tokenization correctness bug." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer wrong-ID regression and tokenizer performance regression are different bugs and fixes." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility is not the same bug as a model-specific dtype mismatch under autocast." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline return_language breakage and generic tokenization-change reports are not the same underlying issue." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different code paths and symptoms: PagedAttentionCache group-type crash vs template-node compilation error." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both involve loading/custom model behavior, but one is base_model_prefix handling and the other is notebook-defined custom model initialization; not the same bug." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline behavior and from_pretrained inside Python classes are unrelated loading/pipeline issues." + }, + { + "left": "issue:43425", + "right": "issue:46869", + "accept": false, + "reason": "Torch incompatibility and YOLO multi-GPU device-map failure are different problems." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to Gemma4 device mismatch during forward." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "No shared bug: torch version compatibility vs Mistral-3 image preprocessor/model dtype mismatch." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Different failures with different triggers and fix scopes." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "A video-input regression and a mask-shape IndexError are not the same code-path problem." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Completely different failure modes and subsystems." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of an older issue is not the same as a Qwen3.5 cache/group-type crash." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but they cover different pipeline/API removals and are not duplicate documentation bugs." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Incorrect Siglip2 outputs and ModelOutput key assignment are unrelated issues." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer special-token conflict and gguf loading failures are distinct loading bugs." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export failures with torch_compilable_check and Perceiver resolution failures are different problems." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline regression and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized save/load breakage and mistral_common continue_final_message behavior do not share the same underlying bug." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Qwen composite config unwrapping are separate loading issues." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "No concrete overlap beyond both involving Transformers APIs." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT MLM regression and tokenizer load failure for a different model are not duplicates." + }, + { + "left": "issue:45823", + "right": "issue:46899", + "accept": false, + "reason": "Both involve Gemma4, but one is a device-map forward mismatch and the other is an unsafe dtype cast in multimodal embedders." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizers, different failure points, different fixes." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Same model family, but one is a general loading breakage and the other is a specific weight-shape mismatch; not enough to treat as the same bug." + }, + { + "left": "issue:45479", + "right": "issue:46894", + "accept": false, + "reason": "A classification-loss degeneration and an ASR training-loss shift bug are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Training memory inefficiency and mmap OOM avoidance are different issues." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "extra_state save/load failure and a TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:45479", + "right": "issue:46897", + "accept": false, + "reason": "Different model families and different loss bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed zero3/BertModel index error and compiled flex_attention on torch>=2.9 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": 25, + "review_unit_index": 10, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43262", + "issue:43317", + "issue:43377", + "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:43883", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44462", + "issue:44488", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44811", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45507", + "issue:45584", + "issue:45859", + "issue:46178", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489" + ], + "soft_pairs": [ + "issue:44488|issue:44987", + "issue:44488|issue:45081", + "issue:43611|issue:43644", + "issue:44462|issue:46489", + "issue:42796|issue:43404", + "issue:36032|issue:42915", + "issue:46285|issue:46402", + "issue:39404|issue:44671", + "issue:43716|issue:45084", + "issue:44737|issue:45859", + "issue:43643|issue:44530", + "issue:44661|issue:44683", + "issue:40990|issue:43421", + "issue:42915|issue:44530", + "issue:44451|issue:44987", + "issue:34567|issue:40444", + "issue:43644|issue:43646", + "issue:43541|issue:45084", + "issue:40444|issue:46326", + "issue:43582|issue:43632", + "issue:42175|issue:46178", + "issue:44206|issue:44220", + "issue:36032|issue:43645", + "issue:44811|issue:45584", + "issue:43452|issue:43482", + "issue:44987|issue:44991", + "issue:43883|issue:44488", + "issue:45084|issue:45507", + "issue:43317|issue:43482", + "issue:43262|issue:43377" + ], + "prepared_review_unit_hash": "3a7f7a8258fbef0a2f905cf57de4d3b2c4b26ebc6ce99565177ba9ef0fb7d9f3", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "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": "These items are mostly unrelated closed issues. The soft-similarity pairs share broad themes like tokenizer loading, Whisper, or Transformers v5 regressions, but they do not look like the same underlying bug and should not be merged as duplicates.", + "confidence": 0.87, + "canonical_issue_reason": "issue:44488 is the most central tokenizer-loading report in this set: it is broad, well-discussed, has inbound references, and sits closest to several other tokenizer-related reports by symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44488 is the best representative issue for this cluster because it is the broadest and most connected loading/tokenizer regression among the candidates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are tokenizer/model loading failures, but they affect different repos and likely different code paths." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "One is a generic tokenizer loading issue; the other is a Mistral regex patch crash with a different failure mode." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Both mention tokenizer loading/output on specific models, but the symptoms and likely fixes differ." + }, + { + "left": "issue:42796", + "right": "issue:43404", + "accept": false, + "reason": "Tokenizer BPE marker handling and Mistral lm_head tying are unrelated bugs." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer init conflict vs FineGrainedFP8Config failure; different subsystems and root causes." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "Both are model-structure issues, but CLIP structure and Sam3 weight loading are separate problems." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language behavior and CamemBERT masked-LM predictions are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor dtype mismatch vs template compilation error are different bugs." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both are device/offload related, but they hit different model components and failure modes." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss and PagedAttention linear_attention crash are unrelated." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES handling and flex_attention compilation are different code paths." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression and runtime post-processor refresh are different tokenizer issues." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Both involve Qwen variants, but one is FP8 config loading and the other is a PagedAttention cache crash." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are model-loading failures, but they target different repositories and likely different causes." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both touch Transformers 5 initialization behavior, but one is about non-persistent buffers and the other custom model init." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch Dynamo grouped_mm tracing and non-template-node compilation are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "VL iterable-dataset batching and continue_final_message are different multimodal tokenizer flows." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "AppleSilicon allocator warmup and the _is_hf_initialized flag are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Both are backend/dependency issues, but they concern different backends and installation paths." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "FeatureExtractor center-arg handling and _torch_extract_fbank_features are different audio preprocessing bugs." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and notebook custom initialization are unrelated." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens and empty-transcription generation are different issues." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "GGUF loading via AutoTokenizer/AutoModel and Qwen2.5-GGUF v5 loading are distinct failures." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are v5+ tokenizer/loading regressions, but they affect different repositories and likely need different fixes." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Molmo tied-weights attribute error and tokenizer repository loading are unrelated." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error and GraniteMoEHybrid invalid method are not the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Dequantized model offload failure and GGUF loading failure are distinct problems." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask batching are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 11, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:41720", + "issue:42915", + "issue:43089", + "issue:43322", + "issue:43377", + "issue:43404", + "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:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44849", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45987", + "issue:46315", + "issue:46326", + "issue:46661", + "issue:46821", + "issue:46869", + "issue:46899" + ], + "soft_pairs": [ + "issue:43746|issue:46326", + "issue:43646|issue:44534", + "issue:43645|issue:46661", + "issue:44403|issue:45081", + "issue:42915|issue:45084", + "issue:43792|issue:46869", + "issue:44849|issue:45325", + "issue:43377|issue:43819", + "issue:43404|issue:44466", + "issue:44464|issue:45987", + "issue:28530|issue:34567", + "issue:44987|issue:45081", + "issue:43994|issue:44610", + "issue:43883|issue:44991", + "issue:44008|issue:46899", + "issue:42915|issue:44291", + "issue:44451|issue:45081", + "issue:36032|issue:43742", + "issue:44062|issue:44987", + "issue:44265|issue:44610", + "issue:43089|issue:45987", + "issue:43322|issue:43540", + "issue:44479|issue:45325", + "issue:43632|issue:46821", + "issue:42915|issue:44589", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:44991|issue:46315", + "issue:45216|issue:45357" + ], + "prepared_review_unit_hash": "a695015a51ad6377c8b28ef4d878cf12b598b5efa0ad27c23836702140300518", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36615, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The packet supports one clear duplicate-like pair: issue:45216 and issue:45357, both about Qwen3.5 `save_pretrained` producing incorrect checkpoint/visual-encoder keys. The other soft pairs are related at a broad subsystem level but describe different defects, so they should remain rejected.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45216 is the best canonical issue because it is a concrete, specific regression report with a clear failure mode and version window, and 45357 reads like a later follow-up on the same Qwen3.5 `save_pretrained` bug.", + "canonical_pr_reason": null, + "best_issue_reason": "45216 is the strongest representative issue: it names the exact API, affected model family, and regression behavior, and it has the only soft edge that looks like a true duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "Different failures: PEFT/local checkpoint loading vs mistral_common tokenizer/chat-template behavior." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but one is custom model init and the other is non-persistent buffer corruption; not the same bug." + }, + { + "left": "issue:43645", + "right": "issue:46661", + "accept": false, + "reason": "Notebook custom-model initialization and remote-code `config_class` handling are separate code paths and symptoms." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Logging/noise during loading is unrelated to the Mistral regex patch crash." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 failure and template compilation TypeError are different defects." + }, + { + "left": "issue:43792", + "right": "issue:46869", + "accept": false, + "reason": "Whisper loading failure and YOLO multi-GPU device-map failure are different models and failure modes." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen-family issues, but one is hidden-state output behavior and the other is rope/position-id scaling." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI batching/padding-mask mismatch and DAC STE mismatch are unrelated model-specific logic bugs." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both involve tied weights, but one is untied weights in a model class and the other is device-dependent serialization." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is not the same as `get_seq_length()` returning the wrong type." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early stopping metric selection and `num_input_tokens_seen` tracking are unrelated trainer issues." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "A specific model load failure and a tokenizer regex patch crash are different problems." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Siglip2 inference/pipeline failure and OmDet-Turbo processor size mismatch are separate issues." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Molmo tied-weights attribute error and tokenizer load failure for est-roberta are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:46899", + "accept": false, + "reason": "Gemma3n variable collision and Gemma4 quantized multimodal dtype-cast failure are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and `init_empty_weights` `_is_hf_initialized` handling are distinct initialization issues." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "ScandiBERT loading failure and Mistral regex patch crash are unrelated tokenizer/model loading bugs." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer special-token conflict and MobileLLM key error are different tokenizer/model issues." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken construction error and physical-intelligence/fast loading failure are not the same bug." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export exportability via `torch_compilable_check` is unrelated to processor input-size mismatch." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead and cache-length return-type correctness are different issues." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next load segfault and Qwen3OmniMoe video-processing ValueError are separate multimodal failures." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Video-input regression across Qwen models is not the same as still-image temporal position scaling in Qwen2.5-VL." + }, + { + "left": "issue:43632", + "right": "issue:46821", + "accept": false, + "reason": "`_is_hf_initialized` breakage and TimesFM `window_size` AttributeError are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and granite_speech config loading are different code paths." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer load failure and `convert_tokens_to_ids` slowdown are unrelated." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both describe the same Qwen3.5 `save_pretrained` regression producing incorrect checkpoint/visual-encoder keys; 45357 is a later follow-up on the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Reject: the response overstates the evidence in the summary. The packet only clearly supports one duplicate-like pair, issue:45216 \u2194 issue:45357. The claim that there are \"two soft matches\" that look like true duplicates is not grounded, because the other v5/custom-model pair is explicitly marked non-duplicate in the soft verdicts and the titles point to different defects. The rest of the pairwise rejections are conservative enough." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 12, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:40444", + "issue:40990", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43381", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43646", + "issue:43716", + "issue:43746", + "issue:43792", + "issue:43856", + "issue:44079", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44991", + "issue:45081", + "issue:45325", + "issue:45491", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:46491", + "issue:46710", + "issue:46899" + ], + "soft_pairs": [ + "issue:40444|issue:45325", + "issue:43611|issue:44534", + "issue:43425|issue:44403", + "issue:43716|issue:46899", + "issue:43296|issue:43531", + "issue:44220|issue:44683", + "issue:45812|issue:46491", + "issue:42796|issue:43746", + "issue:43381|issue:45923", + "issue:44403|issue:44991", + "issue:43632|issue:44756", + "issue:33666|issue:40444", + "issue:40990|issue:44811", + "issue:44291|issue:44589", + "issue:43792|issue:44589", + "issue:42796|issue:46710", + "issue:44079|issue:45663", + "issue:35707|issue:40444", + "issue:43482|issue:43525", + "issue:43632|issue:44534", + "issue:43531|issue:44291", + "issue:43421|issue:44568", + "issue:43482|issue:45538", + "issue:42175|issue:45542", + "issue:41720|issue:45081", + "issue:44079|issue:44534", + "issue:43317|issue:43856", + "issue:44462|issue:45812", + "issue:41553|issue:43582", + "issue:43377|issue:45491", + "issue:43611|issue:43646" + ], + "prepared_review_unit_hash": "95b4d6ab86e3a94a107f98b9d4f813fdbb8faad1f048019034aaf2a3cd07d1a6", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36940, + "estimated_input_tokens": 9235, + "estimated_eval_tokens": 18726 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose set of unrelated issues, with a smaller subtheme around Transformers v5 loading/tokenizer regressions. I did not find any true duplicate pairs; the soft links are mostly superficial similarity.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43611 is the broadest and most central representative of the v5-loading regression subgroup, but the cluster overall does not contain a clear single duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "43611 is the best umbrella issue for the cluster\u2019s main theme because it describes a general Transformers v5 model-loading breakage rather than a narrow, model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL training/position-id issues, but one is about iterable datasets with multiple images and the other is about still-image temporal scaling; different bugs." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is about loading models with base_model_prefix and the other about non-persistent buffer initialization; distinct code paths." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "One is a hard Torch 2.10 compatibility problem, the other is noisy loading output; not the same underlying bug." + }, + { + "left": "issue:43716", + "right": "issue:46899", + "accept": false, + "reason": "Both involve multimodal Gemma/Mistral-style processing and dtype issues, but one is an image preprocessor/model dtype mismatch and the other is a 4-bit quantization cast bug in multimodal embedders." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different model families and failure modes: PaddleOCR-VL loading in vLLM versus Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction bug vs compiled flex_attention failure on newer Torch; no shared code-path evidence." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but one is a silent wrong-ID issue for Granite models and the other is a leading-space round-trip bug in CodeLlama; different tokenizer logic." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "Tokenizer BPE-marker handling and PEFT adapter checkpoint loading are unrelated bugs." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing in eval mode is not the same as a missing supports_gradient_checkpointing flag, even if both touch checkpointing APIs." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "One is general loading noise, the other is tokenizer.json being ignored; not the same failure." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "Both are loading/runtime regressions, but one concerns _is_hf_initialized and the other mmap OOM on Strix Halo; different problems." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both mention Qwen2-VL/Qwen2.5-VL fine-tuning, but one is multi-GPU training and the other is an iterable dataset/multi-image failure." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity evaluation issue vs Whisper batch_decode skip_special_tokens handling; unrelated." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Both are Transformers v5-related instantiation failures, but one is about an unexpected _is_hf_initialized argument and the other about missing Float8 storage support; not mergeable as one fix." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper model loading failure and Float8 storage lookup error are different code paths and symptoms." + }, + { + "left": "issue:42796", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer marker handling for Ministral vs accuracy regression on DeepSeek after upgrade; no evidence of the same bug." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug and Gemma4 FSDP2 shared_kv_states issue are unrelated." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Both are generation/training-related Qwen issues only at a very broad level; one is inputs_embeds/past_key_values generation, the other is multi-image finetuning." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and missing pad_token_id on Llama4Config are separate model/config issues." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but the failing mechanisms are different: initialization flag handling versus non-persistent buffer filling." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and _is_hf_initialized constructor failure are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token post-processing updates and add_special_tokens BOS/EOS insertion are different tokenizer behaviors." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and CLIPTokenizer model_max_length behavior are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Backend installation packaging issue vs missing TensorFlow backend detection; not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Auto device mapping cudaErrorAssert and Mistral regex patch crashing are different errors." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key handling and v5 buffer initialization are distinct framework regressions." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "device_map/offload loading failure and Qwen3 MoE memory inefficiency are related only loosely by model placement/memory, not the same bug." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Ignoring tokenizer.json in AutoTokenizer and wrong token IDs for Granite are different tokenizer-loading problems." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and Apple Silicon TypeError in caching_allocator_warmup are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "MIMI encoder padding-mask batching bug and Gemma3 NaN embeddings from sliding-window all-padding windows are both batching issues but not the same code path." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are broad Transformers v5 breakages, but one is about loading models with base_model_prefix and the other about custom model initialization; separate fixes." + } + ] + }, + "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": 25, + "review_unit_index": 13, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43493", + "issue:43550", + "issue:43606", + "issue:43644", + "issue:43701", + "issue:43746", + "issue:43818", + "issue:43819", + "issue:43828", + "issue:43873", + "issue:43994", + "issue:44079", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44509", + "issue:44610", + "issue:44829", + "issue:44877", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45701", + "issue:45823", + "issue:45834", + "issue:46402", + "issue:46821", + "issue:46909" + ], + "soft_pairs": [ + "issue:43296|issue:46821", + "issue:44509|issue:46909", + "issue:43644|issue:44079", + "issue:42175|issue:43296", + "issue:44829|issue:45479", + "issue:43818|issue:46402", + "issue:43818|issue:45357", + "issue:44265|issue:45072", + "issue:43493|issue:43994", + "issue:40444|issue:45381", + "issue:45325|issue:45834", + "issue:45325|issue:45381", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43873|issue:45701", + "issue:44291|issue:44877", + "issue:43701|issue:45325", + "issue:42796|issue:45381", + "issue:40990|issue:45584", + "issue:43606|issue:44265", + "issue:43262|issue:43746", + "issue:44479|issue:45823", + "issue:35707|issue:41720", + "issue:34634|issue:35707", + "issue:43828|issue:46821", + "issue:39692|issue:40444", + "issue:44186|issue:44265", + "issue:43454|issue:43550", + "issue:43606|issue:43819", + "issue:41720|issue:45823", + "issue:43377|issue:44610" + ], + "prepared_review_unit_hash": "c62d8bfaca12a80034e825255b51ebc907ce37dc0ff6cbbe3fc44e2357e8c485", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "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 unrelated issues; the only clear duplicate-looking pair is the SigLIP2 report pair, with 43493 as the better canonical report.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43493 is the more root-cause-oriented SigLIP2 report, describing the HF vs JAX implementation discrepancy directly.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43493 best represents the duplicate pair because it states the underlying model mismatch more generally, while 43994 reads like a downstream symptom in AutoModel/pipeline.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:46821", + "accept": false, + "reason": "Audio processor chat-template sample-rate defaulting is unrelated to a TimesFM window_size AttributeError." + }, + { + "left": "issue:44509", + "right": "issue:46909", + "accept": false, + "reason": "Both are documentation-related, but they concern different docs problems and different fixes." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer junk in v5 is a different bug from ModelOutput key assignment handling." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "TensorFlow backend packaging and audio processor sampling-rate behavior are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both affect training, but one is about flash_attention_3 degeneration and the other about single-label classification loss logic." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "A missing temporal-attention video tower is unrelated to Sam3Model text encoder weight loading." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different multimodal model families and different failures: Video-LLaVA architecture vs Qwen3.5 save_pretrained key handling." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export incompatibility with torch_compilable_check is unrelated to bf16 dtype mismatches in inference." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both describe the same SigLIP2 implementation mismatch causing incorrect/nonsensical outputs compared with the reference behavior." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "IterableDataset multi-image finetuning failure is not the same as Qwen2.5-VL video position-id regression." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL rope-index/position-id scaling and Kosmos2.5 OCR index errors are different bugs." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen2.5-VL position IDs, but one is still-image temporal scaling and the other is video-input position-ID correctness; not clearly the same defect." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Tied lm_head weight serialization and weight tying causing garbage generation are related only loosely, not the same bug report." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to Qwen2.5-GGUF loading with transformers v5." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization/offloading behavior and version-dependent tokenization are different issues." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument in init_empty_weights is not the same as strict config rejection for granite_speech." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to Qwen2.5-VL position-id scaling." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE marker handling is unrelated to Qwen2.5-VL video vision_position_ids." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on GPT-OSS and Whisper empty-transcription failure are different model-specific bugs." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Bark CPU offload device mismatch is unrelated to torch.export failures on torch_compilable_check models." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio processor sample-rate defaults and GraniteSpeech PEFT checkpoint loading are different code paths." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "A video-input regression in Qwen models is unrelated to Gemma4 PLE device mismatch during forward." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "inputs_embeds/past_key_values generation issues are unrelated to Qwen3 auto device mapping cudaErrorAssert." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset failure is unrelated to progressive generation with inputs_embeds and past_key_values." + }, + { + "left": "issue:43828", + "right": "issue:46821", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE has nothing to do with TimesFM window_size." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 documentation example errors and Qwen2.5-VL multi-image finetuning failure are separate problems." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2Tokenizer padding/truncation crashes are unrelated to torch.export and torch_compilable_check." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "lm_head weight tying/serialization and Bamba torch.compile SDPA failure are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Bark CPU offload device mismatch is unrelated to DAC.from_latents vs forward-pass mismatch." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Qwen3 cudaErrorAssert under auto device mapping is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask batching and OmDet-Turbo input-size mismatch are different model/processor bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: most pairs are correctly rejected as unrelated, and 43493/43994 is the only pair that plausibly reads as the same SigLIP2 bug with one report describing the implementation discrepancy and the other its downstream symptom. No clear overstatement stood out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 14, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43065", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43572", + "issue:43720", + "issue:43749", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44387", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45357", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315", + "issue:46402", + "issue:46821" + ], + "soft_pairs": [ + "issue:43257|issue:45923", + "issue:44568|issue:45920", + "issue:46305|issue:46315", + "issue:42796|issue:46209", + "issue:43881|issue:44488", + "issue:43454|issue:45072", + "issue:45823|issue:46821", + "issue:40990|issue:42947", + "issue:43377|issue:44898", + "issue:42617|issue:42915", + "issue:43454|issue:43720", + "issue:43749|issue:44534", + "issue:43317|issue:43873", + "issue:43572|issue:45759", + "issue:43322|issue:46050", + "issue:41720|issue:43322", + "issue:36331|issue:42617", + "issue:44387|issue:45704", + "issue:41720|issue:44977", + "issue:36331|issue:42915", + "issue:44466|issue:45357", + "issue:40990|issue:45701", + "issue:43450|issue:45701", + "issue:44877|issue:44991", + "issue:43065|issue:46402", + "issue:26413|issue:28530", + "issue:41720|issue:42617", + "issue:44186|issue:45072", + "issue:43526|issue:43883", + "issue:43873|issue:44756", + "issue:43881|issue:44991" + ], + "prepared_review_unit_hash": "db0497217f628b466c5af3a2defbf85360d53e5bb249e3b00813640bc554a004", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36963, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly unrelated issue reports across different models and subsystems. None of the soft pairs look like the same underlying bug with sufficient specificity to merge as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:43873 is the broadest and most central candidate in this set: it is open, has multiple inbound references, and covers a general offloading/quantization failure mode that could subsume nearby reports. That said, the rest of the items are mostly distinct bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest representative issue because it is broader than the other reports, already attracting references, and is phrased as a general regression rather than a single model-specific failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Different bugs: one is MoE weight conversion with accelerate+deepspeed, the other is a missing gradient-checkpointing flag on a trust_remote_code variant." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different code paths and model families; not the same concrete bug." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "One is a sliding-window masking off-by-one bug, the other is a tokenizer lookup performance regression; unrelated failure modes." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE marker handling and FP8 inference on B200 are unrelated issues in different components." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are model loading failures, but for different models and likely different causes; too broad to treat as the same bug." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Tie-weight serialization/generation corruption is unrelated to dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:45823", + "right": "issue:46821", + "accept": false, + "reason": "Device-mapping forward mismatch and a TimesFM window_size AttributeError are different code paths and symptoms." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a model and ineffective gradient checkpointing with LoRA are not the same underlying bug." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Batch-vs-single padding mask inconsistency in MIMI is not the same as Perceiver image classification failure at non-default resolution." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "A script execution failure and a Qwen3 MoE FP8 loading issue are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Different components and behaviors: tied-weight serialization vs packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading breakage and non-persistent buffer junk values are different loading regressions, not the same concrete bug." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both concern offloading/quantization, but one is specific to dequantized model loading with device_map=auto and the other is a broader offloading complaint; insufficient evidence they are the same bug." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is unrelated to failing to unwrap text_config for composite Qwen multimodal configs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Segfault loading Llava Next models and sinq save/load corruption are different failure modes." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "CUDA assert on Qwen3 with auto device mapping and a Llava Next load segfault are unrelated and model-specific." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss API breakage is unrelated to not being able to run a parallelism example script." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Int4 quantization memory regression and an apex RMSNorm memory leak are different memory issues with different causes." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Auto device mapping CUDA assert and flash-attention generation issues in Qwen3.5 are distinct bugs." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Unexpected compute_loss keyword argument and Qwen3 MoE FP8 failure are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve save_pretrained regressions, but one is lm_head serialization/tied weights and the other is incorrect visual encoder keys; different code paths." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "A perplexity regression does not match a broad tokenization-change report; too weak to call duplicates." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug and tokenization changes are unrelated subsystems." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config rejection for granite_speech and tokenizer loading failure for est-roberta are unrelated issues." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3Model text encoder weights not loading are separate model components and bugs." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "resume_from_checkpoint epoch_iterator failure and early-stopping metric lookup are different Trainer issues." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Auto device mapping CUDA assert and an example script execution failure do not describe the same bug." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crashes on NER/batched padding are unrelated to dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and missing all_tied_weights_keys on Molmo are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both are memory/offload themed, but one is a quantization/offloading regression and the other is a Strix Halo mmap OOM workaround; not the same concrete bug." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "A model loading failure and a tokenizer loading failure for a different model 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": 25, + "review_unit_index": 15, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28530", + "issue:30064", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:44008", + "issue:44186", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44610", + "issue:44671", + "issue:44734", + "issue:44743", + "issue:44898", + "issue:44913", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:46291", + "issue:46612", + "issue:46665", + "issue:46821", + "issue:46858", + "issue:46869" + ], + "soft_pairs": [ + "issue:46612|issue:46858", + "issue:43550|issue:43720", + "issue:43577|issue:43708", + "issue:43295|issue:45701", + "issue:45685|issue:46821", + "issue:44186|issue:44610", + "issue:43646|issue:46665", + "issue:43550|issue:43606", + "issue:43404|issue:45357", + "issue:36010|issue:42617", + "issue:44610|issue:44898", + "issue:42915|issue:43317", + "issue:44734|issue:46858", + "issue:44423|issue:45406", + "issue:46821|issue:46869", + "issue:44008|issue:45823", + "issue:41720|issue:45161", + "issue:28530|issue:33666", + "issue:44464|issue:46858", + "issue:45632|issue:45684", + "issue:44671|issue:44743", + "issue:43531|issue:43572", + "issue:42915|issue:46291", + "issue:36331|issue:41720", + "issue:44186|issue:44361", + "issue:43425|issue:46821", + "issue:43316|issue:44913", + "issue:42915|issue:43322", + "issue:28530|issue:30064" + ], + "prepared_review_unit_hash": "18fc38277d63ede365838fd87239f305c254bd9f420318b64d2950c063cb541f", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35916, + "estimated_input_tokens": 8979, + "estimated_eval_tokens": 18214 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the issues span unrelated training, tokenizer/processor, model-loading, device-map, serving, and cache bugs. None of the proposed soft pairs look like the same underlying defect, so no soft edges are accepted. There are no PRs in the cluster.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45701 is the best representative because it is the most broadly phrased and has the highest external reference signal, making it the closest thing to a hub in an otherwise mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the best issue candidate for canonicalization, but only as a broad anchor; the cluster does not actually contain a tight duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46612", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache/generation, but one is a model-family beam-search reorder omission and the other is a static-cache seq-length arithmetic bug on non-CUDA backends." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Different code paths: Bamba torch.compile + SDPA versus BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Blip2 dtype propagation and Trainer resume/max_steps calculation are unrelated bugs." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "One is a regression around processor.tokenizer/images passthrough; the other is a broad tokenization-version discrepancy. Related theme, not the same bug." + }, + { + "left": "issue:45685", + "right": "issue:46821", + "accept": false, + "reason": "Completely different models and failures: MoE MPS histogram support versus TimesFM window_size AttributeError." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Both are processor/tokenizer bugs, but they affect different models and different failure modes: NER/batched padding versus image size mismatch." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Custom model initialization breakage in v5 is not the same as inconsistent from_pretrained behavior inside Python classes." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Different model-specific device/offload failures; no shared concrete code path." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Weight tying in Mistral3 conditional generation is unrelated to Qwen3.5 visual encoder key saving." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "Importing GenerationMixin and running 3d_parallel.py are unrelated issues." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision input-shape problems, but they involve different models and different mechanisms: processor resize mismatch versus interpolate_pos_encoding handling." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and dequantized model offload loading failure are different model-loading bugs." + }, + { + "left": "issue:44734", + "right": "issue:46858", + "accept": false, + "reason": "Both touch cache logic, but one is serve-side continuation tensor indexing and the other is prefill static-cache length arithmetic; not the same bug." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve/multimodal-related, but the actual failures differ: string-to-device conversion versus missing _tokenizer on Gemma4Processor." + }, + { + "left": "issue:46821", + "right": "issue:46869", + "accept": false, + "reason": "TimesFM window_size AttributeError and YOLO multi-GPU device_map failure are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n forward variable collision is unrelated to Gemma4 PLE device mismatch under device_map=auto." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Different models and symptoms: Qwen3 auto device mapping cudaErrorAssert versus GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:28530", + "right": "issue:33666", + "accept": false, + "reason": "Early stopping/eval metric selection and Qwen2-VL multi-GPU training are not the same defect." + }, + { + "left": "issue:44464", + "right": "issue:46858", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is a different bug from static-cache seq-length arithmetic on non-CUDA backends." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision for trust_remote_code is unrelated to permission propagation in save_pretrained/register_for_auto_class." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are separate model-specific bugs." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and missing pad_token_idx in StableLmConfig are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to kernels package version/revision import requirements." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer compute_loss signature change and Qwen3 device-map cudaErrorAssert are different layers of the stack." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both tokenizer-related, but the models and concrete errors differ; they are not the same underlying bug." + }, + { + "left": "issue:43425", + "right": "issue:46821", + "accept": false, + "reason": "Torch 2.10 compatibility and TimesFM window_size AttributeError do not share a concrete cause." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both concern config API behavior, but for different configs and unrelated fields." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FP8 loading and Llava Next segmentation fault are distinct model-loading failures." + }, + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early stopping metric resolution and void segmentation map processing 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": 25, + "review_unit_index": 16, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:43329", + "issue:43388", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43644", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:44062", + "issue:44077", + "issue:44387", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44589", + "issue:44625", + "issue:44898", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:46402", + "issue:46665", + "issue:46821", + "issue:46858" + ], + "soft_pairs": [ + "issue:33357|issue:35707", + "issue:43540|issue:43782", + "issue:29159|issue:36032", + "issue:44387|issue:45701", + "issue:43873|issue:45704", + "issue:30064|issue:34634", + "issue:43526|issue:43881", + "issue:43550|issue:43653", + "issue:44534|issue:46665", + "issue:43531|issue:45759", + "issue:43606|issue:43720", + "issue:41720|issue:43746", + "issue:45406|issue:45464", + "issue:44492|issue:44509", + "issue:42490|issue:43329", + "issue:43388|issue:45059", + "issue:43606|issue:43653", + "issue:34567|issue:35707", + "issue:34634|issue:41720", + "issue:43531|issue:46821", + "issue:43540|issue:46858", + "issue:45759|issue:45859", + "issue:44451|issue:44991", + "issue:43531|issue:44589", + "issue:45357|issue:46402", + "issue:44077|issue:44625", + "issue:43644|issue:46665", + "issue:44062|issue:44451", + "issue:44898|issue:45072" + ], + "prepared_review_unit_hash": "5fb21a95aca3ffd9915d2ddf24c2ef7beb374675a9d8f9a4f2ff171a48246fb1", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35883, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose similarity set of unrelated issue reports, with only a few near-duplicate regressions around tokenizer/model-loading/v5 behavior. None of the soft pairs look like the exact same underlying bug, so all are rejected.", + "confidence": 0.63, + "canonical_issue_reason": "Issue 44534 is the clearest broad regression report in the set and is the best representative of the closest duplicate-like theme here (the non-persistent-buffer v5 bug).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44534 is the most suitable anchor for the cluster because its title is broad, specific, and representative of a concrete regression rather than a model- or task-specific edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "Different bugs: MacOS bus error in a CLIP model versus generation behavior with inputs_embeds/past_key_values." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Both involve Qwen multimodal models, but one is a video-processing ValueError and the other is a from_pretrained weight_only loading error." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Tokenizer-related, but not the same bug: slow/fast inconsistency versus a T5Tokenizer load-time AddedToken conflict." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "Unrelated symptoms: int4 quantization OOM versus tokenizer behavior changes across versions." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Different concrete problems: quantization/offloading behavior versus an apex RMSNorm memory leak." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Both are processor bugs, but they affect different processors and failure modes." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated: BeitImageProcessorFast label reduction bug versus a GLM-4V model loading failure." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different code paths: torch.compile/SDPA failure versus tokenizer special-token registration." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Both mention from_pretrained/v5-era behavior, but one is a buffer-initialization regression and the other is inconsistent behavior inside Python wrapper classes." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both are Qwen-family issues, but one is a sliding_window bug and the other is config unwrapping for composite multimodal models." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Accelerate/loading-related, but one is CPU offload device mismatch and the other is packed-weight unpacking for BitNet." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Completely different: Qwen3 device-mapping CUDA assert versus GraniteSpeech local checkpoint loading with PEFT adapters." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Both hit serving/API paths, but the reported failures are different: missing _tokenizer versus streaming inference failure." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache strategy docs is unrelated to stale pipeline-task documentation." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Different issues: local path trailing-slash import warning versus a video multimodal token-counting bug." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different subsystems: gather_for_metrics label truncation versus SAM3 text/box prompting behavior." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "No shared underlying bug; device mismatch during offload is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counting is unrelated to progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "Different model families and failures: BarkProcessor voice preset versus Qwen3 auto device mapping CUDA assert." + }, + { + "left": "issue:43531", + "right": "issue:46821", + "accept": false, + "reason": "Both involve model-specific arguments, but sliding_window on Qwen3-MoE is unrelated to TimesFM window_size AttributeError." + }, + { + "left": "issue:43540", + "right": "issue:46858", + "accept": false, + "reason": "Different code paths: Qwen3OmniMoe video input processing versus cache seq_length tensor arithmetic in _prefill." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both are Qwen 3.5-related, but one is config unwrapping and the other is a CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model-loading failures, but they are different model-specific regressions and not the same concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window handling for Qwen3-MoE is unrelated to a Float8 storage lookup error." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Both concern multimodal model save/load, but they affect different models and different directions (saving visual keys vs loading text encoder weights)." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Different issues: an optional post_init policy problem versus Qwen3.5 num_labels propagation." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent-buffer junk values and inconsistent from_pretrained behavior inside Python classes are not the same bug." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "A tokenizer AddedToken argument conflict is not the same as ScandiBERT loading failure." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different model behaviors: Perceiver image classification with interpolation versus bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + } + ] + }, + "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": 25, + "review_unit_index": 17, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43493", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44186", + "issue:44368", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45507", + "issue:45704", + "issue:45759", + "issue:45907", + "issue:45923", + "issue:45987", + "issue:46032", + "issue:46050", + "issue:46463", + "issue:46655", + "issue:46693", + "issue:46829", + "issue:46832", + "issue:46858" + ], + "soft_pairs": [ + "issue:44987|issue:45042", + "issue:44625|issue:45479", + "issue:43381|issue:43844", + "issue:39692|issue:44610", + "issue:44368|issue:45357", + "issue:41720|issue:45704", + "issue:43746|issue:46050", + "issue:43381|issue:45704", + "issue:46655|issue:46693", + "issue:43421|issue:43927", + "issue:45704|issue:45923", + "issue:44918|issue:44936", + "issue:42886|issue:44843", + "issue:45759|issue:45923", + "issue:43746|issue:43873", + "issue:44514|issue:44545", + "issue:43994|issue:44671", + "issue:46032|issue:46463", + "issue:39692|issue:43493", + "issue:43421|issue:43653", + "issue:43782|issue:45381", + "issue:30064|issue:36010", + "issue:45081|issue:45356", + "issue:43421|issue:44186", + "issue:45507|issue:46858", + "issue:45356|issue:45907", + "issue:39692|issue:43873", + "issue:43720|issue:43819", + "issue:46829|issue:46832", + "issue:43653|issue:43720", + "issue:45987|issue:46829" + ], + "prepared_review_unit_hash": "6ec937c3f12a790d13aeb8df5eb0f6393853a04860f83867a577eb24cc0cd530", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37509, + "estimated_input_tokens": 9378, + "estimated_eval_tokens": 19012 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly unrelated transformer/model bugs that should not be merged; only the Qwen2_5_VLProcessor batched `padding=False` reports are true duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "issue:44545 is an exact duplicate of issue:44514: same `Qwen2_5_VLProcessor.apply_chat_template` crash on batched input with `padding=False`, and it has slightly more discussion/reference activity.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44545 is the best representative issue because it is the clearest same-bug duplicate pair in the set and captures the concrete code-path failure precisely.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Different bugs: one is loading `physical-intelligence/fast`, the other is PIL image processors incorrectly requiring `torchvision`." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both concern classification configs, but one is `num_labels` propagation in Qwen3.5 and the other is a generic zero-loss bug for `num_labels=1`." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Different failure modes: gradient checkpointing in eval mode vs gradient explosion with HfDeepSpeedConfig + ZeRO-3." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 docs/example errors are unrelated to OmDet-Turbo's processor/model image-size mismatch." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Unrelated: a `tie_word_embeddings` warning during LoRA fine-tuning vs incorrect visual encoder keys saved for Qwen3.5." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Different code paths: Qwen3 auto device mapping on A800 vs T5 silently using apex RMSNorm with a memory leak." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "One is PEFT adapters/local checkpoint loading for GraniteSpeech; the other is quantized model save/load corruption with sinq." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Eval-mode gradient checkpointing and T5 apex RMSNorm leak are unrelated bugs." + }, + { + "left": "issue:46655", + "right": "issue:46693", + "accept": false, + "reason": "SmolVLM video padding calculation bug is unrelated to a FA2 long-context performance regression." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both touch tokenizer metadata, but one is runtime post-processor updates and the other is DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "T5 RMSNorm memory leak and missing `supports_gradient_checkpointing` on Nemotron are distinct issues." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: Qwen3.5 embedding unpacking with TRL SFT trainer vs `trainer.evaluate()` after `trainer.train()`." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both involve offline/tokenizer behavior, but one is cache loading offline and the other is an unconditional `model_info()` call in `_patch_mistral_regex`." + }, + { + "left": "issue:45759", + "right": "issue:45923", + "accept": false, + "reason": "Different problems: unwrapping `text_config` in composite Qwen configs vs a missing gradient-checkpointing capability flag." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT/local checkpoint loading is unrelated to quantization offloading behavior." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact duplicate reports of the same `Qwen2_5_VLProcessor.apply_chat_template` batched `padding=False` crash." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and CamemBERT masked-LM prediction regressions are different models and bugs." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "Mamba2 `use_cache` correctness issue vs T5 decoder input validation error message are not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both mention SigLIP2, but one is a documentation example error and the other is an HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Different tokenizer issues: post-processor not updating at runtime vs BigBirdTokenizer mask token not registered as special." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3VL `weight_only=True` loading error and Qwen2.5-VL video `vision_position_ids` bug are separate multimodal issues." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Completely different areas: void segmentation maps in image processing vs an ImportError for `GenerationMixin`." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Mistral regex patch crash is unrelated to the Kimi-K2.5 tokenizer codec/regression report." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime special-token post-processor updates are unrelated to LayoutLMv2Tokenizer crashes on NER/batched padding." + }, + { + "left": "issue:45507", + "right": "issue:46858", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and static-cache sequence-length tensor arithmetic are different code paths." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Kimi tokenizer regression is unrelated to `list_repo_files` failing to forward `revision`." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 example/model mismatch is unrelated to quantization offloading behavior." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet packed-weights loading and DAC `from_latents`/STE mismatch are different model bugs." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight unpacking are unrelated." + }, + { + "left": "issue:46829", + "right": "issue:46832", + "accept": false, + "reason": "Different models and failures: VideoPrism hidden-state output bug vs RT-DETR feature-level mismatch." + }, + { + "left": "issue:45987", + "right": "issue:46829", + "accept": false, + "reason": "StaticCache sequence-length return type issue is unrelated to VideoPrism output wiring." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate pair is issue:44514/issue:44545, and the other soft-edge pairs are appropriately rejected as distinct bugs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 18, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43335", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43611", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43756", + "issue:43819", + "issue:43927", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44361", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44661", + "issue:44671", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45834", + "issue:45987", + "issue:46294", + "issue:46665", + "issue:46693", + "issue:46858", + "issue:46932" + ], + "soft_pairs": [ + "issue:43295|issue:45834", + "issue:44062|issue:44488", + "issue:43550|issue:43927", + "issue:43550|issue:44361", + "issue:43540|issue:46294", + "issue:45381|issue:46693", + "issue:43531|issue:43643", + "issue:43596|issue:44220", + "issue:43611|issue:46665", + "issue:44186|issue:44898", + "issue:44008|issue:46858", + "issue:43653|issue:43819", + "issue:41720|issue:44991", + "issue:45987|issue:46932", + "issue:30064|issue:36331", + "issue:43643|issue:44877", + "issue:44220|issue:44661", + "issue:43618|issue:43697", + "issue:43335|issue:43756", + "issue:43550|issue:43819", + "issue:44843|issue:45081", + "issue:39692|issue:45834", + "issue:44448|issue:44671", + "issue:44821|issue:45084", + "issue:44186|issue:44442", + "issue:43540|issue:44008", + "issue:34634|issue:36010", + "issue:45183|issue:45479", + "issue:45584|issue:45834" + ], + "prepared_review_unit_hash": "98435aeed41754a8c2b11f084f341b5e73986bafb7c498e474ced00e13944b6c", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36647, + "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 mostly false-positive similarity across unrelated Transformers bugs. None of the soft pairs look like the same underlying defect, so all soft edges are rejected. Issue 43550 is the most central item, but it is not a safe duplicate umbrella.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43550 is the most central node among the candidates (highest soft-pair connectivity) and is a concrete runtime/model failure report, so it is the best available issue representative for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "It is the strongest representative only in a graph-centrality sense; the cluster itself is too heterogeneous for a true duplicate canonical, but 43550 is the clearest, most specific issue to anchor on.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Different code paths: processor/tokenizer regression vs long OCR input indexing error." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AddedToken constructor bug and the other is a tokenizer loading failure for a specific model." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Different models and failures: Bamba compile error vs MLuke tokenizer AttributeError." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both involve video, but one is Qwen3OmniMoe input processing and the other is a chat-template feature/token count mismatch in Qwen3.5." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "Video position ID bug in Qwen2.5-VL is not the same as the FA2 long-context performance regression." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to trust_remote_code AutoConfig field loss." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "ZeRO3/BertModel indexing failure and audio fbank feature extraction are unrelated subsystems." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Model loading with base_model_prefix is a different bug from inconsistent from_pretrained behavior inside Python classes." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer crashes on NER padding/truncation are unrelated to Perceiver image classification at non-default resolution." + }, + { + "left": "issue:44008", + "right": "issue:46858", + "accept": false, + "reason": "Gemma3n variable collision in forward() is unrelated to StaticCache sequence-length tensor semantics." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DAC.from_latents missing STE are different bugs." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:45987", + "right": "issue:46932", + "accept": false, + "reason": "StaticCache get_seq_length() returning a tensor is unrelated to Gemma4 quantized embedder dtype casting." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Void segmentation-map processing and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config loading, but missing fields from trust_remote_code and strict granite_speech config rejection are distinct failures." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like failing in TOKENIZER_MAPPING_NAMES are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIP attentions assignment regression and RTDetrV2 output drift are different model behaviors." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Sparse-layer creation at zero layers and Smollm3 RoPE-layer dropping are separate configuration bugs." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba compile/SDPA failure and DAC forward/from_latents mismatch are unrelated model issues." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; not the same defect." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc-example errors are unrelated to Kosmos2.5 long-OCR indexing." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Different models and symptoms: Pegasus output drift vs CamemBERT masked-LM prediction errors." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation errors have no common code path." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Qwen3OmniMoe video input handling and Gemma3n audio-mask attribute collision are different model paths." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset and GenerationMixin import errors are unrelated issues." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Validation XOR error messaging and degenerate zero-loss classification behavior are different bugs." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation and Kosmos2.5 long OCR index errors 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": 25, + "review_unit_index": 19, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36331", + "issue:43322", + "issue:43388", + "issue:43421", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43825", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44008", + "issue:44291", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45381", + "issue:45685", + "issue:45701", + "issue:45834", + "issue:46032", + "issue:46291", + "issue:46294", + "issue:46463", + "issue:46512", + "issue:46665", + "issue:46846" + ], + "soft_pairs": [ + "issue:43644|issue:43749", + "issue:43322|issue:45685", + "issue:43606|issue:45072", + "issue:43388|issue:46032", + "issue:43825|issue:43994", + "issue:44479|issue:45834", + "issue:43550|issue:45072", + "issue:43782|issue:44977", + "issue:43421|issue:44811", + "issue:44561|issue:45020", + "issue:43577|issue:43818", + "issue:43582|issue:46512", + "issue:46294|issue:46463", + "issue:44991|issue:46291", + "issue:43716|issue:44008", + "issue:43931|issue:46846", + "issue:44805|issue:44936", + "issue:43540|issue:45381", + "issue:44610|issue:45072", + "issue:43749|issue:44291", + "issue:43720|issue:45072", + "issue:43825|issue:43827", + "issue:28218|issue:29159", + "issue:43645|issue:46665", + "issue:44671|issue:45701", + "issue:43577|issue:43582", + "issue:34634|issue:36331" + ], + "prepared_review_unit_hash": "e187f4702b6bdd1c4976ce6689395492397001692e4b2e6dd494914e03eb9b5e", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35448, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a set of unrelated issue reports. All soft edges should be rejected; the tokenizer pair is related in theme but not supported strongly enough to treat as a duplicate from titles alone.", + "confidence": 0.66, + "canonical_issue_reason": "Issue #28218 is the most concrete and narrowly scoped report in the set, so it is the best issue anchor even though it should not be merged as a duplicate with #29159 on the available evidence.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #28218 is the clearest, most specific bug report here. It is more precise than the broader tokenizer inconsistency report in #29159, so it is the best representative issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Similar loading/runtime wording, but they describe different failures and different code paths: junk non-persistent buffers vs FSDP CPU RAM-efficient loading." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Different symptoms and subsystems: Llava Next loading segfault vs an MPS histogram op type error in MoE code." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both mention device/dtype mismatches, but they affect different models and different failure modes; not the same concrete bug." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "One is a metrics-label truncation bug, the other is a use_cache correctness issue in Mamba2Mixer; no shared underlying defect." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Pipeline error-message wording and a SigLIP2 bad-output bug are distinct; same API surface is not enough to merge them." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Both involve multimodal/video/OCR inputs, but the models and failures are different; no evidence of the same root cause." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "torch.compile+SDPA failure in Bamba is unrelated to bfloat16 dtype mismatches in other models." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "A from_pretrained weight_only loading error is not the same as a flash-attention generation bug." + }, + { + "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 separate tokenizer/decoding issues." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both reference remote_code regressions, but one is specifically removal of is_torch_fx_available and the other is a broad recent-version breakage report." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype persistence and Video-LLaVA architecture/weight-sharing issues are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Apple Silicon TypeError in caching_allocator_warmup and stale model.config.dtype after from_config are different bugs." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen3.5 video-chat template token/feature mismatch is unrelated to T5 decoder input validation messaging." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading failure for a specific model and kernels package import-version requirements are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "A dtype mismatch between image preprocessor and model is not the same as a Gemma3n variable-name collision AttributeError." + }, + { + "left": "issue:43931", + "right": "issue:46846", + "accept": false, + "reason": "Weight shape mismatch during load and missing TP-plan entries causing OOM/shape errors are different issues, even if both involve Qwen models." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask/tensor shape mismatch and trainer.evaluate failing after train do not share a concrete underlying bug." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs in Qwen-family models, but they target different model variants and different mistakes (processing error vs position-id regression)." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "OmDet-Turbo processor image-size mismatch is unrelated to bfloat16 dtype mismatch failures." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading being broken is not the same as init_empty_weights rejecting an unexpected argument." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading and dtype mismatches are distinct loading issues." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "One is a runtime error-message bug, the other is stale docs referencing removed pipeline() APIs; related topic, but not the same issue." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "The title overlap is not enough: #28218 is a specific added-token spacing bug, while #29159 is a broader slow-vs-fast tokenizer inconsistency report without stronger body-level evidence of the same root cause." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Both involve model initialization patterns, but the notebook custom-model breakage and from_pretrained-inside-classes inconsistency are not clearly the same bug." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM prediction regressions and a general tokenization-change report are too broad and not concrete duplicates." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype persistence and Apple Silicon warmup TypeError are unrelated failures." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The analysis is mostly grounded, but it overstates the evidence for the only accepted soft edge. In the packet, `issue:28218` vs `issue:29159` is explicitly marked `deterministic_accept: false`, and the titles alone do not firmly establish a duplicate relationship\u2014`#28218` is a specific added-token spacing bug, while `#29159` is a broader slow-vs-fast tokenizer inconsistency report. Please keep that verdict conservative and reject it unless there is stronger body-level evidence. The rest of the soft-edge 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": 25, + "review_unit_index": 20, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:42175", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43582", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43697", + "issue:43742", + "issue:43782", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44554", + "issue:44589", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44991", + "issue:45003", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45538", + "issue:45600", + "issue:46175", + "issue:46209", + "issue:46285", + "issue:46291", + "issue:46326", + "issue:46697", + "issue:46728", + "issue:46932" + ], + "soft_pairs": [ + "issue:42175|issue:45081", + "issue:46175|issue:46932", + "issue:46697|issue:46728", + "issue:45003|issue:46285", + "issue:43782|issue:45325", + "issue:43653|issue:44912", + "issue:43262|issue:43819", + "issue:45372|issue:46326", + "issue:43582|issue:44857", + "issue:45081|issue:45406", + "issue:43742|issue:45020", + "issue:44589|issue:45042", + "issue:43550|issue:44912", + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:43526|issue:43697", + "issue:44060|issue:44368", + "issue:36010|issue:43645", + "issue:36010|issue:36331", + "issue:43906|issue:46209", + "issue:43329|issue:45600", + "issue:43526|issue:44554", + "issue:45084|issue:45538", + "issue:43819|issue:43927", + "issue:43638|issue:44936", + "issue:43531|issue:45042", + "issue:42175|issue:46291", + "issue:42490|issue:43257" + ], + "prepared_review_unit_hash": "3c88361a6022902c4dc2d687e6ed4aec616defc60c62b664197c473b54d752cc", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 36030, + "estimated_input_tokens": 9008, + "estimated_eval_tokens": 18272 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft pair looks like the same underlying bug/change; the matches are mostly broad semantic similarity across unrelated regressions. I\u2019d keep the issues separate and use a broad, clearly-scoped issue as the canonical representative.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45020 has the clearest broad regression title and scope, making it the best representative issue in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the strongest standalone issue for tracking because it states a concrete, user-facing breakage with a clear subsystem and version regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Different bugs: backend install/dependency selection vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:46175", + "right": "issue:46932", + "accept": false, + "reason": "Both Gemma4-related, but one is a huge position-embedding materialization issue and the other is dtype casting for quantized embedders." + }, + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Different image processors and symptoms: over-padding vs wrong patch count." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Remote-code/sys.modules safety issue vs CLIP model structure inconsistency." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different model families and failures: Qwen3-VL load error vs Qwen2.5-VL rope indexing bug." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration bug vs quantization fallback/load problem." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sample-rate default vs DAC forward/from_latents mismatch." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both involve mistral_common, but one is an import dependency break and the other is a continue_final_message behavior bug." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon TypeError vs CUDA AMP crash; unrelated execution paths." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Different missing-attribute failures in different code paths: backend_tokenizer vs _tokenizer." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Specific KeyError loading one model vs broad remote_code regression; not the same concrete defect." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Storage-object lookup failure vs PIL/torchvision dependency requirement." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA bug vs MXFP4 quantization load fallback." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Different loading regressions: gguf_file handling across auto classes vs a specific tokenizer load failure." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Accelerate/deepspeed weight conversion bug vs multimodal token counting/video branch undefined names." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Beit label reduction bug vs RTDetrV2 output drift in v5; different model behaviors." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Related warning theme, but one is a model weight-tying bug and the other is a user-facing fine-tuning warning." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Import path regression vs notebook custom-model initialization breakage." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError from generation module vs custom trainer signature mismatch." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "Repro of an older issue vs FP8 inference failure on B200; unrelated symptoms." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video-branch helper bug vs stale auto_mappings config reference." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Image label reduction bug vs MPS attention correctness issue." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation error vs tokenizer model_max_length sentinel issue." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent reconstruction mismatch vs DiaConfig save/load token-ID loss." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3/non-pretrained Bert IndexError vs train-then-evaluate failure." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding-window bug vs PIL backend torchvision requirement." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Torch backend install issue vs kernels dependency version/revision requirement." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash local model import warning vs Qwen3-MoE weight conversion failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 21, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:29870", + "issue:34567", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43454", + "issue:43540", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43696", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43749", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44387", + "issue:44479", + "issue:44568", + "issue:44610", + "issue:44617", + "issue:44756", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44977", + "issue:45042", + "issue:45356", + "issue:45459", + "issue:45759", + "issue:45803", + "issue:45859", + "issue:46421", + "issue:46424", + "issue:46701" + ], + "soft_pairs": [ + "issue:43746|issue:44811", + "issue:43643|issue:45759", + "issue:43854|issue:44977", + "issue:39692|issue:44912", + "issue:43720|issue:44610", + "issue:38175|issue:44610", + "issue:43257|issue:43696", + "issue:43262|issue:44811", + "issue:43257|issue:43931", + "issue:44387|issue:44617", + "issue:43329|issue:43450", + "issue:45459|issue:45803", + "issue:43329|issue:44479", + "issue:42915|issue:43742", + "issue:34567|issue:40990", + "issue:46421|issue:46424", + "issue:43749|issue:44756", + "issue:44291|issue:45042", + "issue:43650|issue:43906", + "issue:43606|issue:45859", + "issue:42915|issue:44821", + "issue:29159|issue:34567", + "issue:43540|issue:46701", + "issue:43927|issue:44912", + "issue:28282|issue:29870", + "issue:43454|issue:43577", + "issue:44568|issue:45356" + ], + "prepared_review_unit_hash": "6566ff3b06efc6684b946a3930e89afa07ffbb9ddf5f1b9a14ba2d5f80e543c5", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35447, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is still mostly heterogeneous. `issue:43329` is the strongest representative for the video/multimodal area, but the related video items are adjacent more than clearly duplicate; I would not merge any of the soft-edge pairs here.", + "confidence": 0.66, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "`issue:43329` is the clearest representative because it is a concrete bug in the multimodal token-counting/video code path, but the surrounding video/VLM reports are only loosely adjacent rather than definite duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Different bugs: PEFT/local checkpoint loading vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config-loading behavior, but they target different code paths and different failures." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different models and failures; one is a unit-test load issue, the other is a flash-attention generation regression." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 docs/example errors are unrelated to GPT-OSS MXFP4 quantization fallback." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to OmDet-Turbo processor image-size mismatch." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Different model-specific problems: zero probabilities vs wrong processor output resolution." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading is unrelated to a GPT-OSS out-of-memory failure." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Both are processor-related, but one is chat-template sample-rate handling and the other is Whisper decoding flags." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different loading bugs: accelerate/deepspeed weight conversion vs Qwen3-VL shape mismatch." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "One is an int4 memory regression and the other is a model-specific CUDA OOM; roots are not the same." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Shared video theme, but token-counting undefined variables and batched-shape handling are different bugs." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both swallow errors, but they are different exception-handling bugs in different components." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "They are thematically close video/VLM regressions, but the packet does not clearly show the same underlying bug or a single code-path fix." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different model-loading failures: FineGrainedFP8Config vs a key error in MobileLLM." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Trainer token-count tracking and GPT-OSS perplexity are unrelated problems." + }, + { + "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 semantics." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are different loading/memory issues." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights argument handling is unrelated to PIL backend torchvision requirements." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "The first title is not a concrete bug report; the second is an isolated reproduction of another issue." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both mention CPU offload, but they affect different models and different code paths." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3Moe FP8 loading and AutoImageProcessor-from-URL are unrelated." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistencies are unrelated to TrainerState token counting." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Different multimodal pipelines and different failure modes." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig token-ID persistence is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:28282", + "right": "issue:29870", + "accept": false, + "reason": "ImportError for PyTorch and Bart generate() UnboundLocalError are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Both affect multimodal model loading/generation, but the underlying bugs are different." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regressions on different models; same subsystem, not the same bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level summary is mostly conservative, but one soft-edge verdict is too strong: issue:43329 and issue:44479 are similar video/VLM regressions, yet the packet does not clearly show they are the same bug. Also, saying 43329 is the only item with a plausible duplicate partner slightly overstates the evidence, since other video/multimodal items are at least thematically adjacent even if not duplicates." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 22, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43404", + "issue:43425", + "issue:43452", + "issue:43525", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44849", + "issue:44857", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45127", + "issue:45137", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45538", + "issue:45701", + "issue:45907", + "issue:46291", + "issue:46512", + "issue:46693", + "issue:46821", + "issue:46832", + "issue:46869" + ], + "soft_pairs": [ + "issue:43525|issue:46291", + "issue:44987|issue:45538", + "issue:28282|issue:34689", + "issue:44857|issue:46832", + "issue:43697|issue:45357", + "issue:33357|issue:34567", + "issue:43404|issue:45127", + "issue:43425|issue:46291", + "issue:45137|issue:45405", + "issue:45538|issue:46869", + "issue:42175|issue:43824", + "issue:28282|issue:29127", + "issue:42915|issue:45538", + "issue:44479|issue:46693", + "issue:44062|issue:44991", + "issue:43742|issue:44291", + "issue:44936|issue:46821", + "issue:42915|issue:43749", + "issue:43452|issue:43582", + "issue:43577|issue:46512", + "issue:24643|issue:29159", + "issue:44849|issue:45701", + "issue:27957|issue:28282", + "issue:41720|issue:43540", + "issue:44987|issue:45907", + "issue:43540|issue:44560", + "issue:44479|issue:45381", + "issue:42175|issue:44488", + "issue:34567|issue:36331", + "issue:43540|issue:44479" + ], + "prepared_review_unit_hash": "77cbe9260bc4b344d2e4ad1a6ac21f279fc7854afdb74b1a540745f95a490254", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, with many unrelated model-loading, trainer, tokenizer, and video-input issues. The only clearly plausible duplicate subcluster is the Qwen video regression pair 44479/45381; the rest should stay separate.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 44479 is the best canonical anchor because it is the broadest regression ticket in the only coherent subcluster and aligns with the more specific video-position-id symptom reported in 45381.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the best overall representative for triage: it names the regression, spans multiple affected Qwen video models, and is the strongest match for a mergeable duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Both are loading-related, but they describe different failures in different subsystems: Llama4 pad_token_id vs kernels revision forwarding." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "One is a model-loading failure for a specific repo; the other is a tokenizer max_length bug. No shared code-path." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Both mention model loading, but one is a missing PyTorch import error and the other is a Llama 3.2 vision-loading regression." + }, + { + "left": "issue:44857", + "right": "issue:46832", + "accept": false, + "reason": "Different RT-DETR bugs: AMP loss crash versus forward shape/feature-level mismatch." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different issues in different paths: output differences in RTDetrV2 vs incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Completely unrelated: MacOS bus error with CLIP vs TrainerState token counting." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both involve weight tying, but they are different model families and different failure modes; not the same concrete bug." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch compatibility and kernels revision handling are unrelated issues." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error and PEFT version bump are unrelated." + }, + { + "left": "issue:45538", + "right": "issue:46869", + "accept": false, + "reason": "Tokenizer max_length behavior and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging/import problem vs missing Qwen2.5-VL symbol; different causes." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Generic missing-PyTorch import error vs LayoutLMv3 box-validation messaging; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FP8/Qwen3Moe failure and CLIPTokenizer max_length are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Video-input regression and FA2 long-context performance regression are different problems." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken keyword collision vs tokenizer loading failure for a specific model repo; no shared code-path." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Both are model-loading errors, but one is a key lookup issue and the other is an init_empty_weights argument mismatch." + }, + { + "left": "issue:44936", + "right": "issue:46821", + "accept": false, + "reason": "Trainer lifecycle bug vs TimesFM window_size attribute error; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and FSDP CPU RAM efficient loading are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading API break and Apple Silicon warmup TypeError are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both concern dtype handling, but one is a Blip2 loading issue and the other is a generic from_config config-staleness bug; not the same change." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed weight-dimension error and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states bug and version-dependent tokenization changes are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to missing PyTorch import handling." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Both are Qwen-family runtime failures, but they are different code paths and symptoms: auto device mapping vs video input processing." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Loading a model repo and list_repo_files revision forwarding are different bugs." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both involve video inputs, but they affect different Qwen model families and do not clearly share the same concrete bug." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "These look like the same underlying Qwen video-input regression: 44479 reports the broader v5.3.0 regression and 45381 describes the specific vision_position_ids symptom." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Packaging/backend inclusion issue vs a specific model load failure; unrelated." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token accounting and custom compute_loss signature mismatch are separate trainer API issues." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-input related, but the model families and reported failures differ enough that they do not clearly represent one mergeable bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft edge, 44479/45381, is plausibly the same Qwen video-input regression based on the titles, and the remaining pair verdicts are appropriately cautious. No clear 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": 25, + "review_unit_index": 23, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:24643", + "issue:29870", + "issue:38175", + "issue:43257", + "issue:43452", + "issue:43531", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43716", + "issue:43742", + "issue:43756", + "issue:43906", + "issue:44062", + "issue:44291", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44877", + "issue:44898", + "issue:44964", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45405", + "issue:45488", + "issue:45588", + "issue:45600", + "issue:45663", + "issue:45672", + "issue:45674", + "issue:45823", + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491", + "issue:46693", + "issue:46701", + "issue:46821", + "issue:46829", + "issue:46846" + ], + "soft_pairs": [ + "issue:38175|issue:44898", + "issue:44792|issue:44964", + "issue:44568|issue:45488", + "issue:43257|issue:43756", + "issue:44062|issue:45081", + "issue:44291|issue:44821", + "issue:45042|issue:45081", + "issue:43582|issue:46821", + "issue:43756|issue:46399", + "issue:43742|issue:44589", + "issue:43756|issue:45325", + "issue:43716|issue:46701", + "issue:44568|issue:46491", + "issue:45405|issue:45600", + "issue:44737|issue:46463", + "issue:45672|issue:45674", + "issue:43577|issue:45072", + "issue:43531|issue:44821", + "issue:45823|issue:46463", + "issue:43906|issue:46693", + "issue:43906|issue:46468", + "issue:24643|issue:29870", + "issue:43452|issue:44488", + "issue:44625|issue:44877", + "issue:43540|issue:46693", + "issue:45663|issue:46846", + "issue:43697|issue:46829", + "issue:44737|issue:45588" + ], + "prepared_review_unit_hash": "906e5ecab1fe092900c04bb9fbb65498bac9069dd47f2dda89781d797b663567", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35945, + "estimated_input_tokens": 8987, + "estimated_eval_tokens": 18230 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most pairs share only broad subsystem keywords (tokenizers, multimodal loading, dtype, config, training) but not the same concrete bug or code path, so I would not merge any of the soft edges.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the items span many unrelated bugs, and none of the soft pairs is close enough to serve as a single duplicate anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "No clear best issue: there is no strong representative issue for the cluster because the candidate pairs are not true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model inference bugs, but they involve different models and different failure modes (zero probabilities vs resize/interpolation behavior)." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both are multimodal loading/generation reports, but they target different models and distinct breakages." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both concern tokenizer behavior, but one is missing BOS/EOS insertion and the other is a pre-tokenizer override regression in a different tokenizer path." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "These are unrelated Qwen/SmolLM issues: one is MoE weight conversion under accelerate+deepspeed, the other is RoPE layer selection." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer errors, but they come from different code paths and unrelated root causes." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "These address completely different subsystems: init_empty_weights loading versus AutoImageProcessor URL loading." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image processor torchvision dependency and Mistral regex patching are unrelated bugs." + }, + { + "left": "issue:43582", + "right": "issue:46821", + "accept": false, + "reason": "AppleSilicon allocator warmup and TimesFM window_size handling are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Different models and different failures: RoPE layer dropping versus kernel registration/decorator error." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Both are model-loading errors, but the specific causes and affected paths are different." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both mention RoPE/position handling, but they affect different models and different logic." + }, + { + "left": "issue:43716", + "right": "issue:46701", + "accept": false, + "reason": "Both involve multimodal pipelines, but one is dtype mismatch and the other is a broken image-text-to-text output path." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different round-trip behaviors." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "A PEFT version pin and a removed config reference are unrelated CI/config issues." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "XLNet CPU allocation bug and T5 decoder error-message bug are unrelated." + }, + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Both are 4-bit inference issues, but one is a Gemma4 audio-module uint8/finfo problem and the other is a BitsAndBytesConfig exclusion bug." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype mismatches, but they are different models and different inference paths." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "Gemma4 device_map forward mismatch and T5 validation messaging are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "The first is an isolated reproduction of another issue; the second is a FA2 performance regression, so they do not describe the same bug." + }, + { + "left": "issue:43906", + "right": "issue:46468", + "accept": false, + "reason": "An isolated reproduction of a separate issue is not enough to match an official Nemotron example breakage." + }, + { + "left": "issue:24643", + "right": "issue:29870", + "accept": false, + "reason": "DeepSpeed weight-dimension training error and Bart decoder_attention_mask UnboundLocalError are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file breaks from_pretrained APIs, while the other reports a model-specific loading failure." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "One is Qwen3.5 label propagation, the other is strict config loading for granite_speech; different fixes and code paths." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "Video input processing in Qwen3OmniMoe is unrelated to FA2 long-context performance regression." + }, + { + "left": "issue:45663", + "right": "issue:46846", + "accept": false, + "reason": "Both are parallelism/training issues, but they affect different models and different mechanisms (FSDP2 shared_kv_states vs TP plan entries)." + }, + { + "left": "issue:43697", + "right": "issue:46829", + "accept": false, + "reason": "Different models and different symptoms: output drift versus an output field mis-assignment." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU tensor creation and flash_attention s_aux=None handling are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 24, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28282", + "issue:32090", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43819", + "issue:43994", + "issue:44361", + "issue:44442", + "issue:44466", + "issue:44493", + "issue:44811", + "issue:44843", + "issue:44898", + "issue:44913", + "issue:44936", + "issue:45005", + "issue:45200", + "issue:45216", + "issue:45356", + "issue:45357", + "issue:45372", + "issue:45375", + "issue:45479", + "issue:45542", + "issue:45920", + "issue:46050", + "issue:46121", + "issue:46170", + "issue:46178", + "issue:46463", + "issue:46701", + "issue:46829", + "issue:46832", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:45357|issue:45920", + "issue:46463|issue:46899", + "issue:28282|issue:32090", + "issue:44913|issue:46121", + "issue:43697|issue:45216", + "issue:43526|issue:45479", + "issue:43720|issue:44898", + "issue:43643|issue:45375", + "issue:45200|issue:46463", + "issue:44466|issue:45005", + "issue:41762|issue:44936", + "issue:44493|issue:46050", + "issue:43761|issue:46829", + "issue:43452|issue:44843", + "issue:44361|issue:44442", + "issue:43688|issue:43749", + "issue:43653|issue:44361", + "issue:38175|issue:43295", + "issue:43819|issue:44811", + "issue:43994|issue:44898", + "issue:45372|issue:46170", + "issue:43525|issue:46701", + "issue:46832|issue:46932", + "issue:43531|issue:43742", + "issue:45542|issue:46178", + "issue:43295|issue:45356" + ], + "prepared_review_unit_hash": "740d7d9bc1e53f04cd54ce01bb7e7a7a2b120f25ef02afd8cf16843cf38d699e", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 36113, + "estimated_input_tokens": 9029, + "estimated_eval_tokens": 18314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issues with no single cluster-wide theme. The only plausible duplicate pair is the tied-weights/save_pretrained regression in v5 (#44466 and #45005); the rest are separate bugs across unrelated code paths.", + "confidence": 0.63, + "canonical_issue_reason": "#45005 is the broader umbrella report for the tied-weights regression in v5, and #44466 reads like a specific lm_head/device manifestation of the same save_pretrained problem.", + "canonical_pr_reason": null, + "best_issue_reason": "#45005 is the best representative issue in this set because it captures the broader tied-weights failure mode that can subsume the more specific serialization symptom in #44466.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Different bugs: Qwen3.5 visual-encoder key serialization vs broad AutoTokenizer token-ID regression." + }, + { + "left": "issue:46463", + "right": "issue:46899", + "accept": false, + "reason": "Unrelated failures: T5 decoder-input validation message vs Gemma4 4-bit multimodal dtype casting." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "ImportError for missing PyTorch vs Trainer GPU broadcast NoneType error are unrelated." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both touch RoPE-related config handling, but one is GPTNeoX config reload and the other is a validation helper TypeError; not the same concrete bug." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RT-DETR output differences and Qwen3.5 checkpoint saving are separate regressions." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT image label reduction bug and sequence-classification zero-loss bug are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver interpolation failure are different model/code paths." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Config-field loss from trust_remote_code and Qwen3.5 vision config field dropping are different issues." + }, + { + "left": "issue:45200", + "right": "issue:46463", + "accept": false, + "reason": "Gemma4 mm_token_type_ids defaulting and T5 decoder validation error are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both describe the same v5 tied-weights/save_pretrained regression; #44466 is a specific lm_head/device symptom, while #45005 is the broader tied-weights translation-model report." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading index error and trainer.evaluate() failing after train() are unrelated." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected position-id key warnings and sinq quantized save/load breakage are different bugs." + }, + { + "left": "issue:43761", + "right": "issue:46829", + "accept": false, + "reason": "Both concern output fields, but CLIPVision hidden_states regression and VideoPrism output misassignment are different model implementations and not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf_file loading breakage and mistral regex offline failure are separate AutoTokenizer code paths." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer load failure are unrelated tokenizer bugs." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug and MLukeTokenizer task AttributeError do not share the same underlying cause." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 zero-probabilities output and processor.tokenizer regression are different symptoms and code paths." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC from_latents mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and Perceiver image-classification resolution handling are different bugs." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Gemma4 processor import failure from mistral_common and Windows ROCm torch.distributed dependency problems are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:46701", + "accept": false, + "reason": "Llama4Config pad_token_id missing and DiffusionGemmaGenerationOutput pipeline failure are separate issues." + }, + { + "left": "issue:46832", + "right": "issue:46932", + "accept": false, + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedder dtype casting are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM load key error are different bugs." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "TensorBoard-without-TensorFlow backend error and tensorflow_text backend mapping error are separate optional-dependency issues." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Processor/tokenizer regression in v4.57.5 and Kimi-K2.5 codec handling regression are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet titles, and the only accepted pair (#44466/#45005) is a conservative match on the tied-weights/save_pretrained regression. The other soft-edge rejections are appropriately cautious and do not overclaim shared causality." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 25, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29159", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:36010", + "issue:41720", + "issue:42898", + "issue:43452", + "issue:43504", + "issue:43550", + "issue:43577", + "issue:43611", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43927", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44484", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44743", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45440", + "issue:45491", + "issue:45701", + "issue:45704", + "issue:46002", + "issue:46050", + "issue:46402", + "issue:46491", + "issue:46612", + "issue:46693", + "issue:46821", + "issue:46846" + ], + "soft_pairs": [ + "issue:43653|issue:44568", + "issue:43646|issue:43749", + "issue:29159|issue:33357", + "issue:43688|issue:45704", + "issue:42898|issue:46491", + "issue:44843|issue:44991", + "issue:43452|issue:43577", + "issue:43756|issue:46402", + "issue:43688|issue:45701", + "issue:44464|issue:44743", + "issue:43653|issue:43927", + "issue:44479|issue:44625", + "issue:45491|issue:46002", + "issue:44821|issue:46821", + "issue:44484|issue:44704", + "issue:33666|issue:36010", + "issue:43688|issue:45440", + "issue:32090|issue:33666", + "issue:44466|issue:46050", + "issue:44977|issue:46693", + "issue:44734|issue:46612", + "issue:44464|issue:44977", + "issue:41720|issue:44464", + "issue:43504|issue:43550", + "issue:43504|issue:43720", + "issue:43577|issue:44554", + "issue:43611|issue:46846", + "issue:43782|issue:46693", + "issue:43643|issue:44561" + ], + "prepared_review_unit_hash": "2a34ac50cd34fcc38da09f488353b9ca77f355d008c083f626df99f0e40795e5", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36488, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab bag of distinct regressions across tokenization, loading, generation, and model-specific code paths. The soft pairs only share broad subsystem similarity, not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.42, + "canonical_issue_reason": "issue:45701 is the broadest umbrella here ('transformers version changes the tokenization') and works best as a representative bucket for the tokenizer-regression cluster; the other tokenizer issues are narrower and appear to have different root causes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the strongest global representative because it is generic enough to cover the shared tokenizer-regression theme without overcommitting to one specific model or failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both involve token handling, but one is BigBird special-token registration and the other is mDeBERTa add_special_tokens/BOS-EOS behavior; different model paths and symptoms." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model init under v5 and FSDP CPU-efficient loading are unrelated failure modes with different code paths." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency and a MacOS bus error on a CLIP model are not the same bug." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Auxiliary-loss normalization in MoE models and a T5/apex RMSNorm memory leak are different issues." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "Both concern tokenization behavior changes, but one is clean_up_tokenization_spaces and the other is CodeLlama leading-space round-trip regression; distinct fixes." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline-mode breakage from an unconditional model_info call is a different root cause than EMBEDDIA/est-roberta tokenizer loading failure." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading breakage and Blip2 dtype not propagating are separate model-loading bugs." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and Sam3Model text-encoder weights not loading are unrelated model-specific defects." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "A MoE auxiliary-loss normalization bug is not the same as a generic tokenization-version regression." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Compiled chunked generation inconsistency and recurrent-state reset with cache are both generation bugs, but they hit different mechanisms and code paths." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird special-token registration and DiaConfig save/load token-id loss are different token/config persistence problems." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "A video-input regression in Qwen VL models and missing num_labels propagation in Qwen3.5 are separate bugs." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on GPU and the other is classification-head init/checkpoint NaNs; different root causes." + }, + { + "left": "issue:44821", + "right": "issue:46821", + "accept": false, + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size behavior and AutoProcessor.cached_file kwargs passing are separate serialization/loading concerns." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are completely different issues." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE auxiliary-loss normalization and DeepseekV3MoE divergence from remote implementation are not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer gpu_broadcast NoneType failure is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Device-dependent lm_head serialization and sinq quantization save/load breakage are different persistence bugs." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Flash-attention generation correctness for Qwen3.5 and long-context FA2 performance regression are different failure modes." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "KV-cache continuation indexing in transformers serve and cache-reorder being skipped for legacy models are separate generation-cache bugs." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked compiled generation inconsistency is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto device-mapping cudaErrorAssert and chunked generation inconsistency are different problems." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field loading and Bamba torch.compile+SDPA failure are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Beit preset loading and BitNet packed-weight unpacking during accelerate loading are different code paths." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation and an MPS attention correctness issue are not the same bug." + }, + { + "left": "issue:43611", + "right": "issue:46846", + "accept": false, + "reason": "base_model_prefix loading breakage and Qwen3.5 tensor-parallel plan omissions are separate v5 regressions." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "weight_only=True loading failure and FA2 long-context performance regression are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code and removal of is_torch_fx_available are both v5 trust_remote_code regressions, but they are different root causes and not one concrete 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": 25, + "review_unit_index": 26, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:39692", + "issue:42796", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43688", + "issue:43701", + "issue:43720", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44361", + "issue:44451", + "issue:44466", + "issue:44488", + "issue:44514", + "issue:44554", + "issue:44568", + "issue:44610", + "issue:44625", + "issue:44792", + "issue:44843", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45290", + "issue:45405", + "issue:45584", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46365", + "issue:46489", + "issue:46710" + ], + "soft_pairs": [ + "issue:33666|issue:36331", + "issue:43278|issue:43688", + "issue:43262|issue:43295", + "issue:43720|issue:43927", + "issue:43504|issue:44610", + "issue:43606|issue:43927", + "issue:43493|issue:43825", + "issue:44514|issue:45290", + "issue:43720|issue:44361", + "issue:44964|issue:45405", + "issue:44554|issue:46365", + "issue:44625|issue:45584", + "issue:43525|issue:43572", + "issue:44977|issue:46209", + "issue:43452|issue:44451", + "issue:45823|issue:46050", + "issue:28218|issue:33357", + "issue:43994|issue:46192", + "issue:42796|issue:44977", + "issue:43493|issue:43844", + "issue:39692|issue:43295", + "issue:43278|issue:43844", + "issue:44568|issue:46489", + "issue:43646|issue:44466", + "issue:46209|issue:46710", + "issue:33357|issue:33666", + "issue:43278|issue:43701", + "issue:44568|issue:45812", + "issue:44488|issue:44843", + "issue:44849|issue:46209", + "issue:44792|issue:45405", + "issue:43927|issue:44361" + ], + "prepared_review_unit_hash": "ad3b3b9a0012997e1016af03b2e91ed7078dcee5ea00705065b74f502bd3c291", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37420, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly false-positive similarity: the issues span unrelated tokenizer, config, loading, training, and hardware-specific bugs. No soft pair looks like the same concrete underlying change. Issue 45812 is the best representative because it is the broadest tokenizer regression in the set.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45812 is the most central and broadly scoped bug here: a silent v4\u2192v5 tokenizer regression affecting all Granite models. It best represents the tokenizer-compatibility subset of this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the strongest standalone tracker because it describes a clear, user-facing, high-blast-radius tokenizer bug and is broader than the narrower model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Unrelated problems: Qwen2-VL multi-GPU training versus a CustomTrainer compute_loss signature mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Different bugs: embedding dtype drift in evaluate versus auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio processor sampling-rate default issue versus a multimodal processor.tokenizer regression; different code paths." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weights loading failure and DiaConfig custom token ID persistence are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT pretrained preset loading bug versus OmDet-Turbo image-size mismatch; different models and failure modes." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to DiaConfig token-ID loss on save/load." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 model-vs-JAX discrepancy is not the same as a pipeline error message about translation support." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both are apply_chat_template-related, but one is batched padding=False crash and the other is assistant tool-call/no-content handling." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Packed-weight loading in BitNet and MLukeTokenizer task AttributeError are different bugs." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal loading failure is unrelated to a bumped MIN_PEFT_VERSION dependency issue." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and Trainer import failure due to datasets are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription handling are different model-specific bugs." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both involve missing pad token fields, but they affect different configs and fields in different models." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug and FP8 inference on B200 are unrelated hardware/back-end issues." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file loading regression is not the same as failing to load a specific pretrained model." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 device-map mismatch and sinq save/load quantization break are different problem classes." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing after added token is unrelated to a MacOS bus error loading a CLIP model." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Granite token IDs regression and return_tensors being ignored when text_kwargs is passed are different API bugs." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Ministral BPE-marker handling and Qwen3.5 flash-attention generation are unrelated tokenizer vs attention issues." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is not the same as gradient explosion under HfDeepSpeedConfig + ZeRO-3." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 documentation/example errors and processor.tokenizer access regression are different bugs." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype change in evaluation is unrelated to abnormal gradient growth during ZeRO-3 training." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Different tokenizer regressions: missing BOS/EOS for mdeberta-v3-base versus incorrect DeepSeek-Coder v1 output." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage and lm_head.weight serialization inconsistency are separate v5 issues." + }, + { + "left": "issue:46209", + "right": "issue:46710", + "accept": false, + "reason": "FP8 inference failure on B200 and DeepSeek-R1-Distill-Llama accuracy regression are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error in CLIP loading is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype drift and resume_from_checkpoint key mismatch do not share an underlying bug." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different models and mechanisms: special-token addition versus Granite token ID corruption." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model loading failure for sleng-bert is unrelated to an unconditional model_info() call breaking offline mode." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 output_hidden_states bug and FP8 inference on B200 are different issues." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A janus test failure is not the same as the MIN_PEFT_VERSION dependency bump." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig custom token IDs are a serialization/config issue, not the same as MLukeTokenizer task AttributeError." + } + ] + }, + "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": 25, + "review_unit_index": 27, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:42490", + "issue:43257", + "issue:43299", + "issue:43334", + "issue:43504", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43653", + "issue:43696", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43937", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44464", + "issue:44493", + "issue:44589", + "issue:44849", + "issue:44869", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45200", + "issue:45229", + "issue:45245", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46285", + "issue:46399", + "issue:46424", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46821", + "issue:46832", + "issue:46846" + ], + "soft_pairs": [ + "issue:43575|issue:45405", + "issue:44849|issue:45357", + "issue:43824|issue:44589", + "issue:43504|issue:43606", + "issue:46821|issue:46832", + "issue:43257|issue:43299", + "issue:45245|issue:46710", + "issue:44493|issue:46285", + "issue:43334|issue:45759", + "issue:46399|issue:46846", + "issue:43334|issue:43572", + "issue:43606|issue:44112", + "issue:44589|issue:45520", + "issue:43572|issue:44291", + "issue:43504|issue:43653", + "issue:45072|issue:46832", + "issue:44464|issue:46424", + "issue:43937|issue:46701", + "issue:43844|issue:44964", + "issue:43873|issue:46050", + "issue:43994|issue:44493", + "issue:45229|issue:45823", + "issue:45200|issue:45229", + "issue:44869|issue:46752", + "issue:45588|issue:46050", + "issue:42490|issue:43696", + "issue:29159|issue:32090", + "issue:29942|issue:30333", + "issue:44977|issue:46701" + ], + "prepared_review_unit_hash": "6d55b4561c121930326fbf6ee6597c35e9e658a0e1f3cbb53d23eab7c01c847d", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36481, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose thematic grouping of unrelated Transformers issues: model-loading regressions, config/version mismatches, device/offload problems, and a few CI/runtime failures. The soft pairs are generally similar by subsystem or model family, but not the same underlying bug/change.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue clearly subsumes the cluster; 43299 is the closest representative regression-style report, but the items are not true duplicates of one bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43299 is the most representative high-signal report in the set: a clear Transformers dev regression affecting model loading, with a concrete failure mode and broad enough framing to anchor the cluster better than the more niche reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Different problems: TP loading OOM vs an unreleased PEFT version bump." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Same model family, but different bugs and code paths: hidden-states handling vs save_pretrained key mapping." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated failures: missing Qwen2.5-VL import vs a Float8 storage lookup/type error." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Different models and failure modes: BEiT legacy-field loading vs bark-small CPU offload device mismatch." + }, + { + "left": "issue:46821", + "right": "issue:46832", + "accept": false, + "reason": "Different architectures and bugs: TimesFM window_size vs RT-DETR feature-level mismatch." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE loading, but the triggers differ materially: accelerate+deepspeed conversion vs dev-model loading regression." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "Completely different issues: category-cardinality limit vs accuracy regression after upgrade." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Different structural regressions: unexpected position-id key vs CLIP text-model class structure change." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both are Qwen multimodal loading issues, but one is a missing config attribute and the other is a factory/unwrapping bug." + }, + { + "left": "issue:46399", + "right": "issue:46846", + "accept": false, + "reason": "Related model family, but different concrete fixes: kernel registration decorator vs TP plan coverage for linear_attn." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different models and different missing config fields." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Bark device-mismatch bug vs a GraniteSpeech CI test issue; not the same underlying defect." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different runtime failures: Float8 storage missing vs flash_attn import-utils KeyError on Python 3.13." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Both are version-related breakages, but one is a missing config field and the other is an init_empty_weights signature mismatch." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Different components and symptoms: BEiT preset loading vs BigBirdTokenizer special-token registration." + }, + { + "left": "issue:45072", + "right": "issue:46832", + "accept": false, + "reason": "Different bug classes: dtype mismatch in inference vs RT-DETR backbone feature-count mismatch." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Both touch generation/cache behavior, but one is compiled-forward inconsistency and the other is cache-length config handling." + }, + { + "left": "issue:43937", + "right": "issue:46701", + "accept": false, + "reason": "Unrelated failures: invalid GenerationConfig vs a broken image-text-to-text pipeline." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "Different models and root causes: ZeRO-3 gradient anomaly vs latest-transformers model loading failure." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Both involve quantization/offloading, but one is offload behavior and the other is save/load corruption after sinq quantization." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Different bugs: Siglip2 pipeline/AutoModel behavior vs unexpected position-id key handling." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but one is OOM from scale and the other is a device mismatch during forward." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only fine-tuning defaults vs multi-GPU inference OOM are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46752", + "accept": false, + "reason": "Different decoding/input-edge-case bugs in unrelated components." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention sink-less-model crash vs quantized model save/load breakage are separate bugs." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Import warning for trailing-slash paths is unrelated to a GPT-oss CUDA OOM." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer fast/slow inconsistency is unrelated to the trainer _gpu_broadcast_one NoneType error." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 CI failures and MLflow reporting on failed jobs are different issues." + }, + { + "left": "issue:44977", + "right": "issue:46701", + "accept": false, + "reason": "Different models and code paths: Qwen3.5 flash-attention generation vs DiffusionGemma image-text pipeline failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 28, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29870", + "issue:29942", + "issue:34567", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43295", + "issue:43404", + "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:44554", + "issue:44610", + "issue:44734", + "issue:44987", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45356", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46121", + "issue:46178", + "issue:46192", + "issue:46369", + "issue:46710", + "issue:46821" + ], + "soft_pairs": [ + "issue:45084|issue:45520", + "issue:28218|issue:34567", + "issue:43065|issue:46369", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:44423|issue:44734", + "issue:41720|issue:46710", + "issue:43493|issue:43701", + "issue:45406|issue:46121", + "issue:43404|issue:46710", + "issue:45072|issue:45198", + "issue:43295|issue:43450", + "issue:37428|issue:43957", + "issue:45588|issue:45759", + "issue:43540|issue:45200", + "issue:43295|issue:43873", + "issue:44186|issue:46192", + "issue:45823|issue:46192", + "issue:43295|issue:44610", + "issue:43572|issue:46821", + "issue:43550|issue:43844", + "issue:45200|issue:45823", + "issue:29870|issue:29942", + "issue:43450|issue:43575", + "issue:45520|issue:46178", + "issue:43572|issue:44987", + "issue:37428|issue:43824", + "issue:42490|issue:43493", + "issue:44554|issue:45356", + "issue:39692|issue:46192" + ], + "prepared_review_unit_hash": "27d192f34deb934a7e94bcd819bff452281477b4ad1bfba9388b5e42901fbdd9", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36330, + "estimated_input_tokens": 9083, + "estimated_eval_tokens": 18422 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. A few items are thematically adjacent around multimodal processor/tokenizer behavior, but none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43295 is the broadest representative of the recurring multimodal processor/tokenizer regression theme and sits closest to the few adjacent soft matches.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43295 best represents the main subtheme in this cluster; it is broad, user-facing, and closer to multiple nearby processor-related reports than the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Different failures: template-node compilation vs import-time backend lookup. Same broad area only by coincidence." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing bug and TrainerState counter update are unrelated code paths." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Both mention Sam3, but one is a dummy layer concern and the other is a memory overflow report; not the same bug." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability behavior." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "Both touch attention/flex attention, but one is an MPS correctness issue and the other is a version-check import bug." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve/runtime crashes, but one is multimodal input handling and the other is KV-cache continuation indexing." + }, + { + "left": "issue:41720", + "right": "issue:46710", + "accept": false, + "reason": "A device-mapping CUDA assert on Qwen3 is not the same as an accuracy regression on DeepSeek after upgrade." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor _tokenizer crash and rope-param validation type error are different subsystems." + }, + { + "left": "issue:43404", + "right": "issue:46710", + "accept": false, + "reason": "Untied lm_head weights in Mistral3 is not the same problem as a model accuracy regression." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "dtype mismatch in inference and Wav2Vec2 save/tokenization failure are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both are processor-related, but one is processor.tokenizer/image handling and the other is batched video output shape; different bugs." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for a flash-attention symbol and meta-device loading failures are different issues." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Both involve multimodal/serve or composite configs, but the crash conditions and code paths differ." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video input processing in Qwen3OmniMoe is unrelated to Gemma4 mm_token_type_ids defaults." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression and quantization offloading behavior are not the same bug." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2Tokenizer crashes on NER/padding are unrelated to return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device_map forward mismatch and return_tensors handling are different failure modes." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Both are preprocessing/model-shape mismatches, but they affect different models and input pipelines." + }, + { + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "StableLmConfig pad_token_idx and TimesFM window_size AttributeError are unrelated config bugs." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile/SDPA failure and DeepSpeed gradient overflow are different training/runtime issues." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Same model family (Gemma4) but distinct bugs: token-type defaults vs device placement." + }, + { + "left": "issue:29870", + "right": "issue:29942", + "accept": false, + "reason": "Bart decoder_attention_mask assignment and Flash Attention 2 test failures are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video processor shape bug and tensor-parallel OOM on Qwen2-57B are unrelated." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both are import/backend mapping issues, but they involve different backend registries and different symptoms." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and loading physical-intelligence/fast on 5.1.0 are not the same bug." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different missing imports: flash-attention helper vs Qwen2_5_VLForConditionalGeneration." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Trailing-slash import warning and SigLIP2 implementation discrepancy are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness and tokenizer codec regression are different problems." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 documentation/model-processor mismatch and return_tensors handling are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 29, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:41720", + "issue:42915", + "issue:43232", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43335", + "issue:43450", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43550", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43856", + "issue:43901", + "issue:43950", + "issue:43975", + "issue:44112", + "issue:44442", + "issue:44466", + "issue:44479", + "issue:44521", + "issue:44737", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45325", + "issue:45381", + "issue:45412", + "issue:45563", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46512", + "issue:46894", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:43232|issue:44849", + "issue:43295|issue:43493", + "issue:44521|issue:45290", + "issue:44466|issue:45127", + "issue:45381|issue:46209", + "issue:43278|issue:43856", + "issue:44737|issue:46894", + "issue:28218|issue:35141", + "issue:43653|issue:44112", + "issue:42915|issue:43824", + "issue:43824|issue:43975", + "issue:43950|issue:46285", + "issue:43452|issue:46512", + "issue:43262|issue:43450", + "issue:43334|issue:43525", + "issue:45200|issue:46899", + "issue:44442|issue:45198", + "issue:45200|issue:46932", + "issue:44792|issue:45563", + "issue:44898|issue:45412", + "issue:44479|issue:46192", + "issue:43819|issue:44112", + "issue:45200|issue:45325", + "issue:43335|issue:43550", + "issue:41720|issue:46209", + "issue:43827|issue:43901" + ], + "prepared_review_unit_hash": "d3e54d621410d443c06f487aeb437afd505349409aaf467067b41b4d43d2a1e1", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35340, + "estimated_input_tokens": 8835, + "estimated_eval_tokens": 17926 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft links are broad topical similarities across different regressions, not true duplicates. I would not merge any of the proposed soft pairs.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 44521 is the strongest representative of the cluster\u2019s most active multimodal/template-related slice, with high discussion and inbound references; if one issue must serve as canonical, it is the best central example.", + "canonical_pr_reason": null, + "best_issue_reason": "44521 is the most broadly representative and active issue in the set, but only as a loose canonical for the multimodal `apply_chat_template` area; the overall cluster is not a duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Different bugs: generation kwargs handling after `sync_gpus` vs `output_hidden_states` behavior in Qwen3.5." + }, + { + "left": "issue:43295", + "right": "issue:43493", + "accept": false, + "reason": "Different model paths: processor/tokenizer image passthrough regression vs SigLIP2 implementation discrepancy." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both involve `apply_chat_template`, but one is zeroed assistant masks in multimodal inputs and the other is a crash on tool-call messages; not the same concrete bug." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both mention tied weights, but one is serialization of `lm_head.weight` and the other is LoRA merge collapse with extended vocab; different failure modes." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Different issues: Qwen2.5-VL video position IDs vs FP8 inference failure on B200." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "One is dtype drift between training/eval for embeddings; the other is MoE training memory inefficiency." + }, + { + "left": "issue:44737", + "right": "issue:46894", + "accept": false, + "reason": "Unrelated code paths: XLNet positional encoding device placement vs CohereASR double-shift training loss bug." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug is unrelated to the GraniteSpeech CI device-override test failure." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Different failures: Qwen3Moe FP8 config issue vs missing Qwen2_5_VL import." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Importing Qwen2_5_VL is unrelated to DeepSeek detokenization in v5." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Both are v5 regressions, but one corrupts non-persistent buffers on load and the other is a CLIP text-model structure inconsistency." + }, + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "GGUF loading/from_pretrained breakage is unrelated to `from_config(dtype=...)` leaving `config.dtype` stale." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio `apply_chat_template` sampling-rate defaulting and video batched-shape output are different processor bugs." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Similar symptom (`pad_token_id` missing), but different model configs and likely separate fixes; not enough evidence of one shared bug." + }, + { + "left": "issue:45200", + "right": "issue:46899", + "accept": false, + "reason": "Text-only `mm_token_type_ids` defaulting and quantized multimodal embedder dtype casting are different Gemma4 problems." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer auto-loading failure and Wav2Vec2 save/tokenization failure are different model-specific issues." + }, + { + "left": "issue:45200", + "right": "issue:46932", + "accept": false, + "reason": "`mm_token_type_ids` defaults vs unsafe dtype casting in quantized vision/audio embedders are distinct bugs." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image generation test failure is unrelated to a stale paged-generate warning." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver non-default-resolution classification failure is unrelated to RT-DETR memory not being released." + }, + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Video-input regression for Qwen2.5/Qwen3 variants is unrelated to `return_tensors` being ignored when `text_kwargs` is passed." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent conversion mismatch and a GraniteSpeech CI test flake are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma4 token-type defaults and Qwen2.5-VL rope-index scaling are different model bugs." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and Bamba SDPA torch.compile failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Different models and failure modes: Qwen3 auto device mapping assert vs FP8 inference on B200." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but one is about pipeline removal and the other about `return_all_scores` wording; not the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 30, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:37428", + "issue:38175", + "issue:43278", + "issue:43335", + "issue:43388", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43825", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44448", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44855", + "issue:44857", + "issue:44995", + "issue:45083", + "issue:45290", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:45593", + "issue:45759", + "issue:46170", + "issue:46234", + "issue:46285", + "issue:46326", + "issue:46424", + "issue:46512", + "issue:46821", + "issue:46858" + ], + "soft_pairs": [ + "issue:44561|issue:46234", + "issue:44610|issue:44857", + "issue:43819|issue:45083", + "issue:44857|issue:45593", + "issue:43388|issue:43526", + "issue:45759|issue:46285", + "issue:46512|issue:46821", + "issue:44995|issue:46858", + "issue:46424|issue:46858", + "issue:43606|issue:45083", + "issue:44206|issue:45584", + "issue:38175|issue:45412", + "issue:43335|issue:43653", + "issue:37428|issue:43425", + "issue:29942|issue:32090", + "issue:28282|issue:33357", + "issue:43550|issue:44112", + "issue:43611|issue:44466", + "issue:43335|issue:43504", + "issue:44448|issue:44743", + "issue:45520|issue:46170", + "issue:44855|issue:45083", + "issue:43278|issue:43825", + "issue:43335|issue:45083", + "issue:43335|issue:45593", + "issue:43335|issue:44112", + "issue:44514|issue:44521", + "issue:29127|issue:32090", + "issue:45290|issue:46326", + "issue:37428|issue:44351" + ], + "prepared_review_unit_hash": "249e0c9235468b103b0893e19b790de568fdf6b549bcdb5be26ee629c43e416b", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36514, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are only loosely similar by embedding; they span many unrelated bugs across import utils, caching, processors, trainers, and model-specific regressions. None of the soft-pair candidates look like the same underlying bug/change, so all proposed duplicate links should be rejected.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43335 is the densest hub in the similarity graph, so it is the least arbitrary anchor for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43335 is the most connected candidate, but the surrounding items are too heterogeneous for this to represent a true duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Different missing-symbol import regressions in import_utils; not the same code path." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Processor image-size mismatch vs AMP/CUDA loss crash; unrelated failures." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to qwen3_omni_moe feature-length behavior." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "One is a float16 CUDA crash, the other is missing auxiliary losses; different bugs." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Tuple-label truncation in metrics vs BEiT fast processor label reduction; different logic." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Composite Qwen text_config unwrapping is not the same as CLIP text-model structure drift." + }, + { + "left": "issue:46512", + "right": "issue:46821", + "accept": false, + "reason": "Stale dtype metadata vs TimesFM window_size AttributeError; different subsystems." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Stale indexer cache on second forward pass vs static-cache seq-length tensor arithmetic." + }, + { + "left": "issue:46424", + "right": "issue:46858", + "accept": false, + "reason": "Quantized-cache max length handling and static-cache seq-length device tensor are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to the qwen3_omni_moe helper function." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported feature-extractor arg vs empty-transcription generation failure; different paths." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Zero probabilities in SigLIP2 and RT-DETR memory release are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Sparse-layer creation bug vs tokenizer special-token registration bug; not the same issue." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Missing flash-attention utility import is not the same as Torch 2.10 incompatibility." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and NoneType GPU broadcast are different failures." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Missing PyTorch import and Mac bus error in CLIP model are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "torch.compile/SDPA model failure vs stale CI device-override test; different bugs." + }, + { + "left": "issue:43611", + "right": "issue:44466", + "accept": false, + "reason": "base_model_prefix loading regression does not match lm_head.weight serialization inconsistency." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and BEiT legacy-field load failure are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus output regression vs Qwen recurrent-state reset bug; different models and code paths." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 flash_attn import error and ROCm/Windows distributed dependency issue are distinct." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 JIT parse/import problem is unrelated to qwen3_omni_moe length helper behavior." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "dtype drift between train/eval is unrelated to chat-template translation messaging." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Sparse-layer config creation and helper output-length behavior are different problems." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Different models and failure modes; not a shared bug." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Batch padding crash and all-zero assistant masks are related area but not the same concrete bug." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity issue is unrelated to Trainer `_gpu_broadcast_one` TypeError." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Tool-call/no-content chat-template crash vs mistral_common continue_final_message bug; different triggers." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Missing internal flash-attention import and missing HybridCache export are separate import 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": 25, + "review_unit_index": 31, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "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:43526", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43819", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44617", + "issue:44779", + "issue:44857", + "issue:45072", + "issue:45083", + "issue:45127", + "issue:45245", + "issue:45310", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45823", + "issue:45859", + "issue:45920", + "issue:46192", + "issue:46234", + "issue:46431", + "issue:46566", + "issue:46697" + ], + "soft_pairs": [ + "issue:45588|issue:45859", + "issue:42490|issue:43854", + "issue:43720|issue:44112", + "issue:34689|issue:35141", + "issue:43819|issue:44493", + "issue:39692|issue:43526", + "issue:46431|issue:46566", + "issue:43606|issue:44857", + "issue:44448|issue:44779", + "issue:45310|issue:45478", + "issue:26413|issue:31501", + "issue:28218|issue:32090", + "issue:45127|issue:45245", + "issue:45459|issue:46234", + "issue:39692|issue:44493", + "issue:44617|issue:46697", + "issue:43701|issue:45083", + "issue:44206|issue:45823", + "issue:44857|issue:45072", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:43550|issue:45083", + "issue:29942|issue:35141", + "issue:27957|issue:34689", + "issue:38175|issue:44493", + "issue:44360|issue:44561", + "issue:43278|issue:43329", + "issue:29870|issue:32090", + "issue:45920|issue:46192" + ], + "prepared_review_unit_hash": "584599ab22f9b7e994de378dc4e8dc5f2fd2e4fa3231de2038ea2b6d31082e43", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36471, + "estimated_input_tokens": 9118, + "estimated_eval_tokens": 18492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Almost all soft pairs are false positives: they point to different bugs, models, or subsystems. The only clear duplicate is the repeated Qwen3.5 MoE from_pretrained regression in issue:45310 and issue:45478.", + "confidence": 0.97, + "canonical_issue_reason": "issue:45310 is the better canonical issue: it is the earlier report, has a slightly fuller description, and matches issue:45478 exactly in title and bug report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best issue to represent the duplicate pair because it is the original report for the same Qwen3.5 MoE from_pretrained failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both involve attention-related model code, but one is a flash-attention sinkless-model crash and the other is a CPU offload rotary-embedding incompatibility; different failure paths." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path trailing slash warning is unrelated to GLM-4.7-Flash loading correctness in tests; different surfaces and symptoms." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet packed-weight loading failure is unrelated to a GraniteSpeech CI device-override test regression." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision model loading breakage is unrelated to embedding reinitialization after resize_token_embeddings with tie_word_embedding=False." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC latent/forward mismatch is a model math bug; unexpected position-id keys is a separate loading/input-key regression." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs example errors and BeitImageProcessorFast reduce_labels behavior are different components and different bug classes." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "Missing chat template error is unrelated to DiffusionGemma staying on meta with device_map=\"auto\"." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch in bark-small is unrelated to LwDetrImageLoss crashing under float16 AMP on CUDA." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output drift across v4/v5 is a model behavior regression, while DeepSeek tokenizer incorrect IDs is a tokenizer regression." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same title, same Qwen3.5 MoE from_pretrained error, and both describe the same underlying regression." + }, + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "resume_from_checkpoint epoch_iterator failure and FineWeb SLM training not starting are different training issues with no concrete shared code path shown." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer added-space bug and trainer _gpu_broadcast_one NoneType error are unrelated." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge with extended vocabulary causing collapse is unrelated to the category-count limit error." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf error handling and missing is_falcon_mamba_ssm_available helper are unrelated import/runtime issues." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Documentation/example issues do not match a position-id-key regression in model loading." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "Sam3Video OOM and Swin2SR over-padding are unrelated model/image-processing bugs." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is a training/checkpoint issue, not the qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to Gemma4 PLE device-map forward mismatch." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: AMP CUDA crash in LwDetrImageLoss versus bfloat16 inference dtype mismatch for SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "NoneType broadcast error in Trainer is unrelated to a MacOS bus error loading a CLIP community model." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Trailing-slash import warning is unrelated to tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile+SDPA failure in Bamba-9B-v2 is unrelated to the qwen3_omni_moe helper function behavior." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization after resizing token embeddings." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "XLMRoberta with Flash Attention 2 and Llama 3.2 Vision model loading are distinct model-specific problems." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to unexpected position-id keys." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "A missing ReLU discussion in the DSA indexer is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in evaluate is unrelated to undefined video-branch helpers in multimodal token counting." + }, + { + "left": "issue:29870", + "right": "issue:32090", + "accept": false, + "reason": "BART decoder_attention_mask assignment bug is unrelated to the _gpu_broadcast_one NoneType type error." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Wrong AutoTokenizer token IDs across multiple models is unrelated to return_tensors being ignored when text_kwargs is passed." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative overall. The only clearly supported duplicate is issue:45310 / issue:45478 with identical titles and closely matching descriptions. The soft-pair rejections are mostly justified by distinct titles and bug classes, with no obvious overstatement." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 32, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:10000", + "issue:15354", + "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:44448", + "issue:44466", + "issue:44561", + "issue:44568", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45290", + "issue:45381", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46399", + "issue:46421", + "issue:46655", + "issue:46858" + ], + "soft_pairs": [ + "issue:44448|issue:44977", + "issue:45290|issue:45381", + "issue:10000|issue:15354", + "issue:44855|issue:46399", + "issue:44977|issue:45229", + "issue:43278|issue:43381", + "issue:29870|issue:33290", + "issue:40444|issue:43526", + "issue:43927|issue:45698", + "issue:45290|issue:46655", + "issue:44857|issue:45588", + "issue:37428|issue:42175", + "issue:32090|issue:33290", + "issue:46421|issue:46858", + "issue:44561|issue:44779", + "issue:43873|issue:45229", + "issue:42490|issue:43873", + "issue:44448|issue:44466", + "issue:25251|issue:32090", + "issue:43976|issue:45440", + "issue:44568|issue:45491", + "issue:29127|issue:34567", + "issue:45588|issue:45823", + "issue:44871|issue:44977", + "issue:43381|issue:43701", + "issue:33357|issue:39692", + "issue:43278|issue:43575", + "issue:41720|issue:45229", + "issue:44977|issue:45563", + "issue:45198|issue:45803", + "issue:44186|issue:45198" + ], + "prepared_review_unit_hash": "0cdabcb724e49e6308e80727e73d8524298a1c7e651fced0acddb6395a686d88", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37334, + "estimated_input_tokens": 9334, + "estimated_eval_tokens": 18924 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab-bag of unrelated Transformers bugs across generation, tokenizers, multimodal processors, Trainer, cache/offloading, and serialization. I don\u2019t see any soft pair that is clearly the same underlying defect; the strongest representative issue is the active custom-module save/load regression.", + "confidence": 0.62, + "canonical_issue_reason": "issue:45698 is the most representative and actionable bug here: it is open, has multiple inbound references, and describes a concrete save_pretrained/from_pretrained regression with broad impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 has the clearest reproducible failure, active discussion, and broader downstream impact than the other candidates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Both are v5 output regressions, but they hit different models and different code paths (Pegasus generation vs Qwen3.5 flash-attention)." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Different multimodal failures: chat template tool-call handling vs Qwen2.5-VL video vision_position_ids." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Different problem classes: a DistilBert model error vs torch.jit.script GeneratorExp export failure." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both involve model import/runtime errors, but the root causes are unrelated: Python 3.13 TorchScript parsing vs a missing decorator in Qwen3.5 kernel registration." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "One is a generation correctness issue, the other is a multi-GPU CUDA OOM; same model family is not enough to make them duplicates." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Both mention eval-mode behavior, but one is dtype drift and the other is gradient checkpointing being unsupported in eval." + }, + { + "left": "issue:29870", + "right": "issue:33290", + "accept": false, + "reason": "UnboundLocalError during BART generation and DeepSpeed Adafactor OOM are unrelated failure modes." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Different subsystems and symptoms: IterableDataset with multi-image prompts vs BeitImageProcessorFast label reduction bug." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load behavior, but one corrupts config token IDs while the other loads the wrong custom module after save_pretrained." + }, + { + "left": "issue:45290", + "right": "issue:46655", + "accept": false, + "reason": "Both are preprocessing crashes, but the underlying bugs are different: chat-template tool calls vs video padding calculation." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "Different crash sites and causes: AMP/CUDA loss computation vs flash_attention handling of s_aux=None." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "Import failure for a missing flash-attention helper is unrelated to TensorFlow not being installed in a torch-only backend setup." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer distributed broadcast of NoneType and Adafactor OOM in DeepSpeed are separate issues." + }, + { + "left": "issue:46421", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache-related, but one is recompilation from lazy init and the other is device-tensor return value breaking slice arithmetic." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "A removed utility breaking trust_remote_code models is not the same as DeepSeek tokenizer behavior changes in v5." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading not working and Gemma4 multi-GPU OOM are different resource-management bugs." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "A path-trailing-slash import warning is unrelated to quantization offloading behavior." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is output mismatch and the other is serialization of tied weights; different code paths." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior and Trainer broadcast errors are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility for Transformers 5.1.0 is not the same as DeepSeekV3 native-vs-remote implementation divergence." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer BOS/EOS insertion bug and Gemma3 NaN embeddings are distinct tokenizer/model execution problems." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error messaging and TrainerState token counter not updating are different concerns." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Both touch Gemma4/flash attention/device placement, but one is a flash-attention AttributeError and the other is a forward device mismatch." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos_token_id config mismatch and Qwen3.5 flash-attention generation issues do not share the same root cause." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode gradient checkpointing limitation and resume_from_checkpoint key mismatch are unrelated Trainer issues." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error with a community CLIP model and SigLIP2 documentation example errors are not the same defect." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16-to-FP32 dtype drift in eval is unrelated to Qwen2-57B tensor-parallel OOM." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Auto device mapping cudaErrorAssert and Gemma4 multi-GPU OOM are different runtime failures." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation failure and stale paged-generate warning are different behaviors." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure and Fuyu bare-except swallowing errors are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and Wav2Vec2 save/tokenization failure are distinct tokenizer 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": 25, + "review_unit_index": 33, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "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:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43994", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44743", + "issue:45072", + "issue:45198", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45712", + "issue:46121", + "issue:46369", + "issue:46693", + "issue:46701" + ], + "soft_pairs": [ + "issue:44361|issue:45198", + "issue:43611|issue:46121", + "issue:43653|issue:44743", + "issue:29870|issue:34689", + "issue:45372|issue:45588", + "issue:44206|issue:44479", + "issue:43756|issue:45440", + "issue:29942|issue:30064", + "issue:42915|issue:42994", + "issue:43653|issue:44442", + "issue:45459|issue:45712", + "issue:45491|issue:46693", + "issue:44265|issue:45198", + "issue:43412|issue:45588", + "issue:43065|issue:43994", + "issue:29127|issue:33357", + "issue:45372|issue:45406", + "issue:43065|issue:45412", + "issue:45245|issue:45563", + "issue:43575|issue:43701", + "issue:45440|issue:45593", + "issue:42490|issue:43381", + "issue:29127|issue:35141", + "issue:45372|issue:45459", + "issue:43421|issue:45491", + "issue:43994|issue:45412", + "issue:45406|issue:46701", + "issue:43066|issue:44462", + "issue:45412|issue:46369", + "issue:43575|issue:44387", + "issue:44442|issue:45072", + "issue:44464|issue:45584" + ], + "prepared_review_unit_hash": "146ba47a4fd9ad9d9e18451336c012dbdbc5831fc082ade3f935c4dbc7c81b1d", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37374, + "estimated_input_tokens": 9344, + "estimated_eval_tokens": 18944 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated Transformers bug reports spanning tokenizers, model loading, generation, quantization, export, and memory issues. None of the soft pairs looks like the same concrete bug, so this cluster should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue cleanly represents the cluster because the items are heterogeneous. If forced, issue 44462 is the broadest user-facing loading bug, but it is not a true umbrella duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44462 is the most generally phrased and broadly reusable bug report in the set, but the cluster is too diverse for it to be a strong canonical duplicate anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model families and different failures: MLukeTokenizer AttributeError vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Model loading with base_model_prefix is unrelated to a TypeError in rope validation list handling." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug is unrelated to Qwen recurrent-state reset logic." + }, + { + "left": "issue:29870", + "right": "issue:34689", + "accept": false, + "reason": "BART generate() decoder mask assignment error is unrelated to Llama 3.2 vision model loading regression." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import/dependency failure is a different bug from flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported center argument in a feature extractor is unrelated to a Qwen video-input regression." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are different model implementation bugs." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems and symptoms." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Qwen3Moe FP8 loading failure and quantized model saving failure are both quantization-related but not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer-related, but BigBird mask-token registration and FastSpeech2ConformerTokenizer loading are distinct issues." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Hiding tokenizer decode errors when protobuf is missing is unrelated to leftover dummy classes in dummy_pt_objects.py." + }, + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Gemma3 NaNs in mixed-length batching are a correctness bug, while the FA2 issue is a performance regression." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures from torch_compilable_check are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/erfinv support and flash_attention sinkless crashes are different code paths." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "A dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to SigLIP2 producing nonsense outputs through AutoModel/pipeline." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMV3 error-message clarity and a MacOS bus error in CLIP loading are not the same underlying bug." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma 4 processor import failure and serve crashing due to missing _tokenizer are different defects in different paths." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy Conv2d placeholders and RT-DETR memory leaks are unrelated issues." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count runtime limit and stale warning behavior in paged generate() are unrelated generation bugs." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM on load and resume_from_checkpoint key mismatch are distinct failures." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE divergence from remote implementation and D-FINE auxiliary-loss omission are unrelated model bugs." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Import warning for local paths ending in slash is unrelated to gradient checkpointing being disallowed in eval mode." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMV3 box-information messaging and post_init reinitializing output embeddings are different bugs." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 loading import failure and protobuf error masking in tokenizer code are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token/post-processor update behavior is unrelated to Gemma3 NaNs with mixed-length batching." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 pipeline/AutoModel output issues are unrelated to RT-DETR memory retention after deletion." + }, + { + "left": "issue:45406", + "right": "issue:46701", + "accept": false, + "reason": "Gemma4Processor serve crash and DiffusionGemmaGenerationOutput pipeline failure are different issues." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and AutoTokenizer ignoring tokenizer.json are related broadly, but not the same concrete bug." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "RT-DETR garbage-collection memory leak and sam3-litetext memory overflow are different memory problems." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Tensor-parallel OOM on model load and int4 quantization reserved-memory regression are distinct bugs." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward and Whisper empty-transcription failure after align_special_tokens are different generation 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": 25, + "review_unit_index": 34, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29870", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43334", + "issue:43381", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43646", + "issue:43720", + "issue:43761", + "issue:43824", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44186", + "issue:44206", + "issue:44279", + "issue:44442", + "issue:44514", + "issue:44743", + "issue:44811", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45245", + "issue:45250", + "issue:45381", + "issue:45440", + "issue:45588", + "issue:45834", + "issue:45859", + "issue:46121", + "issue:46234", + "issue:46369", + "issue:46399", + "issue:46640", + "issue:46693", + "issue:46821" + ], + "soft_pairs": [ + "issue:43526|issue:45834", + "issue:42915|issue:43931", + "issue:43526|issue:46693", + "issue:43824|issue:44279", + "issue:27957|issue:29127", + "issue:27957|issue:33290", + "issue:43994|issue:46369", + "issue:43526|issue:44186", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:43334|issue:43531", + "issue:46121|issue:46399", + "issue:45250|issue:46640", + "issue:43381|issue:43873", + "issue:45588|issue:46821", + "issue:43761|issue:45859", + "issue:44514|issue:45381", + "issue:43646|issue:44077", + "issue:44977|issue:45245", + "issue:44743|issue:44811", + "issue:44442|issue:44898", + "issue:43262|issue:44206", + "issue:42490|issue:44743", + "issue:45003|issue:45440", + "issue:43720|issue:44442", + "issue:42175|issue:46234", + "issue:27957|issue:29870", + "issue:43493|issue:43827" + ], + "prepared_review_unit_hash": "03780444fb37c5e268a14e6bf721581b3c14a59d3058461818dd89cd6c6a0458", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35971, + "estimated_input_tokens": 8993, + "estimated_eval_tokens": 18242 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No strong duplicate clusters emerge from the provided issues; most soft pairs are thematically related at a high level only and do not describe the same concrete bug or change.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46640 is the broadest and most current Flash Attention 2 report among the set, so it works best as a representative umbrella issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46640 has the clearest generic FA2 framing, recent activity, and inbound references, making it the best single issue to anchor this group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "Different components and failures: BeitImageProcessorFast label reduction vs Kosmos2.5 OCR long-input indexing." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Different models and bugs: Qwen3MoE FP8 config failure vs Qwen3-VL weight-shape loading mismatch." + }, + { + "left": "issue:43526", + "right": "issue:46693", + "accept": false, + "reason": "Unrelated issues: image processor label handling vs FA2 long-context performance regression." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "ImportError for a missing Qwen2.5-VL class is not the same as a generic dependency issue." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Different subsystems: XLM-RoBERTa Flash Attention 2 vs LayoutLMv3 error-message clarity." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "Flash Attention 2 model support is unrelated to Adafactor OOM in DeepSpeed." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "Different model families and symptoms: SigLIP2 bad outputs vs SAM3 LiteText memory overflow." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "Different tokenizer/processor bugs in different models; not the same code-path." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Optimizer memory issue is unrelated to Llama 3.2 Vision loading breakage." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM in DeepSpeed Adafactor is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Both are Qwen-related, but one is a missing config attribute and the other is a sliding-window behavior issue." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Both mention rope/rotary code, but they are different failures in different paths and models." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "Too generic to treat as duplicates; both mention FA2, but the concrete bug/change is not the same from the titles alone." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization/offloading are separate behaviors." + }, + { + "left": "issue:45588", + "right": "issue:46821", + "accept": false, + "reason": "Different AttributeError sites and different models; no shared underlying bug is evident." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIP hidden_states regression is unrelated to CPU-offload incompatibility in Qwen rotary embeddings." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL, but batched chat-template padding crashes and wrong video vision_position_ids are distinct bugs." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "The second is a specific patchtsmixer post_init constraint, not clearly the same concrete issue as the broad v5 custom-init regression." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure is unrelated to the category-cardinality runtime error." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different model paths and failures: recurrent-state reset vs Whisper batch_decode token skipping." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure and Perceiver image-resolution failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default and unsupported center arg in feature extraction are different bugs." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning is unrelated to recurrent-state cache behavior." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access and DeepSeekV3 divergence are different code paths." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to FastSpeech2ConformerTokenizer loading." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Packaging/backend dependency coverage and a missing import_utils helper are not the same bug." + }, + { + "left": "issue:27957", + "right": "issue:29870", + "accept": false, + "reason": "Different model/task failures: XLM-RoBERTa FA2 vs BART generate() decoder_attention_mask bug." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and docs still referencing pipeline() 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": 25, + "review_unit_index": 35, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42617", + "issue:42907", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43525", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43701", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43825", + "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:43316|issue:43525", + "issue:45325|issue:46697", + "issue:44062|issue:45907", + "issue:43873|issue:44871", + "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:43749|issue:43931", + "issue:43316|issue:45759", + "issue:44442|issue:45803", + "issue:44987|issue:46178", + "issue:44190|issue:44496", + "issue:43761|issue:44079", + "issue:43232|issue:43388", + "issue:43931|issue:45216", + "issue:43761|issue:44188", + "issue:46829|issue:46932", + "issue:43701|issue:46506", + "issue:42617|issue:43575", + "issue:42907|issue:45698", + "issue:43575|issue:43742", + "issue:43550|issue:43701", + "issue:27957|issue:32090", + "issue:43606|issue:43701", + "issue:25251|issue:28282", + "issue:43295|issue:43825", + "issue:16998|issue:22355", + "issue:43295|issue:44871" + ], + "prepared_review_unit_hash": "ce19fc3d99b54ed64bbd1a2ff91bf4c0d28bb61b9a99a49e072ae66b254ca79a", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36167, + "estimated_input_tokens": 9042, + "estimated_eval_tokens": 18340 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the issues span unrelated bugs in loading, generation, tokenization, config handling, offloading, and training utilities. None of the soft pairs appears to describe the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No clear canonical issue: the set mixes unrelated failures across many subsystems with no shared root cause or duplicate thread.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a strong global representative; 43873 has the broadest/highest-signal discussion, but it still describes a different offloading/quantization problem than the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Different problems: a patchtsmixer post_init policy question versus a video-input regression in Qwen-like models." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API mismatches, but they involve different missing fields and different models, so not the same bug." + }, + { + "left": "issue:45325", + "right": "issue:46697", + "accept": false, + "reason": "One is a Qwen2.5-VL rope/position_ids scaling bug; the other is Swin2SR image over-padding." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer AddedToken argument collision is unrelated to list_repo_files not forwarding revision." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Offloading/quantization behavior is unrelated to a Gemma-3 eos_token_id configuration inconsistency." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 TorchScript parsing error is unrelated to a DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are reload/config issues, but one drops fields with trust_remote_code while the other resets rotary_pct on reload." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading failure and Qwen3-VL weight-shape mismatch are different loading bugs." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "A Gemma3TextConfig field discrepancy is not the same as failing to unwrap text_config in composite Qwen multimodal configs." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure and a bare-except swallowing errors are unrelated code paths." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "A model-loading regression is unrelated to a missing BACKENDS_MAPPING entry for tensorflow_text." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in a training script is unrelated to an unrecognized model config/model_type error." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel failing to return hidden_states is not the same as generic ModelOutput key assignment behavior." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs syncing and gather_for_metrics label truncation are different training utilities with different failures." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Weight-shape mismatch on load is unrelated to save_pretrained producing a wrong checkpoint for Qwen3.5." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Hidden_states output regression is unrelated to attention-kernel divergence under torch.compile." + }, + { + "left": "issue:46829", + "right": "issue:46932", + "accept": false, + "reason": "A wrong hidden_states field in classification output is unrelated to dtype casting for quantized Gemma4 embedders." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Checkpoint key mismatch has no relation to CPU/CUDA result divergence in PPDocLayoutV3." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py failing and Qwen2-57B tensor-parallel OOM are different distributed-memory issues." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving a dequantized model and loading the wrong custom module after save_pretrained are different save/load bugs." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "An OOM during tensor-parallel load is not the same as a key error when loading MobileLLM-125M." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA behavior is unrelated to resume_from_checkpoint key mismatches." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "XLMRoberta flash-attention support and a NoneType GPU broadcast error are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to checkpoint resume key mismatches." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k nesting behavior is unrelated to missing PyTorch dependency detection." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor/tokenizer regression in v4.57.5 is unrelated to a pipeline error-message wording issue in v5." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to the missing transformers.onnx import error." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to Gemma-3 eos_token_id inconsistency." + } + ] + }, + "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": 25, + "review_unit_index": 36, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:43065", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43646", + "issue:43650", + "issue:43688", + "issue:43716", + "issue:43761", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44464", + "issue:44479", + "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:46821", + "issue:46894" + ], + "soft_pairs": [ + "issue:44871|issue:45584", + "issue:46234|issue:46291", + "issue:44464|issue:46421", + "issue:44077|issue:44534", + "issue:43316|issue:43716", + "issue:43316|issue:46821", + "issue:43646|issue:44913", + "issue:44871|issue:46050", + "issue:43452|issue:44496", + "issue:45803|issue:46365", + "issue:44162|issue:44792", + "issue:43404|issue:43441", + "issue:44242|issue:46894", + "issue:45632|issue:46424", + "issue:45803|issue:46170", + "issue:44610|issue:46369", + "issue:43650|issue:46202", + "issue:43540|issue:44514", + "issue:43761|issue:44479", + "issue:43065|issue:44617", + "issue:45542|issue:46234", + "issue:43504|issue:45698", + "issue:43572|issue:46178", + "issue:44792|issue:44977", + "issue:43388|issue:43688", + "issue:44792|issue:45600", + "issue:45127|issue:45200" + ], + "prepared_review_unit_hash": "604c7ac5d159a1f6220561239c318f2c2103031f75710d40d18a9684db147652", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35742, + "estimated_input_tokens": 8936, + "estimated_eval_tokens": 18128 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of semantically similar but actually unrelated Transformers issues. The soft pairs share broad themes like loading, config, cache, or multimodal regressions, but not the same concrete bug/change path, so I would not merge any of them.", + "confidence": 0.83, + "canonical_issue_reason": "There is no single true canonical issue here because the cluster is heterogeneous. If an anchor is required, issue 45698 is the broadest representative of the loading/save-pretrained regression subtheme, but it should not be treated as a duplicate umbrella for the rest.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45698 is the best global anchor: it has the strongest discussion/references and captures a common serialization/loading failure mode that appears more central than the other one-off regressions. Still, it is not a valid duplicate representative for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Different models and failure modes: Gemma EOS config inconsistency vs Whisper empty-transcription generation after token alignment." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "Both are import/runtime dependency problems, but one is a missing utility symbol and the other is a pipelines import error from kernels versioning; not the same bug." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked generation/cache, but one is wrong outputs under compiled forward and the other is spurious recompilation from lazy cache init." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Different regressions: optional post_init validation for patchtsmixer vs non-persistent buffers being filled with junk in v5." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Unrelated model/config discrepancy versus image preprocessor dtype mismatch; same release family at most, not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:46821", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to TimesFM window_size AttributeError." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "General v5 custom model init breakage vs GPTNeoX rotary_pct reload bug; different code paths and symptoms." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "EOS token config mismatch is unrelated to sinq quantization save/load corruption." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both are from_pretrained loading issues, but one is gguf_file handling and the other is missing model_type in config; different root causes." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare except in tensor conversion is unrelated to Trainer import failure caused by datasets dependency interaction." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 model breakage and Janus image-generation test failure are different subsystems and bugs." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Mistral3 weight tying bug and Ministral-3 FlashAttention failure are distinct model issues." + }, + { + "left": "issue:44242", + "right": "issue:46894", + "accept": false, + "reason": "Load-balancing loss gating is unrelated to CohereASR training-loss double-shift behavior." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Cache path collision for trust_remote_code is unrelated to max_cache_length being honored only for quantized cache." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Different failure classes: exception swallowing in tensor conversion vs Windows/ROCm import error due to torch.distributed dependency." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Processor image-size mismatch is not the same as Sam3LiteText memory overflow." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "No meaningful relation: placeholder issue title vs Gemma roundtrip failure on a specific token." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Different Qwen multimodal bugs: video input ValueError vs batched chat-template crash with padding=False." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to Qwen video-input regression." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and Sam3Video OOM are not the same concrete defect." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend mismatch is unrelated to missing falcon mamba availability helper." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy-field Beit load failure and wrong custom module after save/load are both loading bugs, but not the same code path." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm pad_token_idx regression is unrelated to tensorflow_text backend mapping." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus test failure and Qwen3.5 flash-attention generation issue are different model behaviors." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Mask2Former metric gathering bug is unrelated to auxiliary loss normalization in OLMoE/GPT Oss." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Image generation test failure is unrelated to auto_mappings referencing removed Sam3LiteText configs." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with extended vocabulary is unrelated to Gemma 4 mm_token_type_ids defaulting." + } + ] + }, + "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": 25, + "review_unit_index": 37, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:29870", + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43335", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43644", + "issue:43673", + "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:44933", + "issue:44964", + "issue:45072", + "issue:45245", + "issue:45362", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45632", + "issue:46082", + "issue:46421", + "issue:46617", + "issue:46640", + "issue:46697", + "issue:46832", + "issue:46858", + "issue:46897", + "issue:46932" + ], + "soft_pairs": [ + "issue:43673|issue:44743", + "issue:43352|issue:46640", + "issue:44190|issue:44964", + "issue:44534|issue:44913", + "issue:41950|issue:43475", + "issue:43525|issue:43531", + "issue:44734|issue:45406", + "issue:35141|issue:42490", + "issue:45072|issue:46932", + "issue:43827|issue:43994", + "issue:44855|issue:45588", + "issue:45245|issue:46858", + "issue:44869|issue:46832", + "issue:46082|issue:46697", + "issue:45632|issue:46617", + "issue:29870|issue:35141", + "issue:43673|issue:46421", + "issue:44617|issue:45584", + "issue:43931|issue:44060", + "issue:44242|issue:46897", + "issue:43644|issue:44913", + "issue:25251|issue:27957", + "issue:43335|issue:44855", + "issue:45563|issue:45584", + "issue:44933|issue:45362", + "issue:44077|issue:44877", + "issue:43550|issue:44190" + ], + "prepared_review_unit_hash": "4e1b6ed0f388e63195a8a0cd96ca876919cdc0c7e68a60443a846073589939c0", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35852, + "estimated_input_tokens": 8963, + "estimated_eval_tokens": 18182 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy: the items span many unrelated bugs across generation, caching, model configs, pipelines, and serving. I found no reliable duplicate pair among the soft edges, so this should not be merged into a single canonical triage bucket.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache/chunked prefill, but one is a missing GenerationMixin cache in v5.0.0 and the other is recurrent state reset in Qwen3 modular code; different bugs and code paths." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "One says NemotronHForCausalLM does not support Flash Attention 2.0; the other is a generic Flash Attention 2 issue. Too vague and not the same concrete failure." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Different surfaces entirely: a local dataset loading failure in an example script versus a model loading error for Phi-4 multimodal." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is a global non-persistent buffer initialization regression; the other is GPTNeoXConfig serialization/reload losing rotary_pct. Different defects." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification pipeline processor selection and a SAM3 video encoder output attribute error are unrelated model/pipeline bugs." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Llama4Config pad_token_id absence and Qwen3-MoE sliding_window behavior are separate configuration issues." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve-related, but one crashes on KV-cache continuation indexing while the other lacks a tokenizer on Gemma4Processor. Different code paths." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Resizing token embeddings reinitializing output embeddings is unrelated to local-path import warnings with a trailing slash." + }, + { + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "Both involve dtype behavior, but one is bfloat16 inference mismatches across models and the other is quantized Gemma4 vision/audio embedder casting; not the same bug." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing removed pipeline APIs is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "An import-time Python 3.13 IndentationError in DebertaV2 is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:45245", + "right": "issue:46858", + "accept": false, + "reason": "Category-cardinality overflow and static-cache slice arithmetic with device tensors are distinct runtime errors." + }, + { + "left": "issue:44869", + "right": "issue:46832", + "accept": false, + "reason": "Whisper trailing replacement-character decode failure and RT-DETR feature-level mismatch are unrelated model-specific bugs." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are different configuration/processing issues." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both involve trust_remote_code caching, but one is a leaf-directory name collision and the other is a FileNotFoundError when loading from cache; not the same concrete defect." + }, + { + "left": "issue:29870", + "right": "issue:35141", + "accept": false, + "reason": "Bart generate() decoder_attention_mask initialization bug and embedding reinitialization after resize are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill/cache, but one is a missing cache in v5.0.0 and the other is lazy cache initialization causing recompilation on non-CUDA backends; different symptoms and fixes." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video CUDA OOM and Whisper empty-transcription handling are completely different failures." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and Qwen3-Next tied-weights warning are separate model-loading issues." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "Load-balancing loss omission in MoE routing and Florence2/Moonshine double-shift training loss are unrelated training bugs." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "A global v5 non-persistent buffer regression is not the same as GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Top_k pipeline nesting bug and XLMRoberta Flash Attention 2 support are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and DebertaV2 Python parsing/import failure are unrelated." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Paged generate warning behavior and Whisper empty transcription failure do not share the same underlying bug." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A nonexistent image_utils import and a Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Optional post_init policy for patchtsmixer and strict config loading for granite_speech are different configuration-policy issues." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and local dataset loading in an image-classification script 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": 25, + "review_unit_index": 38, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:25251", + "issue:29870", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43526", + "issue:43540", + "issue:43643", + "issue:43673", + "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:44737", + "issue:44779", + "issue:44938", + "issue:44964", + "issue:45003", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45290", + "issue:45357", + "issue:45381", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:46234", + "issue:46369", + "issue:46489", + "issue:46655", + "issue:46697", + "issue:46752", + "issue:46858" + ], + "soft_pairs": [ + "issue:45381|issue:46369", + "issue:41720|issue:45245", + "issue:44479|issue:44734", + "issue:43335|issue:44190", + "issue:43065|issue:44610", + "issue:39692|issue:44190", + "issue:44464|issue:45563", + "issue:45127|issue:45357", + "issue:45381|issue:46655", + "issue:44617|issue:46369", + "issue:43295|issue:45276", + "issue:44779|issue:46489", + "issue:44112|issue:45600", + "issue:45290|issue:46752", + "issue:43526|issue:46697", + "issue:45276|issue:45584", + "issue:43540|issue:44734", + "issue:44737|issue:45803", + "issue:44521|issue:45381", + "issue:44938|issue:45542", + "issue:25251|issue:29870", + "issue:44462|issue:45907", + "issue:43854|issue:44964", + "issue:43673|issue:46858", + "issue:43643|issue:44496", + "issue:45632|issue:46234", + "issue:44246|issue:45003", + "issue:44964|issue:45600", + "issue:44112|issue:44964" + ], + "prepared_review_unit_hash": "343ac7dc26d7be1d1ee8b5abe07b4f57bc07fea46728f9eaf9540ebeb246c421", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36608, + "estimated_input_tokens": 9152, + "estimated_eval_tokens": 18560 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: the issues span unrelated bugs in generation, multimodal preprocessing, model loading, CI, and config handling. None of the proposed soft pairs looks like the same underlying bug, so no merges are warranted.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Different bugs: Qwen2.5-VL video position IDs vs Sam3LiteText memory overflow." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Unrelated failures: Qwen3 auto device mapping CUDA assert vs category-count limit error." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Different code paths: multimodal video regression vs serve KV-cache continuation indexing." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Sparse-layer config bug and local dataset loading are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Different model issues: dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo input sizing." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "SigLIP2 docs example errors are unrelated to dataset loading failure." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Chunked-generation inconsistency and stale warning behavior are not the same bug." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA/vocab collapse and visual-encoder save_pretrained keys are distinct issues." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-related, but the model/code paths and failure modes differ." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Sam3Video OOM and Sam3LiteText memory overflow are different variants and symptoms." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor.tokenizer regression and gemma4 embedding resize behavior are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:46489", + "accept": false, + "reason": "Related family, but one is a broad tokenizer regression and the other a specific PR gap." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Stale GraniteSpeech CI test vs removed Sam3LiteText config references are unrelated." + }, + { + "left": "issue:45290", + "right": "issue:46752", + "accept": false, + "reason": "Both hit apply_chat_template, but one is tool-call handling and the other empty-list handling." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "BEiT label reduction and Swin2SR over-padding are different image-processor bugs." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resize and Whisper empty transcription are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are separate problems." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet device placement bug and Fuyu bare-except handling are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal template issues, but they affect different outputs and models." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import failure and tensorboard/TensorFlow backend issue are different environment problems." + }, + { + "left": "issue:25251", + "right": "issue:29870", + "accept": false, + "reason": "Pipeline top_k nesting behavior and BART UnboundLocalError are unrelated generate-time bugs." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "AutoTokenizer tokenizer.json lookup and list_repo_files revision forwarding are unrelated API bugs." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different model-load regressions affecting different models and loaders." + }, + { + "left": "issue:43673", + "right": "issue:46858", + "accept": false, + "reason": "Both involve cache, but missing cache in chunked_prefill and device-tensor seq_length arithmetic are distinct." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "trust_remote_code config-field loss and unrecognized model_type are different config-loading failures." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Local trust_remote_code cache collision and missing import_utils symbol are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import latency and unsafe sys.modules access are different import-system issues." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Phi-4 multimodal loading and removed Sam3LiteText auto_mappings references are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "GraniteSpeech CI flake and Phi-4 loading failure do not share a concrete 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": 25, + "review_unit_index": 39, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:42907", + "issue:43295", + "issue:43381", + "issue:43404", + "issue:43441", + "issue:43479", + "issue:43526", + "issue:43531", + "issue:43582", + "issue:43650", + "issue:43673", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44220", + "issue:44279", + "issue:44295", + "issue:44336", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44792", + "issue:44987", + "issue:45042", + "issue:45083", + "issue:45216", + "issue:45335", + "issue:45375", + "issue:45446", + "issue:45588", + "issue:45593", + "issue:45600", + "issue:45759", + "issue:46082", + "issue:46121", + "issue:46424", + "issue:46506", + "issue:46932" + ], + "soft_pairs": [ + "issue:27957|issue:35141", + "issue:42175|issue:44279", + "issue:42796|issue:44521", + "issue:46082|issue:46121", + "issue:36246|issue:41628", + "issue:32090|issue:35141", + "issue:43404|issue:45216", + "issue:43531|issue:45600", + "issue:36246|issue:38617", + "issue:41950|issue:44016", + "issue:44295|issue:46506", + "issue:43650|issue:44220", + "issue:43673|issue:46424", + "issue:43582|issue:46932", + "issue:43295|issue:45335", + "issue:43976|issue:44987", + "issue:45083|issue:46506", + "issue:44016|issue:44220", + "issue:44496|issue:45042", + "issue:43526|issue:43901", + "issue:44683|issue:45446", + "issue:43381|issue:45593", + "issue:42907|issue:43441", + "issue:43901|issue:44496", + "issue:43673|issue:44336", + "issue:43479|issue:45375", + "issue:44496|issue:45759", + "issue:43976|issue:44279", + "issue:44016|issue:44062", + "issue:43441|issue:45588", + "issue:44792|issue:46424" + ], + "prepared_review_unit_hash": "6582b5dbf8c760da96a85912c9ecfb7054a637abc866ba9de31795e661bb294b", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37053, + "estimated_input_tokens": 9264, + "estimated_eval_tokens": 18784 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; the items span many unrelated Transformers bugs, regressions, and documentation issues. None of the soft pairs look like the same underlying bug/change, so all are rejected.", + "confidence": 0.93, + "canonical_issue_reason": "No true canonical issue stands out because the cluster is heterogeneous. If one must be chosen, issue:44521 is the most active and broadly referenced, but it is still only a specific multimodal `apply_chat_template` bug and does not represent the rest of the items.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the best single issue to keep as a representative ticket because it is open, has the most inbound references and discussion, and describes a concrete reproducible bug with clear user impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Both mention embedding/Flash Attention related wording only at a very high level; one is about XLM-RoBERTa with Flash Attention 2, the other about embedding reinitialization after resizing. Different bugs and code paths." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "One is about TensorFlow not being included when installing the torch extra; the other is a vague dependency issue. No concrete shared bug." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer BPE marker handling and multimodal assistant mask generation are unrelated problems in different subsystems." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Llama head_dim validation and rope-parameter JSON list handling are both config-validation topics, but they are distinct failures with different inputs and code paths." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules (`Qwen2_5_VLImageProcessor` vs `AutoImageProcessor`). Not the same underlying bug." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both involve model configuration/loading for multimodal models, but one is about untied lm_head weights in Mistral3 and the other about Qwen3.5 checkpoint saving regression. Different concrete defects." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Sliding-window behavior for Qwen3-MoE and auto_mappings referencing removed configs are separate issues with no shared code-path." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import-related breakages, but they concern different missing exports from different files." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline looking for image processors is unrelated to a notebook syntax error." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "Registering `position_ids` as a buffer causing read errors and CPU/CUDA output differences in PPDocLayoutV3 are different behavioral bugs." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "The titles are generic, but one concerns missing data and the other `_torch_extract_fbank_features()`; there is no evidence they are the same bug." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Generation cache missing in v5.0.0 during chunked_prefill and cache max length only honored for Quantized Cache are related to caching, but they are distinct features and failure modes." + }, + { + "left": "issue:43582", + "right": "issue:46932", + "accept": false, + "reason": "Apple Silicon TypeError in `caching_allocator_warmup` and Gemma4 dtype casting for quantized embedders are different issues in different subsystems." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor/tokenizer regression and t5gemma decoder embedding resize behavior are not the same underlying bug." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version incompatibility and model-loading failure for `physical-intelligence/fast` are different release/load issues." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "A helper length calculation bug in qwen3_omni_moe is unrelated to CPU/CUDA result divergence in PPDocLayoutV3." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "A notebook syntax error and an fbank feature extraction issue are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model type in config loading and PIL image processors requiring torchvision are different bugs." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "A BEiT image-processor label reduction bug and docs mismatch for `return_all_scores` are unrelated." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention failing on torch>=2.9 and incorrect PyTorch version check for AuxRequest import are distinct flex_attention-related issues, but not the same code-path problem." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary losses when denoising is off are unrelated model-training bugs." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and Ministral-3 flash-attention failure are different issues in different parts of the stack." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs mentioning `return_all_scores` and an unrecognized model config error are not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Cache behavior in generation and ANSI codes in loading_report are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Phi4 multimodal default config initialization and missing `deepstack_visual_indexes` in Qwen3_5MoeVisionConfig are different config-construction bugs." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "An unrecognized model type error and `from_config` not unwrapping `text_config` for composite Qwen configs are distinct model-loading problems." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version support and dependency issue are too vague and clearly not the same failure." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and `AddedToken()` duplicate `special` keyword error are unrelated." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Ministral flash-attention failure and `flash_attention.py` crashing on `s_aux=None` are both flash-attention themed but involve different model assumptions and code paths." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "A failed janus image-generation test and cache max-length 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": 25, + "review_unit_index": 40, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:39692", + "issue:41950", + "issue:42796", + "issue:43295", + "issue:43299", + "issue:43479", + "issue:43525", + "issue:43577", + "issue:43650", + "issue:43653", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44327", + "issue:44351", + "issue:44368", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44779", + "issue:44933", + "issue:45020", + "issue:45081", + "issue:45200", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45584", + "issue:45593", + "issue:45812", + "issue:45823", + "issue:46082", + "issue:46123", + "issue:46209", + "issue:46369", + "issue:46399", + "issue:46512", + "issue:46515", + "issue:46693", + "issue:46752" + ], + "soft_pairs": [ + "issue:43824|issue:44933", + "issue:44521|issue:46693", + "issue:44279|issue:46515", + "issue:45335|issue:45584", + "issue:44327|issue:46752", + "issue:46123|issue:46369", + "issue:43299|issue:45216", + "issue:46082|issue:46399", + "issue:44279|issue:45405", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:44075|issue:44368", + "issue:43577|issue:45216", + "issue:42796|issue:43650", + "issue:46082|issue:46512", + "issue:44351|issue:45278", + "issue:43976|issue:45405", + "issue:41950|issue:45020", + "issue:44779|issue:45812", + "issue:43299|issue:45357", + "issue:44466|issue:45276", + "issue:43299|issue:43854", + "issue:43653|issue:45593", + "issue:44242|issue:45200", + "issue:41950|issue:45823", + "issue:43650|issue:46209", + "issue:43479|issue:43525", + "issue:44462|issue:45081" + ], + "prepared_review_unit_hash": "31b4743f0eb0bdbc4df714c6a8020cc9cefb79b8706ae3efa90fe619776411ac", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35588, + "estimated_input_tokens": 8897, + "estimated_eval_tokens": 18050 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft pairs share broad themes like tokenizers, import errors, configs, and multimodal regressions, but they do not look like the same concrete bug/change. I would use issue 45020 as the best anchor because it is the most general, well-described regression report in the set.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45020 is the broadest and most representative regression report here, with detailed repro context and high discussion activity, so it works best as an anchor for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the strongest standalone issue in the set: it is detailed, high-signal, and describes a concrete regression with broader applicability than the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they concern different missing symbols/modules and different failure modes; not the same underlying bug." + }, + { + "left": "issue:44521", + "right": "issue:46693", + "accept": false, + "reason": "One is a multimodal masking bug in chat templating, the other is a long-context FA2 performance regression; unrelated code paths." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "A dependency/install issue is not the same as a torch-version incompatibility; the underlying problems differ." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "These are different model behaviors in different subsystems; resize_token_embeddings and empty-transcription generation are unrelated." + }, + { + "left": "issue:44327", + "right": "issue:46752", + "accept": false, + "reason": "Both are crashes, but they occur in different pipelines with different inputs and stack paths; not a duplicate bug." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Batch-finalization logic in a pipeline is unrelated to a model memory overflow issue." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen models, but one is a load-time regression and the other is a save_pretrained serialization bug; different code paths." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "Config validation for head_dim is not the same as a kernel registration/decorator crash during rotary embedding setup." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Dependency issues and an unreleased minimum-version bump are different release/package problems." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and distinct behaviors (processor/tokenizer API vs DeepSeek tokenization output)." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Documentation mistakes in different pipelines are not the same issue; the examples and broken behavior differ." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "An optimizer argument bug is unrelated to a model warning about tied word embeddings." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Loading dtype propagation in BLIP2 is unrelated to Qwen3.5 checkpoint serialization." + }, + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE marker handling has no relation to the unrelated placeholder issue 'ADD THE DATA'." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "These both mention config fields, but one is a validation error and the other is stale dtype metadata; different bugs." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "45278 is a broad umbrella over many import errors, but 44351 is a specific missing import; not enough evidence they are the same concrete defect." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "A Python-version compatibility report and an unreleased dependency version bump are different problems." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video pipeline processor selection and remote_code model breakage are unrelated subsystems." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "These are different tokenizer regressions affecting different model families and token-ID behaviors." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a loading failure and the other is incorrect save_pretrained visual-encoder keys; different bugs." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both concern weight/tied-embedding serialization or resizing, but the specific failure modes and affected components differ." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models, different loading failures, and no shared concrete code path are evident." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "A tokenizer special-token registration bug and a missing auxiliary-loss issue are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "These are both model-training logic bugs, but one is router loss handling and the other is defaulting multimodal token type IDs; different fixes." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video classification processor lookup and Gemma4 device placement are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "A placeholder/triage issue is not the same as an FP8 inference failure on B200." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config classes and different missing/default-field behavior; not the same underlying defect." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Both touch tokenizer loading, but one is AutoTokenizer repository resolution and the other is a Mistral regex patch crash; 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": 25, + "review_unit_index": 41, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:33290", + "issue:42175", + "issue:42796", + "issue:42994", + "issue:43317", + "issue:43441", + "issue:43525", + "issue:43643", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44291", + "issue:44336", + "issue:44361", + "issue:44589", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45081", + "issue:45103", + "issue:45362", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45859", + "issue:45907", + "issue:46851" + ], + "soft_pairs": [ + "issue:44016|issue:44291", + "issue:44683|issue:45542", + "issue:44361|issue:45593", + "issue:43441|issue:45834", + "issue:45081|issue:45907", + "issue:33290|issue:45656", + "issue:45103|issue:45372", + "issue:42994|issue:43317", + "issue:44336|issue:46851", + "issue:45859|issue:46851", + "issue:44623|issue:44749", + "issue:44261|issue:44263", + "issue:45600|issue:45712", + "issue:42175|issue:43976", + "issue:44671|issue:44779", + "issue:43901|issue:43994", + "issue:44623|issue:45397", + "issue:43673|issue:43761", + "issue:43901|issue:44246", + "issue:43723|issue:44162", + "issue:43525|issue:43881", + "issue:43824|issue:43976", + "issue:42796|issue:43441", + "issue:44016|issue:45823", + "issue:45542|issue:45800", + "issue:44188|issue:44336", + "issue:43525|issue:43643", + "issue:44589|issue:45362" + ], + "prepared_review_unit_hash": "d76137ee7c6f60395b452c9a4d5bddd9e427506744e3d57526f1d157df51a103", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35434, + "estimated_input_tokens": 8859, + "estimated_eval_tokens": 17974 + }, + "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, mostly v5 regressions and model-specific bugs. None of the proposed soft pairs look like true duplicates, so they should stay separate.", + "confidence": 0.9, + "canonical_issue_reason": "issue:44291 is a clear, concrete core-library regression with a reproducible loading failure and a well-defined API break, making it the strongest single anchor for this noisy cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44291 is the best standalone representative: specific, actionable, and centered on a fundamental model-loading path rather than a niche model or docs-only problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights TypeError; different surfaces and failure modes." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention failure vs missing TensorFlow backend dependency; unrelated bugs." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; no shared code path." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention model failure vs long-input OCR index error; different model-specific issues." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Mistral tokenizer regex patch crash vs repo file revision forwarding; unrelated functionality." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention DeepSpeed, but one is Adafactor OOM and the other is double optimizer stepping; distinct bugs." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotations crash vs missing ReasoningEffort import; separate failure paths." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized save failure vs dequantized load/offload failure; related theme but different operations and root causes." + }, + { + "left": "issue:44336", + "right": "issue:46851", + "accept": false, + "reason": "ANSI code output in loading_report vs Qwen3.5 variable-length position_ids bug; unrelated." + }, + { + "left": "issue:45859", + "right": "issue:46851", + "accept": false, + "reason": "Same model family, but CPU-offload rotary embedding issue and variable-length position_ids are different code paths." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save_pretrained file emission bug vs performance regression in filtering; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA layernorm epsilon precision issue vs torch.split return handling in GlmMoeDsaIndexer; different components." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Stale auto_mappings config references vs leftover dummy classes; both hygiene issues, not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow backend in torch install vs Python 3.9/3.10 compatibility; unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked LM regression vs Deepseek tokenizer regression; different models and layers." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores vs incorrect model outputs; not the same issue." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor save_pretrained file bug vs gemma-4 zero3 from_pretrained bug; different surfaces." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "GenerationMixin cache/chunked_prefill regression vs CLIPVisionModel hidden_states regression; unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Pipeline docs mismatch vs import slowness; different category entirely." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading failure vs ESM2 model breakage; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Llama4 pad_token_id AttributeError vs glm-4v loading failure; different model-specific bugs." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Qwen2.5-VL import error vs Python version compatibility; no overlap in root cause." + }, + { + "left": "issue:42796", + "right": "issue:43441", + "accept": false, + "reason": "Ministral tokenizer BPE marker handling vs FlashAttention loading failure; not the same bug." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device_map auto device mismatch; unrelated." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "Missing TensorFlow backend vs torch/transformers incompatibility; different compatibility problems." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile vs ANSI-code output issue; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config vs missing fields from AutoConfig with trust_remote_code; different config-loading bugs." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; unrelated runtime 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": 25, + "review_unit_index": 42, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43299", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43784", + "issue:43976", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44451", + "issue:44485", + "issue:44496", + "issue:44779", + "issue:44871", + "issue:45003", + "issue:45072", + "issue:45083", + "issue:45103", + "issue:45292", + "issue:45325", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:46132", + "issue:46291", + "issue:46399", + "issue:46506", + "issue:46614", + "issue:46655", + "issue:46665", + "issue:46726", + "issue:46752", + "issue:46821", + "issue:46829", + "issue:46851" + ], + "soft_pairs": [ + "issue:46291|issue:46399", + "issue:44336|issue:44373", + "issue:45325|issue:46851", + "issue:46655|issue:46726", + "issue:44246|issue:44496", + "issue:41628|issue:43976", + "issue:45083|issue:46851", + "issue:43784|issue:45542", + "issue:44297|issue:45397", + "issue:46399|issue:46851", + "issue:45561|issue:45563", + "issue:45103|issue:46399", + "issue:43526|issue:45003", + "issue:43540|issue:46752", + "issue:45072|issue:46829", + "issue:42491|issue:43299", + "issue:46614|issue:46665", + "issue:43531|issue:45362", + "issue:45588|issue:46851", + "issue:44246|issue:44871", + "issue:44242|issue:45593", + "issue:44295|issue:45397", + "issue:45003|issue:46821", + "issue:44336|issue:44485", + "issue:42886|issue:46132", + "issue:43550|issue:46506", + "issue:44451|issue:46752", + "issue:44779|issue:45292", + "issue:36246|issue:42757" + ], + "prepared_review_unit_hash": "041b409d2b7c0ba93665fa07ca0819aa72f2a49c40559ec8309e7fdac576647f", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36460, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports with a weak Qwen/model-loading overlap; no soft pair looks like the same concrete bug/change, so this cluster should not be merged into a single duplicate bucket.", + "confidence": 0.78, + "canonical_issue_reason": "issue:46851 is the cleanest representative of the strongest recurring theme in the cluster: a Qwen3.5 position_ids / variable-length-inputs regression. It is still not a duplicate of the others, but it is the most central and actionable issue among the set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46851 is the best overall issue to anchor the cluster because it is specific, well-scoped, and closest to the cluster\u2019s only repeated technical motif (Qwen position_ids / model-loading regressions).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both touch imports/kernels, but one is a kernels package version/revision dependency failure and the other is a Qwen3.5 rotary registration crash. Different code paths and fixes." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI codes in loading_report and a wrong docstring for position_ids are unrelated documentation/util issues." + }, + { + "left": "issue:45325", + "right": "issue:46851", + "accept": false, + "reason": "Both mention position_ids, but one is Qwen2.5-VL temporal scaling and the other is Qwen3.5 variable-length inputs. Different models and bugs." + }, + { + "left": "issue:46655", + "right": "issue:46726", + "accept": false, + "reason": "Both are vision/processor issues, but one is a SmolVLM video padding bug and the other is an AutoProcessor missing-backend error." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import-time slowness is unrelated to an unrecognized model/config loading failure." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "One is an AutoImageProcessor import failure and the other is a Python-version compatibility problem; not the same bug." + }, + { + "left": "issue:45083", + "right": "issue:46851", + "accept": false, + "reason": "Both are Qwen-family issues, but they affect different helpers/code paths (_get_feat_extract_output_lengths vs position_ids handling)." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different import/backend errors in different integrations; no shared underlying defect." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save metadata mismatch is unrelated to gemma-4 zero3 from_pretrained behavior." + }, + { + "left": "issue:46399", + "right": "issue:46851", + "accept": false, + "reason": "Qwen3.5 kernel registration crash vs variable-length position_ids crash: different failure sites and fixes." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "A pytest-xdist file race and a stale generate warning are unrelated testing/runtime issues." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Auto-docstring annotation parsing and Qwen3.5 kernel registration are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are different bugs in different subsystems." + }, + { + "left": "issue:43540", + "right": "issue:46752", + "accept": false, + "reason": "Video-input processing in Qwen3OmniMoe and empty-list chat templating are unrelated input-validation failures." + }, + { + "left": "issue:45072", + "right": "issue:46829", + "accept": false, + "reason": "Different model outputs are mismatched in different classes; these are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE loading across versions, but one is LoRA compatibility on hf4.x/hf5.x and the other is Qwen3VL-MoE loading on 5.0.0.dev0. Similar theme, not the same concrete defect." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer backend state and inconsistent from_pretrained behavior inside classes are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Both mention Qwen, but one is sliding_window behavior in Qwen3-MoE and the other is a chat crash in Qwen3.5. Different code paths." + }, + { + "left": "issue:45588", + "right": "issue:46851", + "accept": false, + "reason": "Flash-attention s_aux=None crash is unrelated to Qwen3.5 variable-length position_ids handling." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import latency and eos_token_id config inconsistency are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Load-balancing loss omission and D-FINE auxiliary-loss handling are different model-training bugs." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Position_ids buffer access error and gemma-4 zero3 from_pretrained are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:46821", + "accept": false, + "reason": "Unsafe sys.modules access and TimesFM window_size AttributeError are different defects." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading and AttentionInterface.register behavior changes are separate issues." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and PPDocLayoutV3 CPU/CUDA divergence are not the same bug." + }, + { + "left": "issue:44451", + "right": "issue:46752", + "accept": false, + "reason": "ScandiBERT loading failure and empty chat-template input crash are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Deepseek tokenizer regression and resize_token_embeddings output mismatch are different problems." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Two import errors from different packages and APIs; no shared underlying code-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": 25, + "review_unit_index": 43, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:36246", + "issue:41950", + "issue:42199", + "issue:43257", + "issue:43262", + "issue:43295", + "issue:43475", + "issue:43526", + "issue:43540", + "issue:43596", + "issue:43901", + "issue:44038", + "issue:44162", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44351", + "issue:44466", + "issue:44479", + "issue:44496", + "issue:44623", + "issue:44749", + "issue:44857", + "issue:44877", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45459", + "issue:45561", + "issue:45584", + "issue:45600", + "issue:45759", + "issue:45859", + "issue:46170", + "issue:46369", + "issue:46399", + "issue:46506", + "issue:46655", + "issue:46726", + "issue:46851", + "issue:46897" + ], + "soft_pairs": [ + "issue:43901|issue:45003", + "issue:36246|issue:44351", + "issue:43257|issue:45137", + "issue:41950|issue:44220", + "issue:44496|issue:45003", + "issue:43540|issue:46655", + "issue:45003|issue:45561", + "issue:43262|issue:44246", + "issue:45375|issue:45759", + "issue:44162|issue:45200", + "issue:44479|issue:46655", + "issue:45859|issue:46399", + "issue:44295|issue:44749", + "issue:45137|issue:45600", + "issue:43475|issue:46369", + "issue:43295|issue:45292", + "issue:45292|issue:45584", + "issue:44466|issue:45335", + "issue:44877|issue:45375", + "issue:36246|issue:43596", + "issue:44242|issue:44857", + "issue:44623|issue:46726", + "issue:43526|issue:46897", + "issue:44987|issue:45459", + "issue:44038|issue:46506", + "issue:29870|issue:30333", + "issue:42199|issue:46851", + "issue:44479|issue:46170" + ], + "prepared_review_unit_hash": "2596e20c4a26a81368cea371f7b04b270127de9f028794eae644f11b11c45ea4", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36336, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft pair looks like the same underlying bug/change, so I would keep all issues separate. The set spans unrelated regressions across generation, import paths, video/audio processors, model configs, training, and DeepSpeed behavior.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45561 is the best representative to anchor if needed: it is open, has the most inbound references, and reports a concrete, reproducible race in patched testing utilities.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for TextClassificationPipeline and unsafe sys.modules access are unrelated problems in different code paths." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they refer to different missing symbols and different transformers modules." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed/large-model startup, but one is a Qwen3 MOE weight conversion issue and the other is a ZeRO3 deque IndexError." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification processor behavior and _torch_extract_fbank_features are distinct subsystems with different failures." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unknown-model config loading and sys.modules import safety are different bugs." + }, + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both mention video inputs, but the Qwen3OmniMoe ValueError and SmolVLM padding bug are different code paths." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and pytest-xdist file races are unrelated testing/runtime issues." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sample-rate defaults and import latency are not the same bug." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing strict config field and AutoModelForCausalLM text_config unwrapping are different composite-config problems." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific bugs." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-related regressions, but they affect different models and different failure mechanisms." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "CPU offload incompatibility in rotary embedding and a missing decorator registration crash are not the same issue." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access and multilingual slowdown after upgrade are unrelated symptoms." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "Zero3 deque failure and stale auto_mappings references are different failures." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Sam3VisionEncoderOutput attribute error and SAM3-LiteText memory overflow are distinct bugs." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression and resize_token_embeddings output-embedding update are different code paths." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize behavior and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both concern weight/embedding consistency, but one is lm_head serialization and the other is decoder.embed_tokens resizing." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict config loading and a missing config field are different mechanisms, even though both are config-related." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "A missing Qwen2.5-VL image processor import and a ZeRO3 BertModel IndexError are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing load-balancing loss and an AMP/CUDA crash in LwDetrImageLoss are different training bugs." + }, + { + "left": "issue:44623", + "right": "issue:46726", + "accept": false, + "reason": "processor.save_pretrained missing files and AutoProcessor backend-missing import handling are separate processor issues." + }, + { + "left": "issue:43526", + "right": "issue:46897", + "accept": false, + "reason": "Reduce-labels returning one label and Florence2 double-shift training loss are different label/loss bugs." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Physical-intelligence/fast model loading failure and tokenizer protobuf error masking are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "Qwen3-VL-Moe transformers 5.0 bug and PPDocLayoutV3 CPU/CUDA divergence are different regressions." + }, + { + "left": "issue:29870", + "right": "issue:30333", + "accept": false, + "reason": "Bart generate() UnboundLocalError and MLFlow job-state reporting are unrelated issues." + }, + { + "left": "issue:42199", + "right": "issue:46851", + "accept": false, + "reason": "DETR cardinality error and Qwen3.5 variable-length position_ids are different model-specific failures." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regression in multimodal models and Windows ROCm torch.distributed import errors 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": 25, + "review_unit_index": 44, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:42886", + "issue:43065", + "issue:43352", + "issue:43408", + "issue:43475", + "issue:43526", + "issue:43540", + "issue:43653", + "issue:43701", + "issue:43723", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44877", + "issue:44977", + "issue:44995", + "issue:45103", + "issue:45200", + "issue:45229", + "issue:45250", + "issue:45397", + "issue:45412", + "issue:45563", + "issue:45588", + "issue:45656", + "issue:45685", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46170", + "issue:46202", + "issue:46421", + "issue:46431", + "issue:46710", + "issue:46726", + "issue:46832", + "issue:46858", + "issue:46894" + ], + "soft_pairs": [ + "issue:44060|issue:44077", + "issue:43526|issue:46894", + "issue:44368|issue:44977", + "issue:45412|issue:46123", + "issue:43653|issue:44162", + "issue:43352|issue:45250", + "issue:42886|issue:45229", + "issue:44877|issue:46170", + "issue:43475|issue:43540", + "issue:46170|issue:46832", + "issue:45397|issue:45656", + "issue:43065|issue:43408", + "issue:43723|issue:46710", + "issue:44368|issue:45563", + "issue:44995|issue:45200", + "issue:43931|issue:44368", + "issue:45987|issue:46082", + "issue:44188|issue:46202", + "issue:43906|issue:46202", + "issue:44483|issue:46202", + "issue:44188|issue:46421", + "issue:44995|issue:45588", + "issue:45685|issue:46431", + "issue:45103|issue:46858", + "issue:43065|issue:46123", + "issue:44361|issue:44749", + "issue:44297|issue:44623", + "issue:43994|issue:46726", + "issue:44749|issue:45656", + "issue:43701|issue:44749" + ], + "prepared_review_unit_hash": "7328c05bccba0f61f52cc70ba10c820f17181f4b2cef34a44ac8058ce03a18d0", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36786, + "estimated_input_tokens": 9197, + "estimated_eval_tokens": 18650 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a true duplicate set: it spans many unrelated bugs and regressions across tokenizers, vision models, generation, config loading, cache behavior, and serialization. None of the soft pairs look like the same concrete underlying issue/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are heterogeneous and should be split by subsystem/topic rather than merged as duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global best issue exists here because the soft pairs are mostly superficial matches, not same-bug duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Different models and failure modes: tied-weights warning vs optional post_init restriction." + }, + { + "left": "issue:43526", + "right": "issue:46894", + "accept": false, + "reason": "Different tasks and bugs; one is image processor label reduction, the other ASR training loss shifting." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a config warning and the other is a flash-attention generation regression." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "Unrelated: RT-DETR memory cleanup vs MaskGenerationPipeline final-batch handling." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration bug is not the same as the ESM2 breakage." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "One says Flash Attention 2.0 is unsupported for NemotronH; the other is a generic Flash Attention 2.0 topic." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Tokenizer cache/offline loading bug is unrelated to Gemma multi-GPU OOM." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading vs Windows/ROCm import dependency issue are different problems." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 missing attribute vs Qwen3OmniMoe video-input ValueError are different code paths." + }, + { + "left": "issue:46170", + "right": "issue:46832", + "accept": false, + "reason": "ROCm/Windows import failure is unrelated to RT-DETR forward shape handling." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Different DeepSpeed-related symptoms; from_pretrained loading vs duplicate optimizer step." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Both are SAM3-related, but one is a dummy layer and the other is a model-type mismatch warning." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading in v5 is not the same as an accuracy regression after version upgrade." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Tie_word_embeddings warning is unrelated to stale num_return_sequences warning in paged generate()." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale cache crash is unrelated to Gemma4 mm_token_type_ids defaults." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Weight-shape mismatch in Qwen3-VL is unrelated to Qwen3.5 tie-word-embeddings warnings." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache return-type bug and LlamaConfig head_dim validation are different issues." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "Torch.compile branching bug is unrelated to Gemma roundtrip tokenization failure." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "Insufficient evidence they share the same underlying bug; one is just an isolated reproduction." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat-completions request rejection is unrelated to Gemma token roundtrip issues." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both mention cache/compile, but one is divergent attention kernels and the other is recompilation from lazy cache init." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Different failures: stale indexer cache on second forward vs flash_attention.py AttributeError on s_aux." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MoE MPS histogram error is unrelated to missing chat template in processor." + }, + { + "left": "issue:45103", + "right": "issue:46858", + "accept": false, + "reason": "Auto-docstring annotations crash is unrelated to static-cache seq-length tensor arithmetic." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 pixel-decoder dummy layer and MaskGenerationPipeline batching bug are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError and slow filtering after upgrade are different problems." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizers save_pretrained mismatch is not the same as processor save_pretrained missing files." + }, + { + "left": "issue:43994", + "right": "issue:46726", + "accept": false, + "reason": "Nonsensical outputs with AutoModel/pipeline are different from missing vision-backend error handling." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Transformer filtering slowdown is unrelated to double optimizer stepping under DeepSpeed." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to the filtering performance regression." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 45, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:39692", + "issue:41628", + "issue:42371", + "issue:42994", + "issue:43232", + "issue:43441", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43723", + "issue:43784", + "issue:43873", + "issue:43986", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44373", + "issue:44623", + "issue:44869", + "issue:44871", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45092", + "issue:45137", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45397", + "issue:45406", + "issue:45478", + "issue:45656", + "issue:46132", + "issue:46202", + "issue:46326", + "issue:46399", + "issue:46463", + "issue:46726", + "issue:46752", + "issue:46821", + "issue:46851", + "issue:46877" + ], + "soft_pairs": [ + "issue:43646|issue:44162", + "issue:44623|issue:45656", + "issue:44871|issue:46132", + "issue:43441|issue:46752", + "issue:43475|issue:43643", + "issue:42994|issue:43482", + "issue:45375|issue:45478", + "issue:43475|issue:46821", + "issue:28282|issue:30990", + "issue:44188|issue:46851", + "issue:44964|issue:45137", + "issue:43232|issue:43673", + "issue:39692|issue:46726", + "issue:45310|issue:45375", + "issue:46463|issue:46877", + "issue:41628|issue:44933", + "issue:44869|issue:45356", + "issue:43986|issue:46726", + "issue:44373|issue:46851", + "issue:45230|issue:45397", + "issue:44038|issue:44623", + "issue:42994|issue:43525", + "issue:43723|issue:46326", + "issue:42371|issue:45092", + "issue:43873|issue:44945", + "issue:43784|issue:46399", + "issue:45406|issue:46202" + ], + "prepared_review_unit_hash": "d6008f2aec2093183f403d81a1c6bb3b00088d3b88868ce3b89dc1610f06c8b2", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35430, + "estimated_input_tokens": 8858, + "estimated_eval_tokens": 17972 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; the items span unrelated model-, tokenizer-, processor-, generation-, and training-related bugs. None of the soft pairs look like the same concrete underlying defect, so I would not merge any of them.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Different problems: missing fields from trust_remote_code config loading vs an ESM2 regression/broken workflow." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save/export files missing is unrelated to DeepSpeed optimizer step being called twice." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id mismatch is a model config issue; AttentionInterface.register changes callback behavior in a different subsystem." + }, + { + "left": "issue:43441", + "right": "issue:46752", + "accept": false, + "reason": "FlashAttention failure for Ministral-3 is unrelated to apply_chat_template crashing on an empty list." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "Both mention missing attributes, but they affect different objects and code paths: Sam3 vision encoder output vs AutoConfig with trust_remote_code." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Saving a quantized model and loading Qwen2.5-GGUF are opposite operations with different failure modes." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but one is a missing config field dropped by strict loading and the other is a from_pretrained error; not the same concrete bug." + }, + { + "left": "issue:43475", + "right": "issue:46821", + "accept": false, + "reason": "Sam3 attribute error is unrelated to TimesFM window_size AttributeError." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Missing PyTorch import dependency is unrelated to Sentence Transformers getting stuck while loading." + }, + { + "left": "issue:44188", + "right": "issue:46851", + "accept": false, + "reason": "Torch.compile attention-kernel divergence is a different path from Qwen3.5 variable-length position_ids handling." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Model load failure for Phi-4 multimodal is unrelated to a DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both are generation-related, but one is sync_gpus/update_model_kwargs and the other is missing GenerationMixin cache in chunked prefill." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "Documentation example errors are not the same as AutoProcessor backend-missing runtime handling." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both involve Qwen3.5-Moe, but one is a generic from_pretrained error and the other is a missing deepstack_visual_indexes field silently dropped by strict config handling." + }, + { + "left": "issue:46463", + "right": "issue:46877", + "accept": false, + "reason": "Incorrect T5 decoder-input validation messaging is unrelated to auto_docstring printing lint diagnostics at import time." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Different import failures: AutoImageProcessor export missing vs a non-existent image_utils import." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding crash on a trailing replacement character is unrelated to Kimi-K2.5 tokenizer codec regression." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both involve AutoProcessor and vision backends, but one is missing torchvision for video models while the other is unrecognized image processor handling." + }, + { + "left": "issue:44373", + "right": "issue:46851", + "accept": false, + "reason": "A wrong docstring is not the same bug as a position_ids runtime failure." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report has no identifiable overlap with gemma-4 ZeRO3 from_pretrained failure." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe model loading is unrelated to missing files from processor.save_pretrained." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failure is unrelated to Llama4Config missing pad_token_id." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "AutoTokenizer loading in v5 and continue_final_message for mistral_common tokenizers are different tokenizer paths and failures." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 settings guidance is unrelated to remote-code checkpoint meta-initialization incompatibility." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior is a different training/runtime issue than incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Sentence-transformers import NameError is unrelated to Qwen3.5 kernel registration failure." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer during serve is not the same as Gemma pipeline roundtrip failure on \u2581." + } + ] + }, + "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": 25, + "review_unit_index": 46, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43479", + "issue:43504", + "issue:43704", + "issue:43723", + "issue:43824", + "issue:43976", + "issue:44038", + "issue:44222", + "issue:44351", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44792", + "issue:44857", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45405", + "issue:45561", + "issue:45593", + "issue:45600", + "issue:45712", + "issue:45753", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46693", + "issue:46701", + "issue:46829", + "issue:46877" + ], + "soft_pairs": [ + "issue:44857|issue:46877", + "issue:44945|issue:45753", + "issue:41628|issue:44351", + "issue:45081|issue:45103", + "issue:45593|issue:46877", + "issue:36010|issue:41628", + "issue:46202|issue:46701", + "issue:43278|issue:45137", + "issue:44038|issue:45310", + "issue:42673|issue:43704", + "issue:42994|issue:43824", + "issue:44222|issue:45310", + "issue:42831|issue:44222", + "issue:43232|issue:44792", + "issue:43504|issue:45561", + "issue:45561|issue:45600", + "issue:44496|issue:45561", + "issue:42175|issue:44933", + "issue:45405|issue:46877", + "issue:30333|issue:33290", + "issue:43976|issue:44933", + "issue:30990|issue:35141", + "issue:42898|issue:44568", + "issue:42175|issue:45712", + "issue:43479|issue:46829", + "issue:44623|issue:45310", + "issue:46077|issue:46123", + "issue:42673|issue:46693", + "issue:42898|issue:45812", + "issue:43479|issue:44869", + "issue:43723|issue:44568" + ], + "prepared_review_unit_hash": "4c314dda9b0bae2d60aa2508399ff397a815c054dd2a7111a682fbc7a9685d27", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36868, + "estimated_input_tokens": 9217, + "estimated_eval_tokens": 18690 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly unrelated issue reports. The only clear duplicate pair is the repeated Qwen3ForCausalLM VRAM-leak report (issue:42673 and issue:43704), which share the exact same title and underlying symptom. All other soft edges are broad-theme similarities at best, not the same concrete bug.", + "confidence": 0.97, + "canonical_issue_reason": "issue:42673 is the best canonical representative because it has the exact same title as issue:43704 and is the earlier, fuller report with more discussion.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the strongest issue to keep as the representative target because it is the earliest full report of the same Qwen3ForCausalLM VRAM-leak bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44857", + "right": "issue:46877", + "accept": false, + "reason": "Different bugs: AMP/CUDA loss crash vs auto_docstring import-time stdout logging." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Different failure modes: pipeline-parallel output corruption vs a specific multimodal infinite loop." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different code paths." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Different subsystems: tokenizer regex patching vs auto_docstring annotation handling." + }, + { + "left": "issue:45593", + "right": "issue:46877", + "accept": false, + "reason": "Missing auxiliary losses is unrelated to docstring diagnostics printed at import." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Unrelated import problems: GenerationMixin vs AutoImageProcessor." + }, + { + "left": "issue:46202", + "right": "issue:46701", + "accept": false, + "reason": "Both mention Gemma, but the concrete failures are different: token roundtrip vs pipeline integration." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Different bugs: dtype drift in evaluation vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Related model family, but different versions/models and no evidence of the same concrete failure path." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same title and same VRAM-leak symptom; clear duplicate reports of the same bug." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized save failure vs missing exported class import are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Different operations and likely different fixes: FP8 save_pretrained vs from_pretrained load error." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both mention FP8, but one is an accuracy regression and the other is a save/pretrained bug." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "A generation-kwargs bug and a failed Janus test are not the same concrete issue." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Pretrained preset loading vs xdist race in testing utilities are unrelated." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "A test-runner race on captured_info.txt is unrelated to stale auto_mappings references." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config loading error vs testing-utils file race; no shared code path." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Package/backend dependency issue vs missing image_utils import; unrelated." + }, + { + "left": "issue:45405", + "right": "issue:46877", + "accept": false, + "reason": "Release/version pinning issue vs docstring logging problem; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLFlow reporting failure and DeepSpeed OOM are different training issues." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version incompatibility vs missing import from image_utils; not the same bug." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Loading hang and embedding reinitialization are unrelated failures." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different behaviors and models." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Different dependency/packaging problems: missing TensorFlow backend vs dummy classes leaking into namespace." + }, + { + "left": "issue:43479", + "right": "issue:46829", + "accept": false, + "reason": "Different model/config bugs with different outputs and symptoms." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained missing files is a different path than Qwen3.5 MoE from_pretrained loading." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Both touch generation, but one is attention-mask construction and the other is batch result dropping." + }, + { + "left": "issue:42673", + "right": "issue:46693", + "accept": false, + "reason": "Both involve GPU performance, but the concrete bugs differ: VRAM leak in threads vs FA2 long-context regression." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different APIs and models." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Different model-family issues: multimodal config defaults vs Whisper timestamp decoding crash." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Both are v5 tokenizer issues, but one is tokenizer loading and the other is add_special_tokens behavior." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The main duplicate call is plausible, but the canonicalization rationale overstates the packet: issue:42673 and issue:43704 do have the same title and the earlier one is fuller/more discussed, but they do not show more inbound references (both are 0 in the packet). For a precision review, that unsupported detail is enough to reject." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 47, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:34689", + "issue:36331", + "issue:39692", + "issue:42222", + "issue:42796", + "issue:42831", + "issue:42898", + "issue:43116", + "issue:43295", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44190", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44448", + "issue:44623", + "issue:44625", + "issue:44734", + "issue:44779", + "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:46877", + "issue:46909" + ], + "soft_pairs": [ + "issue:44625|issue:46877", + "issue:44016|issue:46231", + "issue:39692|issue:46909", + "issue:42898|issue:44779", + "issue:43479|issue:44079", + "issue:43784|issue:43824", + "issue:44734|issue:45103", + "issue:42222|issue:45200", + "issue:43874|issue:45092", + "issue:27957|issue:30333", + "issue:42898|issue:43295", + "issue:30333|issue:34689", + "issue:42898|issue:45356", + "issue:44295|issue:45656", + "issue:30333|issue:36331", + "issue:45405|issue:45712", + "issue:43116|issue:44190", + "issue:22355|issue:28282", + "issue:44222|issue:44297", + "issue:44075|issue:46231", + "issue:44060|issue:45125", + "issue:44933|issue:45712", + "issue:44448|issue:46132", + "issue:42796|issue:43723", + "issue:42831|issue:44623", + "issue:44297|issue:45656", + "issue:43479|issue:43644", + "issue:44222|issue:44623", + "issue:45245|issue:46132", + "issue:44360|issue:46762" + ], + "prepared_review_unit_hash": "9b647d0f6fd521e66f092127df6f59a61784610b50dc9e47078ad30602798331", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36687, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issues with superficial title similarity. I did not find any soft pair that clearly describes the same concrete bug/change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 44779 is the most representative user-facing regression in the set: a concrete tokenizer breakage in v5 with clear reproduction impact and multiple references. It is broader and more substantive than the mostly narrow doc/import/config issues around it.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44779 is the best overall issue candidate because it is a concrete, high-signal regression with broader applicability than the other items, which are mostly model-specific, documentation-only, or isolated API/import errors.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:46877", + "accept": false, + "reason": "Different problems: Qwen3.5 num_labels propagation vs auto_docstring logging diagnostics to stdout." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "A notebook syntax error and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:46909", + "accept": false, + "reason": "Both are broken examples, but they affect different docs and different model flows; not the same bug." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both involve tokenizer regressions in v5, but one is a specific clean_up_tokenization_spaces behavior change and the other is Deepseek output corruption; too indirect to treat as the same bug." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Different code paths: multimodal config initialization vs ModelOutput key assignment." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and different packages/models." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache indexing in serve is unrelated to the auto_docstring annotations crash." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different multimodal issues affecting different models and different inputs." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Both touch multimodal model loading, but one is a missing image-processor method and the other is remote-code/meta-init incompatibility; not the same fix." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention model support and MLflow reporting on training failure are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Different regressions: tokenizer cleanup behavior vs processor/tokenizer image-passing compatibility." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Failed MLflow reporting and Llama model loading regression are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they appear model/version specific and not the same code-path problem." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Reading a registered buffer and duplicate optimizer stepping under deepspeed are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow failure reporting and CustomTrainer signature mismatch are different bugs." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "A dependency version bump issue and leftover dummy classes are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Both involve example scripts, but they fail in different scripts for different reasons." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Different missing-dependency import errors: transformers.onnx vs torch for AutoModel." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained for MoE and tokenizer_class metadata mismatch are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Unused SGD args is a runtime optimizer issue, not a documentation typo." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both are Qwen-family issues, but one is a tied-weights warning and the other is a missing tensor-parallel plan." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Importing a nonexistent image_utils symbol and dummy-class leakage are different maintenance issues." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus v4/v5 output drift and AttentionInterface.register behavior change are not the same bug." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "A model tokenizer BPE-marker regression and AutoTokenizer loading failure are different tokenizer problems." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy degradation and processor.save_pretrained missing files are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata and deepspeed optimizer-step duplication are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Default config initialization and non-persistent buffer corruption are different subsystems." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "MoE FP8 save behavior and processor serialization missing files are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category cardinality error and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:46762", + "accept": false, + "reason": "A missing ReLU in one implementation and numerical differences in a different attention backend are not the same concrete 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": 25, + "review_unit_index": 48, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42907", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43723", + "issue:44190", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45834", + "issue:46050", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46369", + "issue:46424", + "issue:46431", + "issue:46614", + "issue:46655", + "issue:46710", + "issue:46909", + "issue:46932" + ], + "soft_pairs": [ + "issue:46132|issue:46710", + "issue:44373|issue:44485", + "issue:41628|issue:42757", + "issue:44297|issue:45310", + "issue:43723|issue:44977", + "issue:45071|issue:46082", + "issue:46614|issue:46932", + "issue:44222|issue:44484", + "issue:44190|issue:46082", + "issue:43122|issue:46209", + "issue:44263|issue:45092", + "issue:43232|issue:46424", + "issue:43232|issue:45834", + "issue:43475|issue:46305", + "issue:30990|issue:38175", + "issue:46209|issue:46431", + "issue:46305|issue:46655", + "issue:44295|issue:44623", + "issue:44263|issue:44360", + "issue:44263|issue:44373", + "issue:42907|issue:46050", + "issue:45092|issue:45357", + "issue:46369|issue:46909", + "issue:42831|issue:45310", + "issue:45381|issue:46305", + "issue:44279|issue:46328", + "issue:44483|issue:45406", + "issue:22355|issue:27957", + "issue:44261|issue:45092" + ], + "prepared_review_unit_hash": "13a97349b6146ba78868b4f27d84fc6198fd5924c093363fefc317816686cd91", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35353, + "estimated_input_tokens": 8839, + "estimated_eval_tokens": 17934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is largely heterogeneous; the soft pairs share only broad topical overlap, not the same concrete bug or change. I rejected all soft edges because none look mergeable as duplicates. I picked 45092 as the closest issue representative of the loose v5-compatibility subset, but there is no strong canonical for the full cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single strong canonical: the set is heterogeneous. If one must be chosen, 45092 is the closest representative because it is a concrete v5 regression with a clear compatibility failure.", + "canonical_pr_reason": null, + "best_issue_reason": "45092 is the best representative for the same reason: it is specific, regression-oriented, and more likely to anchor related compatibility reports than the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46132", + "right": "issue:46710", + "accept": false, + "reason": "Different bugs: attention registration behavior vs an accuracy regression in a specific model." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring wording issue vs RoPE implementation bug; not the same code-path failure." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Separate import errors from different packages/modules and different root causes." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata mismatch vs a Qwen3.5 from_pretrained load error; related area but distinct defect." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading in v5 vs flash-attention generation for Qwen3.5; unrelated failures." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Both are config-related, but one is type-checking regression and the other is LlamaConfig head_dim validation." + }, + { + "left": "issue:46614", + "right": "issue:46932", + "accept": false, + "reason": "Phonemizer backend state bug vs Gemma4 quantized embedder dtype casting; different components." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained/moe issue vs a max_shard_size default question; no shared bug." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Dataset loading script failure vs model config validation; unrelated code paths." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization drift across versions vs FP8 inference on B200; not the same underlying problem." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "DSA indexer split behavior vs meta-initialization incompatibility for old checkpoints; distinct issues." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update bug vs cache max length semantics; same broad area, different failure." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation-state update bug vs a long-OCR index error in Kosmos2.5; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM3 video encoder output attribute issue vs MoonshineStreaming mask off-by-one; different models and logic." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Loading hang vs zero-probability inference bug; not the same defect." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 B200 inference failure vs missing chat template error; unrelated." + }, + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-related indexing/padding bugs, but in different processors and different computations." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer access error vs processor.save_pretrained missing files; different save/load paths." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same component family, but one is a torch.split issue and the other is a missing ReLU; not the same bug." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Indexer return-value bug vs a wrong docstring; no shared failure mode." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Different quantization/save issues affecting different flows and serialization behaviors." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both touch multimodal checkpoint compatibility, but one is meta-init loading and the other is incorrect saved visual encoder keys." + }, + { + "left": "issue:46369", + "right": "issue:46909", + "accept": false, + "reason": "Model memory overflow vs documentation example bug; unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy regression vs from_pretrained load error; distinct issues despite both being model-related." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Both mention off-by-one style behavior, but they are different models, tensors, and code paths." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Dependency problem vs pad-performance inefficiency; not the same bug." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat-completions request handling bug vs Gemma4Processor attribute crash; different API failures." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "ONNX import error vs Flash Attention 2 support question; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "RMS norm epsilon precision bug vs old checkpoint meta-init incompatibility; different model behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 49, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43540", + "issue:43576", + "issue:43644", + "issue:43646", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44188", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44488", + "issue:44568", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45278", + "issue:45381", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46305", + "issue:46369", + "issue:46402", + "issue:46909" + ], + "soft_pairs": [ + "issue:44829|issue:46002", + "issue:43644|issue:46305", + "issue:42831|issue:44038", + "issue:41628|issue:43824", + "issue:43408|issue:46402", + "issue:22355|issue:28218", + "issue:42831|issue:45753", + "issue:39401|issue:44568", + "issue:42886|issue:42994", + "issue:42886|issue:43066", + "issue:27957|issue:30990", + "issue:44188|issue:44373", + "issue:42757|issue:45278", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:42796|issue:43122", + "issue:45381|issue:46909", + "issue:43646|issue:46002", + "issue:43540|issue:46909", + "issue:30990|issue:33357", + "issue:42907|issue:44945", + "issue:42907|issue:44964", + "issue:43976|issue:44945", + "issue:39401|issue:44488", + "issue:42371|issue:44263", + "issue:44295|issue:44297", + "issue:43756|issue:44060", + "issue:43576|issue:43784", + "issue:41628|issue:45278" + ], + "prepared_review_unit_hash": "ecacb1bece2382ab78f15ec9a4f13bb73fc2f04fc79af40c0560682b12ea2f3a", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35259, + "estimated_input_tokens": 8815, + "estimated_eval_tokens": 17886 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly broad-topic similarities rather than true duplicates; I would reject all of them. If one issue must serve as the cluster representative, the broad version-upgrade import regression is the best umbrella candidate.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the most umbrella-like report in this set: it describes many import errors after a transformers upgrade, making it the best representative for a heterogeneous import-regression cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the strongest global issue candidate because it is broad, clearly user-facing, and closer to a cluster-level regression than the more model-specific or one-off bug reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Both are training/initialization bugs, but they describe different failure modes: Flash Attention 3 degenerate training vs NaNs after checkout+init." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "One is a v5 non-persistent-buffer corruption issue; the other is a MoonshineStreaming sliding-window off-by-one. Different code paths and symptoms." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FineGrainedFP8 accuracy problems and a Qwen3-VL-Moe bug are unrelated beyond both involving model quality/regression concerns." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and likely different root causes; not the same concrete bug." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Different model families and different problems: a wrong model-type warning vs missing Sam3 text encoder weights." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "A missing module import and a tokenizer extra-space bug are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "Accuracy regression in FP8 is not the same as an infinite loop on a specific image." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenization behavior, but one is offset mapping for Qwen3 and the other is BOS/EOS insertion for mDeBERTa; different bugs." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline cache loading failure and quantized model saving failure are different lifecycle stages and code paths." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading and wrong tokenizer decoder type are unrelated issues." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta and a Sentence Transformers loading hang are different problems." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "A torch.compile attention-kernel divergence bug is unrelated to a docstring issue." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Both mention import breakage, but one is a specific huggingface_hub symbol and the other is a broad transformers upgrade regression; not enough to call them the same bug." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "Model-type warning and memory overflow are different failure modes." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer BPE-marker handling for one model and a general tokenization difference across versions are not clearly the same bug." + }, + { + "left": "issue:45381", + "right": "issue:46909", + "accept": false, + "reason": "A Qwen2.5-VL video-position-id bug and a broken documentation example for DIA are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads both touch init, but they are not the same concrete failure." + }, + { + "left": "issue:43540", + "right": "issue:46909", + "accept": false, + "reason": "Video input ValueError in Qwen3OmniMoe is unrelated to a DIA docs example." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang for Sentence Transformers and a MacOS bus error with a CLIP model are distinct issues." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized model saving failure and incorrect pipeline-parallel outputs are different bugs." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Quantized model saving failure and inability to load Phi-4 multimodal are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility regression and pipeline parallelism output corruption are different problems." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer offset mapping for Qwen3 and failure to load a specific model are not the same underlying bug." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration guidance and torch.split return-value handling are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are different tokenizer/model serialization bugs." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "RoPE layer dropping in Smollm3 and a tied-weights warning in Qwen3-Next are different model-definition issues." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "A broken CLI command and a NameError during import are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Both concern imports after upgrades, but the specific missing symbol and the broad multi-import regression are too different to merge." + } + ] + }, + "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": 25, + "review_unit_index": 50, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:36296", + "issue:39401", + "issue:42175", + "issue:42757", + "issue:42898", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:44038", + "issue:44117", + "issue:44188", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "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:46231", + "issue:46326", + "issue:46328", + "issue:46614", + "issue:46762", + "issue:46877" + ], + "soft_pairs": [ + "issue:46172|issue:46762", + "issue:42907|issue:43937", + "issue:43784|issue:45600", + "issue:39401|issue:44521", + "issue:44117|issue:44704", + "issue:44625|issue:44829", + "issue:43408|issue:43475", + "issue:44704|issue:45290", + "issue:43122|issue:44521", + "issue:44360|issue:44485", + "issue:43122|issue:46326", + "issue:42907|issue:43452", + "issue:44038|issue:44297", + "issue:22355|issue:29159", + "issue:45216|issue:45478", + "issue:44038|issue:44295", + "issue:39401|issue:46614", + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:43116|issue:45200", + "issue:43122|issue:44779", + "issue:43278|issue:44945", + "issue:36296|issue:43116", + "issue:42757|issue:44351", + "issue:22355|issue:29127", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:44188|issue:46328", + "issue:46231|issue:46877", + "issue:42898|issue:46614" + ], + "prepared_review_unit_hash": "82ef5647af3de684fed8e06aad34ac96dbb2d5b70f5a66484f8eedb5350b5f28", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36475, + "estimated_input_tokens": 9119, + "estimated_eval_tokens": 18494 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly share only broad subsystem terms (tokenizers, save/load, multimodal, import errors) but not the same concrete bug. I found no duplicate-worthy pair and no PRs.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone issue to anchor on operationally: it is open, has multiple inbound references, and describes a concrete user-facing regression in multimodal chat templating. It is not, however, a duplicate target for the other items here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "Different model families and different symptoms: deepseek-v4 tensor conversion bug vs Minimax MSA numerical drift." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Unrelated failures: save_pretrained for dequantized models vs invalid GenerationConfig for GLM-5." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Importing sentence-transformers breaks due to missing nn, while the other is a CI break from removed Sam3LiteText config refs." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Different code paths: tokenizer offset mapping for Qwen3 vs assistant_masks generation in multimodal chat templates." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES None handling is a tokenizer registry bug; AutoProcessor kwargs forwarding is a cached_file plumbing issue." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are unrelated bugs." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Both are SAM 3 related, but one is a model-type warning and the other is a missing output attribute; different failures." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and apply_chat_template tool-call crash affect different paths and would not merge into one fix." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer-version regression is not the same bug as multimodal assistant_mask generation." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer activation discussion and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both involve tokenizer behavior, but they describe different regressions in different APIs and code paths." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Save/pretrained failure for dequantized models is unrelated to gguf_file load issues in AutoTokenizer/AutoModel." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe bug and tokenizer_class metadata mismatch are different problems." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Missing transformers.onnx module import and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a bad saved checkpoint while the other is a from_pretrained import error in newer transformers." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug is unrelated to the position_ids buffer access error." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Qwen3 tokenizer offset mapping and stale phonemizer language state are different tokenizer subsystems with different root causes." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe issue is not the same as FP8 save_pretrained for MoE models." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Different missing exports: Qwen2_5_VLForConditionalGeneration vs HybridCache." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script multi-label inference returning empty results is not the same bug as mm_token_type_ids defaults for Gemma 4 fine-tuning." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "General tokenizer regression and Deepseek-specific incorrect tokenization are not clearly the same underlying defect." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift between train/eval is unrelated to incorrect pipeline-parallel LLM outputs." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and example-script classification emptiness are different failure modes." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they refer to different symbols from different packages and different breakages." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Transformers.onnx import failure and LayoutLMv3 error messaging are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Package extras missing tensorflow_backend and Python 3.14 import/load failure are unrelated environment issues." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 numerical discrepancy and Qwen3_5 image-triggered infinite loop are different model-specific bugs." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "torch.compile attention-kernel divergence and SequenceFeatureExtractor.pad performance regression are unrelated." + }, + { + "left": "issue:46231", + "right": "issue:46877", + "accept": false, + "reason": "Documentation mismatch in set_training and auto_docstring stdout diagnostics are separate doc-tooling issues." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer cleanup behavior change and stale phonemizer language state do not share a concrete bug 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": 25, + "review_unit_index": 51, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28348", + "issue:38617", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42796", + "issue:43066", + "issue:43116", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43618", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44263", + "issue:44355", + "issue:44485", + "issue:44610", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44995", + "issue:45072", + "issue:45083", + "issue:45092", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:46002", + "issue:46231", + "issue:46614", + "issue:46640", + "issue:46697", + "issue:46710", + "issue:46726", + "issue:46752", + "issue:46762", + "issue:46829", + "issue:46909", + "issue:46981" + ], + "soft_pairs": [ + "issue:38617|issue:45278", + "issue:43937|issue:44485", + "issue:43066|issue:44779", + "issue:43611|issue:46002", + "issue:42673|issue:45072", + "issue:43976|issue:45446", + "issue:44038|issue:45478", + "issue:43116|issue:43493", + "issue:45357|issue:45478", + "issue:44355|issue:46231", + "issue:44625|issue:46002", + "issue:42371|issue:43874", + "issue:44610|issue:46909", + "issue:45083|issue:45446", + "issue:46726|issue:46981", + "issue:44075|issue:44936", + "issue:43618|issue:46829", + "issue:43116|issue:43825", + "issue:43066|issue:46710", + "issue:43992|issue:44704", + "issue:46614|issue:46752", + "issue:44995|issue:46697", + "issue:44263|issue:46762", + "issue:28348|issue:46640", + "issue:45092|issue:46762", + "issue:42673|issue:42796", + "issue:43827|issue:45446", + "issue:42222|issue:43441" + ], + "prepared_review_unit_hash": "2d937999e793d0b13da86904f910d16f7dce735aadd704c33bd008d49077a2cd", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35794, + "estimated_input_tokens": 8949, + "estimated_eval_tokens": 18154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items mostly look like separate regression reports across different models, tokenizer paths, and runtime behaviors. None of the soft pairs appear to be the same underlying bug/change, so all soft edges are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45278 is the broadest, most umbrella-style report in the set (\u201cmany import errors after update\u201d), so it is the best single issue to anchor the cluster even though most other items are not duplicates of it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the clearest cluster representative: general regression wording, broad scope, and strongest fit as an umbrella issue among otherwise unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but one is a specific missing symbol and the other is a broad upgrade complaint; not the same concrete bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Different GLM-5 failures: invalid generation config vs RoPE implementation concerns." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different models and likely different code paths." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Loading via base_model_prefix and NaNs in classification heads are unrelated bugs." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multithreaded dataloading is unrelated to bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Different environment/version-check issues in separate code paths." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen MoE, but they describe different failures and model variants." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label behavior and SigLIP2 implementation discrepancy are not the same bug." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "save_pretrained key mismatch and from_pretrained loading error are different problems." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled-file execution errors and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation is a config wiring bug, not the same as NaNs in heads." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and GLM46V image-patch helper absence are unrelated." + }, + { + "left": "issue:44610", + "right": "issue:46909", + "accept": false, + "reason": "Processor image-size mismatch and a broken documentation example are different issues." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Feature-extraction length helper behavior and a PyTorch import version check are unrelated." + }, + { + "left": "issue:46726", + "right": "issue:46981", + "accept": false, + "reason": "Missing vision backend handling and HTTP 429 retry behavior in AutoProcessor are different failure modes." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer argument handling and trainer evaluate/train interaction are separate bugs." + }, + { + "left": "issue:43618", + "right": "issue:46829", + "accept": false, + "reason": "Both affect model outputs, but they involve different output fields and different models." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example-script classification behavior and pipeline error messaging are not the same code-path problem." + }, + { + "left": "issue:43066", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer decoder-type bug and DeepSeek accuracy regression are too different to merge." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "A missing weight in UMT5 loading is unrelated to AutoProcessor kwargs forwarding." + }, + { + "left": "issue:46614", + "right": "issue:46752", + "accept": false, + "reason": "Tokenizer phonemizer backend state and empty-chat-template handling are unrelated." + }, + { + "left": "issue:44995", + "right": "issue:46697", + "accept": false, + "reason": "GLM-MoE stale cache crash and Swin2SR over-padding are unrelated model/component bugs." + }, + { + "left": "issue:44263", + "right": "issue:46762", + "accept": false, + "reason": "GlmMoeDsa torch.split handling and Minimax attention numerical differences are different problems." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "A flash-attention feature request for GPT-2 is not the same as a flash-attention bug report." + }, + { + "left": "issue:45092", + "right": "issue:46762", + "accept": false, + "reason": "Meta-init incompatibility for InternVL2 and Minimax attention numerics are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM and Ministral tokenizer marker handling are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still mentioning pipeline() and a flex_attention import version check are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Vitpose breakage and Ministral FlashAttention failure are different model-specific 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": 25, + "review_unit_index": 52, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:33453", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42617", + "issue:42673", + "issue:43064", + "issue:43122", + "issue:43408", + "issue:43450", + "issue:43673", + "issue:43742", + "issue:43824", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44291", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44484", + "issue:44779", + "issue:44829", + "issue:44877", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45405", + "issue:45446", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:45632", + "issue:46002", + "issue:46077", + "issue:46178", + "issue:46231", + "issue:46298", + "issue:46328", + "issue:46512", + "issue:46655", + "issue:46752", + "issue:46981" + ], + "soft_pairs": [ + "issue:33453|issue:39401", + "issue:43064|issue:43673", + "issue:43937|issue:45563", + "issue:43824|issue:46981", + "issue:43673|issue:46328", + "issue:42673|issue:46655", + "issue:42222|issue:46752", + "issue:44291|issue:45071", + "issue:45405|issue:45446", + "issue:44117|issue:45588", + "issue:44380|issue:46077", + "issue:44466|issue:46002", + "issue:43937|issue:44977", + "issue:39401|issue:44779", + "issue:44410|issue:45310", + "issue:45071|issue:46512", + "issue:43450|issue:43986", + "issue:45632|issue:46981", + "issue:45341|issue:45561", + "issue:45341|issue:46328", + "issue:44829|issue:45216", + "issue:44117|issue:45081", + "issue:42617|issue:45341", + "issue:43408|issue:43742", + "issue:43122|issue:46298", + "issue:38617|issue:41628", + "issue:45362|issue:46328", + "issue:42222|issue:46178", + "issue:44877|issue:46231", + "issue:44484|issue:45310" + ], + "prepared_review_unit_hash": "599915ded169037d6984fb02b2166d65841521f36c280c3924cbdebc6729a9b9", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36669, + "estimated_input_tokens": 9168, + "estimated_eval_tokens": 18592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, config/import, processor, training, and logging issues. The soft edges are superficial title similarity only; none appear to be the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits this cluster because the items span multiple unrelated failure modes and code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global representative; if forced, 43122 is the broadest tokenizer-regression report, but it still does not unify the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both mention tokenizer problems, but one is generic loading regression and the other is a Qwen3 offset_mapping bug; different code paths and symptoms." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "One is FSDP/PEFT optimizer-state corruption, the other is a generation cache issue in chunked prefill; not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GLM generation config validation and a stale paged-generate warning are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:46981", + "accept": false, + "reason": "Both involve loading/auto APIs, but one is an import error for a specific model class and the other is repeated HTTP 429 retries in AutoProcessor." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Both touch generation/padding-ish utilities, but one is cache handling and the other is a performance issue in SequenceFeatureExtractor.pad." + }, + { + "left": "issue:42673", + "right": "issue:46655", + "accept": false, + "reason": "VRAM leakage in Qwen3ForCausalLM across dataloader threads is unrelated to a video processor padding bug." + }, + { + "left": "issue:42222", + "right": "issue:46752", + "accept": false, + "reason": "Vitpose model breakage and apply_chat_template empty-list IndexError are different subsystems and failure modes." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights argument mismatch and PreTrainedConfig type checking regression are separate v5 API issues." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "One is a PEFT version pin problem; the other is an incorrect PyTorch version check for flex_attention." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES None handling and flash_attention crash on s_aux=None are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention and Gemma4 bidirectional-mask construction affect different models and logic." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization of lm_head.weight/tied weights is distinct from NaNs in classification heads after checkout/init." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation-related, but one is invalid GenerationConfig handling and the other is Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but the first is offset mapping and the second is Deepseek tokenization output; different concrete defects." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing qkv/gate projections in qwen3next layer 0 is unrelated to Qwen3.5 Moe from_pretrained loading errors." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Type checking regression in PreTrainedConfig is not the same as stale dtype metadata after from_config(dtype=...)." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Video processor shape bug and AutoProcessor crash without torchvision are both video-related but are different code paths and root causes." + }, + { + "left": "issue:45632", + "right": "issue:46981", + "accept": false, + "reason": "trust_remote_code cache-path collision is unrelated to repeated retry behavior on HTTP 429 responses." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "A bug in testing_utils.py is not the same as pytest-xdist workers racing on captured_info.txt." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing-utils bug and SequenceFeatureExtractor.pad performance issue do not share a concrete underlying change." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention-3 degenerate training in sequence classification is unrelated to Qwen3.5 checkpoint save_pretrained regression." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Mapping-name fallback and Mistral regex patch crash are separate tokenizer-loading failures." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Running 3d_parallel.py and a testing_utils.py bug are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "sam3_video/sam3_tracker config mismatch is distinct from a key error loading facebook/MobileLLM-125M." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenization differences across versions and a noisy logging message are not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they involve different symbols and modules (`layer_type_validation` vs `AutoImageProcessor`)." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Qwen3.5 chat crash and SequenceFeatureExtractor.pad performance issue are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "Vitpose breakage and a BACKENDS_MAPPING tensorflow_text error are different loading/backend issues." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config preventing granite_speech loading is unrelated to a docs/parameter-name mismatch in set_training()." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "save_pretrained shard-size question and Qwen3.5 Moe from_pretrained 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": 25, + "review_unit_index": 53, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43208", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43577", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43937", + "issue:44038", + "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:44829", + "issue:44933", + "issue:45003", + "issue:45071", + "issue:45278", + "issue:45341", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45656", + "issue:46172", + "issue:46209", + "issue:46617", + "issue:46762", + "issue:46894", + "issue:46981" + ], + "soft_pairs": [ + "issue:39401|issue:46209", + "issue:42371|issue:44261", + "issue:43937|issue:44496", + "issue:44261|issue:46762", + "issue:42371|issue:46617", + "issue:44410|issue:45478", + "issue:44933|issue:45278", + "issue:44075|issue:45656", + "issue:38617|issue:44220", + "issue:43208|issue:46894", + "issue:43066|issue:43122", + "issue:44263|issue:44485", + "issue:43097|issue:44493", + "issue:44483|issue:45464", + "issue:43673|issue:45341", + "issue:42222|issue:45542", + "issue:44077|issue:46981", + "issue:43931|issue:44410", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:43874|issue:44261", + "issue:44484|issue:45478", + "issue:44263|issue:46172", + "issue:43931|issue:44038", + "issue:44360|issue:46172", + "issue:27957|issue:43352", + "issue:43441|issue:44829", + "issue:43012|issue:43577", + "issue:43097|issue:43452", + "issue:45003|issue:45341" + ], + "prepared_review_unit_hash": "e19bdb7bc7cc15ce84d93efcfdc9638f742e0e9851e438ca9976ce7952d93631", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36157, + "estimated_input_tokens": 9040, + "estimated_eval_tokens": 18336 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: many issues share broad vocabulary around imports, tokenization, model loading, or attention/precision, but they do not look like the same underlying bug. I would not merge any of the soft-edge pairs; the strongest representative is a broad import-regression report.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella here: it reports multiple import errors after a Transformers upgrade, so it best represents the cluster\u2019s general regression theme.", + "canonical_pr_reason": null, + "best_issue_reason": "45278 is the most suitable triage representative because it is the most general and could subsume narrower import-related reports, even though it is not a true duplicate of most others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset-mapping bug vs. FP8 inference failure are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "Both are precision-related, but one is TF32 policy guidance and the other is a model-specific RMSNorm epsilon bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig vs. unrecognized model/config loading are different failure modes." + }, + { + "left": "issue:44261", + "right": "issue:46762", + "accept": false, + "reason": "Both concern numerical differences, but they affect different models and code paths." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 behavior settings and cache FileNotFoundError are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen model families and different loading/projection problems." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import is not the same as a broad upgrade-wide import-regression report." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "One is unused SGD arguments; the other is double optimizer stepping under deepspeed." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for a utility symbol and a fbank feature issue are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:46894", + "accept": false, + "reason": "Different model/training bugs; only the vague word \"training\" overlaps." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer decoder type mismatch and different tokenization output are related thematically but not the same concrete bug." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Moe indexer torch.split semantics and GLM-5 RoPE implementation are different code paths." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed config API and unexpected position-id keys are different regressions." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a general endpoint acceptance issue and the other is a model-specific streaming failure." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache behavior and a testing_utils bug are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and TensorFlow backend detection are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:46981", + "accept": false, + "reason": "PatchTSMixer post_init validation and repeated 429 retries are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight-shape mismatch and missing projections are different model-loading failures." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression are unrelated." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type-checking breakage and missing TensorFlow backend handling are different issues." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Image-patch counting bug and RMSNorm epsilon mismatch are different model-component errors." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Save-pretrained sharding size question and Qwen3.5-Moe loading error are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different GLM/DeepSeek tensor-shape issues with no shared concrete code path." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "DSA ReLU bug and DeepSeek transpose bug are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention FlashAttention 2, but they target different models and different support gaps." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "FlashAttention failures on different models with different symptoms; not a single fixable bug." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Compile-time float32 warning and BLIP2 dtype staying float32 are different dtype issues." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removed config API and gguf_file/pretrained loading breakage are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Sys.modules access in modeling_utils and a small testing_utils 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": 25, + "review_unit_index": 54, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42886", + "issue:42890", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43502", + "issue:43531", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44936", + "issue:44987", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46298", + "issue:46365", + "issue:46369", + "issue:46655", + "issue:46664", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46821", + "issue:46897", + "issue:46981" + ], + "soft_pairs": [ + "issue:43986|issue:45341", + "issue:42886|issue:46981", + "issue:44261|issue:44485", + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43531|issue:45341", + "issue:43874|issue:44263", + "issue:45440|issue:46172", + "issue:43097|issue:45588", + "issue:43673|issue:46123", + "issue:44410|issue:45375", + "issue:43986|issue:45372", + "issue:44038|issue:44410", + "issue:43097|issue:46050", + "issue:38617|issue:42757", + "issue:43064|issue:44315", + "issue:43208|issue:46897", + "issue:43986|issue:44484", + "issue:46655|issue:46664", + "issue:46298|issue:46752", + "issue:46172|issue:46710", + "issue:42890|issue:46298", + "issue:44062|issue:44117", + "issue:43097|issue:43873", + "issue:38617|issue:44351", + "issue:43012|issue:46821", + "issue:44117|issue:46701", + "issue:43502|issue:43992", + "issue:44936|issue:46365" + ], + "prepared_review_unit_hash": "d815a8b4832a59ddac4e65fb9cb1218887decff71b6a93551f764fb6d362f842", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35604, + "estimated_input_tokens": 8901, + "estimated_eval_tokens": 18058 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy and mostly consists of unrelated Transformer/HF issues that only share broad topical similarity. I do not see any soft pair that is clearly the same underlying bug/change, so I would not merge any of them.", + "confidence": 0.69, + "canonical_issue_reason": "If a representative issue must be chosen, issue:43873 has the broadest cross-cutting impact and the most discussion/references; however, the set does not really form a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue:43873 is the strongest single representative by activity and scope, but only as a weak canonical pick because the cluster is heterogeneous.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Both mention crashes, but one is an AutoProcessor/video-model dependency failure and the other is a testing_utils bug; different code paths and symptoms." + }, + { + "left": "issue:42886", + "right": "issue:46981", + "accept": false, + "reason": "Both involve hub/network behavior, but one is offline cache loading and the other is repeated 429 retries via HF_ENDPOINT; not the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "One is a missing rms_norm_eps config causing precision mismatch; the other is a GLM RoPE implementation issue. Different model components." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Both are model runtime failures, but one is an infinite loop on a specific image and the other is a memory overflow in sam3-litetext." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "The first is a tokenizer-mapping None assumption bug; the second is a model-specific load failure on transformers>=5.1.0. Too broad to be the same defect." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and a testing_utils bug are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Both touch GLM-related code, but one is an image processor missing a method and the other is a torch.split return-value issue; different failures." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeekV3 divergence from remote implementation is not the same as the deepseek-v4 comb.transpose bug." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder and flash_attention s_aux=None crash are unrelated change/bug paths." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "One is missing GenerationMixin cache during chunked_prefill; the other is a MaskGenerationPipeline partial-batch bookkeeping bug." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3_5/Qwen3-VL-Moe related, but one is missing projections in layer 0 and the other is a missing config field dropped by strict loading." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Both are import-time crashes, but they fail on different missing dependencies/symbols and in different processor-loading flows." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "General Qwen3-VL-Moe bug report versus a specific missing-attention-projection defect; not the same concrete issue." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "A removed method in v5 and a broken quantized save/load round-trip are different regressions." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different modules/symbols; superficial import-error similarity only." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Optimizer-state corruption with FSDP2/PEFT is unrelated to Liger Kernel not being applied under model_init." + }, + { + "left": "issue:43208", + "right": "issue:46897", + "accept": false, + "reason": "xLSTM training blockers and Florence2 double-shift loss bug are different model-specific training issues." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "AutoProcessor dependency crash and max_shard_size question are not the same bug." + }, + { + "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; related subsystem, different defect." + }, + { + "left": "issue:46298", + "right": "issue:46752", + "accept": false, + "reason": "A noisy logging complaint is unrelated to an apply_chat_template empty-list crash." + }, + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "A deepseek-v4 tensor transform bug is unrelated to an accuracy regression after upgrading Transformers." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Seed-related test flakiness and a noisy logging message do not describe the same problem." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizers.AddedToken keyword duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer initialization bugs." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Removed API in v5 and quantization/offloading issues are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different symbols and modules and likely different release regressions." + }, + { + "left": "issue:43012", + "right": "issue:46821", + "accept": false, + "reason": "A PyTorch warning during bfloat16 compilation is unrelated to a window_size AttributeError in TimesFM 2.5." + }, + { + "left": "issue:44117", + "right": "issue:46701", + "accept": false, + "reason": "Tokenizer mapping assumptions and a diffusion-gemma image-text pipeline failure are different code paths." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only still hitting API requests and UMT5Encoder missing embed_tokens.weight are unrelated loading bugs." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Trainer.evaluate after train and inability to import Trainer with datasets installed are not the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 55, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43761", + "issue:43824", + "issue:43825", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44016", + "issue:44077", + "issue:44162", + "issue:44222", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44704", + "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" + ], + "soft_pairs": [ + "issue:44704|issue:45103", + "issue:44380|issue:46421", + "issue:44938|issue:46178", + "issue:43986|issue:44016", + "issue:42907|issue:44655", + "issue:42907|issue:45753", + "issue:43825|issue:44655", + "issue:43012|issue:43550", + "issue:45712|issue:46298", + "issue:43012|issue:44368", + "issue:43208|issue:43550", + "issue:43992|issue:44484", + "issue:43352|issue:43976", + "issue:43352|issue:44380", + "issue:43066|issue:43906", + "issue:44162|issue:46172", + "issue:43097|issue:46298", + "issue:45584|issue:46365", + "issue:44380|issue:46132", + "issue:44077|issue:46365", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:44380|issue:45479", + "issue:42175|issue:46365", + "issue:42757|issue:43824", + "issue:43064|issue:44380", + "issue:42490|issue:43012", + "issue:44483|issue:44514", + "issue:43618|issue:44380", + "issue:43208|issue:43526", + "issue:38617|issue:43502", + "issue:43618|issue:43761" + ], + "prepared_review_unit_hash": "8dd5521e5b39940ec85fc6b432ecf7dffeb43aa3136cfaff9dbd62de351c2c38", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37196, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous and does not form a single duplicate set. The only plausible duplicate-like subcluster is the CLIP v5 regression pair around missing outputs from CLIPVisionModel forward.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43761 is the best anchor for the only near-duplicate subcluster: it clearly describes a v5 regression in CLIPVisionModel.forward, and 43618 looks like a narrower symptom of the same output-population bug.", + "canonical_pr_reason": null, + "best_issue_reason": "43761 is the most actionable and specific report in the set, with a concrete regression label, model path, and expected-vs-actual behavior that make it a strong triage target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different bugs: AutoProcessor kwarg forwarding vs auto_docstring crash under future annotations." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Different code paths: GPT2 attention scaling/backend behavior vs cache lazy initialization recompilation." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Unrelated import/runtime failures: Python 3.14 load issue vs tensorflow_text backend mapping error." + }, + { + "left": "issue:43986", + "right": "issue:44016", + "accept": false, + "reason": "Video AutoProcessor/torchvision crash is unrelated to a notebook syntax error." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized model persistence and the other is Pipeline.save_pretrained; too different." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Dequantized save failure is unrelated to a Qwen image-generation infinite loop." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Error-message wording for pipeline tasks is unrelated to pipeline serialization." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A precision warning during bfloat16 compilation is not the same as a Bamba torch.compile SDPA failure." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Dummy class cleanup/import leakage is unrelated to a noisy logging message." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Different warnings with different causes: float32 precision during compile vs tie_word_embeddings warning." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers are unrelated to Bamba torch.compile SDPA." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder is unrelated to max_shard_size defaults." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "FlashAttention support in Nemotron is unrelated to the Python version compatibility issue." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Different features and failures: Nemotron FlashAttention support vs GPT2 attention scaling ignored." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage is unrelated to the deepseek-v4 dtype/transpose bug." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Removed config API behavior is unrelated to a noisy logging message." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to Trainer import interference from datasets." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "GPT2 attention scaling bug is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init validation is unrelated to importing Trainer." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both mention MoE/FP8/Qwen, but one is save_pretrained FP8 and the other is a from_pretrained regression; not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config vocab_size AttributeError is unrelated to Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "GPT2 attention backend scaling is unrelated to the zero-loss classification bug." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Backend packaging/install issue is unrelated to Trainer import problems." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Different missing-symbol import errors from different packages/models." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local path warning on trailing slash is unrelated to a bfloat16 compile precision warning." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat completions API rejection is unrelated to Qwen2_5_VLProcessor batched padding behavior." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP output regression and GPT2 attention scaling are different model/code paths." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training blockers are unrelated to BeitImageProcessorFast label reduction." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Missing configuration_utils symbol import is unrelated to local_files_only API requests." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": true, + "reason": "Both describe the same CLIP v5 output regression in the CLIP vision forward path, with attentions/hidden_states no longer populated as expected." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The single accepted soft pair (43618/43761) is plausibly the same CLIP v5 output-regression cluster, and the rejected pairs are justified without overstating evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 56, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29870", + "issue:33453", + "issue:36010", + "issue:36296", + "issue:36683", + "issue:42757", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43867", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44077", + "issue:44230", + "issue:44355", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:44908", + "issue:44938", + "issue:45071", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45290", + "issue:45310", + "issue:46050", + "issue:46077", + "issue:46082", + "issue:46202", + "issue:46365", + "issue:46829", + "issue:46851", + "issue:46877" + ], + "soft_pairs": [ + "issue:43756|issue:44485", + "issue:42994|issue:44655", + "issue:44038|issue:44230", + "issue:36010|issue:42757", + "issue:43299|issue:43867", + "issue:45276|issue:46077", + "issue:44625|issue:46365", + "issue:43611|issue:44655", + "issue:44483|issue:45290", + "issue:43352|issue:43937", + "issue:36683|issue:43475", + "issue:44355|issue:46877", + "issue:44655|issue:46050", + "issue:44625|issue:46077", + "issue:36683|issue:44077", + "issue:43976|issue:44938", + "issue:43867|issue:45216", + "issue:46077|issue:46082", + "issue:43618|issue:45216", + "issue:43208|issue:46829", + "issue:36683|issue:43525", + "issue:36683|issue:44625", + "issue:29870|issue:36296", + "issue:43064|issue:44492", + "issue:43010|issue:45278", + "issue:44075|issue:46365", + "issue:43992|issue:45310", + "issue:44908|issue:46851", + "issue:29870|issue:33453", + "issue:44938|issue:45071", + "issue:44355|issue:44855", + "issue:43867|issue:46202", + "issue:36683|issue:43064" + ], + "prepared_review_unit_hash": "484f0bcc6f3764fadab56912c5904ab231f96330283afd8da7172909c193f8e3", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37133, + "estimated_input_tokens": 9284, + "estimated_eval_tokens": 18824 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a heterogeneous mix of unrelated regression reports, model-specific bugs, and API/import issues. The titles share broad themes like loading/saving/imports, but none of the soft pairs look like the same concrete bug or change, so I would not merge any of them.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 36683 is the highest-signal current report in the set: it is open, has the most discussion, and describes a concrete model/config regression. There is no true duplicate cluster here, but it is a reasonable representative issue for triage.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 36683 is the strongest standalone issue to anchor the cluster because it is active, well-discussed, and has a specific reproducible failure. It is not a duplicate of the others, just the best representative among unrelated issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Different model families and failures: Smollm3 RoPE layer loss vs GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both mention save_pretrained, but one is quantized model saving and the other is Pipeline object serialization; different code paths." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs fp8 inference support request; related model name, but not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Unrelated import errors from different dependencies: GenerationMixin from transformers.generation vs is_offline_mode from huggingface_hub." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are loading regressions, but one is Qwen3-VL-Moe model loading and the other is state_dict ordering during load_model." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4 issues, but one is resizing embeddings and the other is attention-mask construction; separate bugs." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation bug vs a Trainer import conflict with datasets; unrelated." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "base_model_prefix loading regression vs Pipeline save_pretrained failure; different APIs and failure modes." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Both concern chat-related behavior, but one is /v1/chat/completions request handling and the other is apply_chat_template crashing on tool-call assistant messages." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Nemotron flash-attention support limitation vs GLM-5 GenerationConfig validation error; unrelated." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs SAM 3 Video output missing fpn_position_embeddings; different models and attributes." + }, + { + "left": "issue:44355", + "right": "issue:46877", + "accept": false, + "reason": "Compiled Python file import errors vs auto_docstring printing doc-lint diagnostics at import time; not the same issue." + }, + { + "left": "issue:44655", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save_pretrained/from_pretrained, but one is Pipeline serialization and the other is quantized model roundtrip corruption." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 num_labels propagation vs Gemma4 bidirectional-attention mask logic; unrelated." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs patchtsmixer post_init restriction; no shared code path." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version compatibility reports, but they target different versions and different breakages." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "State_dict sorting/load ordering bug vs Qwen3.5 save_pretrained regression; both serialization-related but not the same defect." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention-mask bug vs LlamaConfig head_dim validation error; unrelated model/config issues." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIPOutput attentions field regression vs Qwen3.5 checkpoint save corruption; separate bugs." + }, + { + "left": "issue:43208", + "right": "issue:46829", + "accept": false, + "reason": "xLSTM training blockers vs VideoPrism hidden_states field mismatch; different subsystems and symptoms." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs Llama4Config missing pad_token_id; similar shape, but different models and config fields." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs Qwen3.5 num_labels propagation; unrelated config regressions." + }, + { + "left": "issue:29870", + "right": "issue:36296", + "accept": false, + "reason": "BART generate decoder_attention_mask UnboundLocalError vs tensor-parallel training bug; no shared concrete code path." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption vs a typo in cache strategy docs/code; not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "no_grad decoration for cache update vs broad import failures after upgrade; unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "SGD optimizer args ignored vs inability to import Trainer until datasets is removed; different components." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error; separate model loading bugs." + }, + { + "left": "issue:44908", + "right": "issue:46851", + "accept": false, + "reason": "Scheduler kwargs bug vs Qwen3.5 variable-length position_ids failure; unrelated." + }, + { + "left": "issue:29870", + "right": "issue:33453", + "accept": false, + "reason": "BART generate error vs tokenizer loading regression; both are regressions but not the same bug." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 import/load failure vs PretrainedConfig type-checking breakage in v5.4.0; different failures." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled-file import errors vs DebertaV2Model import failure from torch.jit.script parsing; different root causes." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "State_dict load ordering issue vs Gemma pipeline roundtrip failure on \u2581; unrelated serialization problems." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs Trainer FSDP2/PEFT optimizer-state issue; different layers of the stack." + } + ] + }, + "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": 25, + "review_unit_index": 57, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43065", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43824", + "issue:43867", + "issue:43874", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44734", + "issue:44749", + "issue:44964", + "issue:45125", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46132", + "issue:46285", + "issue:46617", + "issue:46664" + ], + "soft_pairs": [ + "issue:43867|issue:45375", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue:36296|issue:43575", + "issue:41669|issue:44246", + "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:42890|issue:45600", + "issue:43502|issue:44062", + "issue:42898|issue:46664", + "issue:44734|issue:46617", + "issue:43723|issue:44556", + "issue:44222|issue:45125", + "issue:43824|issue:46617", + "issue:43502|issue:44485", + "issue:43874|issue:46617", + "issue:43010|issue:43065", + "issue:44387|issue:44556", + "issue:43576|issue:46132", + "issue:30333|issue:33453", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:45125|issue:45375", + "issue:30990|issue:31795", + "issue:42548|issue:44351", + "issue:43502|issue:44222" + ], + "prepared_review_unit_hash": "808c3b99c4795793dbf1ca5513341c372984fa87d472a6fdcbf3e53248033b08", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37214, + "estimated_input_tokens": 9304, + "estimated_eval_tokens": 18864 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a loose collection of unrelated Transformers regressions: import/export errors, model-loading issues, tensor-parallel bugs, test flakes, and a few docs/behavior changes. The soft links look like keyword overlap rather than true duplicates, so I would not merge them into one cluster.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43723 is the broadest user-facing loading regression in the set and the best available anchor, though the overall cluster is too heterogeneous to have a strong true canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "43723 is the most general, central loading failure among the issues; it is the closest thing to a representative duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Different bugs: state_dict loading order vs missing Qwen3.5 vision config field." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Unrelated test issues: missing set_seed vs xdist capture-file race." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "Specific no_grad cache-update bug vs a broad, vague compatibility complaint." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Both involve tensor parallelism, but different failures and code paths." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Import slowness overlap only; one is a specific import-* change, the other is generic." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Different model/API regressions; no shared concrete bug." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Completely different: cache update semantics vs Python version support." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Both are loading-related, but one is state_dict ordering and the other is save_pretrained key mapping." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different model-specific from_pretrained failures, not one shared defect." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing import error vs tokenizer loading regression; unrelated." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Both are CLIP-related, but the specific regressions differ." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer discussion vs DeepSeekV3 implementation divergence; not the same bug." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Test seeding flake vs model-load failure for Phi-4 multimodal." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are test failures, but they have different root causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Missing set_seed in tests vs stale auto_mappings config references." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Network request bug under local_files_only vs AddedToken constructor argument error." + }, + { + "left": "issue:42898", + "right": "issue:46664", + "accept": false, + "reason": "Tokenization-space behavior change vs video num_frames semantics; unrelated." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "Serve KV-cache indexing crash vs cache FileNotFoundError on trust_remote_code." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer load regression vs checkpoint reload compatibility after upgrade." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained bug vs missing tensor-parallel plan; different fixes." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "Import error for a model class vs cache loading FileNotFoundError." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only networking bug vs GLM-5 RoPE discussion." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Missing image-patch helper vs cache FileNotFoundError; not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache no_grad issue vs dummy Conv2d placeholder; unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Different version regressions: int4 memory growth vs checkpoint reload failure." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken env command vs AttentionInterface behavior change; separate issues." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow reporting bug vs tokenizer loading regression; no overlap." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Related model family, but not enough to say it is the same concrete bug." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled-file errors vs slowdown after upgrade; unrelated." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Different missing top-level imports for different symbols." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "local_files_only network bug vs FP8 save_pretrained 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": 25, + "review_unit_index": 58, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:29870", + "issue:30333", + "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:43673", + "issue:43824", + "issue:43827", + "issue:43901", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44938", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45230", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45850", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46617", + "issue:46664", + "issue:46726" + ], + "soft_pairs": [ + "issue:43502|issue:44961", + "issue:46664|issue:46726", + "issue:43827|issue:44961", + "issue:43010|issue:43232", + "issue:30333|issue:36296", + "issue:29870|issue:31795", + "issue:43299|issue:44556", + "issue:33453|issue:34689", + "issue:44355|issue:45071", + "issue:43576|issue:45071", + "issue:42757|issue:43576", + "issue:44315|issue:44360", + "issue:31795|issue:38617", + "issue:43576|issue:43824", + "issue:44961|issue:45468", + "issue:44485|issue:44961", + "issue:44351|issue:46617", + "issue:36296|issue:43901", + "issue:43352|issue:44961", + "issue:43576|issue:44938", + "issue:43475|issue:46617", + "issue:46143|issue:46144", + "issue:44222|issue:44230", + "issue:44315|issue:46209", + "issue:41628|issue:42548", + "issue:43673|issue:44492", + "issue:44485|issue:46664", + "issue:45362|issue:45850", + "issue:44273|issue:45542", + "issue:44623|issue:45230", + "issue:44273|issue:45003", + "issue:44360|issue:46664" + ], + "prepared_review_unit_hash": "008c8bef7d00ca82705fc89a173399a3512d786c1cfca4e9be80098d0915021d", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 35851, + "estimated_input_tokens": 8963, + "estimated_eval_tokens": 18182 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are unrelated loading/import/regression/docs/model-specific issues, so the soft edges do not look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue is a reliable canonical because the set spans unrelated bugs and doc regressions with different code paths and symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global representative for deduping; the cluster does not form one coherent underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only/network behavior and an unrelated placeholder issue are not the same bug." + }, + { + "left": "issue:46664", + "right": "issue:46726", + "accept": false, + "reason": "Video frame-count semantics and missing-vision-backend import handling are different processor problems." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs about pipeline removals vs an unrelated issue; no shared code-path bug." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache no-grad decoration and generation kwargs sync_gpus ordering are distinct generation issues." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow failure reporting and tensor-parallel training are unrelated." + }, + { + "left": "issue:29870", + "right": "issue:31795", + "accept": false, + "reason": "A generate() UnboundLocalError is not the same as confusing forward docs." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading regressions, but for different versions/models and different failure modes." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression and Llama 3.2 model-loading breakage are not the same concrete defect." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-Python-file loading errors and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env CLI and config type-checking are different code paths." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing huggingface_hub symbol import and the env command failure are not clearly the same bug." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel/model_init behavior and the DSA ReLU discussion are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Forward-doc confusion and an ImportError for layer_type_validation are different issues." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "CLI env breakage and missing Qwen2_5_VL export are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No meaningful shared bug; the titles do not describe the same failure." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "RoPE implementation discussion and an unrelated placeholder issue are not duplicates." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "Missing HybridCache export and trust_remote_code cache FileNotFoundError are different loading failures." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training and pipeline docs behavior are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44961", + "accept": false, + "reason": "Flash-Attention support for Nemotron and an unrelated placeholder issue are not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env CLI and Python 3.14 import/load failure are different problems." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "SAM3 attribute failure and cache FileNotFoundError do not share a concrete code-path bug." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "Both touch RoFormer, but one is kwargs propagation and the other is attention-interface usage; related but not one underlying defect." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained behavior and FP8 inference support are different operations." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Model_init/Liger issues and B200 FP8 inference failure are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors from transformers, but for different symbols and likely different export regressions." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Generation cache missing and a typo in cache strategies are not the same bug." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "GLM-5 RoPE implementation and SmolVLM num_frames semantics are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A specific Qwen3.5 chat crash and a broad complaint about many models are not the same concrete issue." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading and tensorboard-without-TensorFlow backend errors are different failures." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Processor.save_pretrained missing files and an empty bug report are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44961", + "accept": false, + "reason": "Missing Qwen2_5_VL export and an unrelated placeholder issue are not duplicates." + }, + { + "left": "issue:45003", + "right": "issue:44355", + "accept": false, + "reason": "Unsafe sys.modules access and compiled-Python-file load errors are different underlying problems." + }, + { + "left": "issue:44360", + "right": "issue:46664", + "accept": false, + "reason": "A DSA/ReLU discussion and a video processor num_frames 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": 25, + "review_unit_index": 59, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43519", + "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:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46218", + "issue:46328", + "issue:46655", + "issue:46894" + ], + "soft_pairs": [ + "issue:44038|issue:45125", + "issue:44393|issue:45071", + "issue:41669|issue:43673", + "issue:44393|issue:45753", + "issue:43576|issue:44556", + "issue:43525|issue:45850", + "issue:43572|issue:45850", + "issue:44556|issue:45092", + "issue:44492|issue:45850", + "issue:45278|issue:45850", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:43408|issue:44315", + "issue:44230|issue:46095", + "issue:44246|issue:44273", + "issue:42890|issue:44908", + "issue:43576|issue:44908", + "issue:43352|issue:44393", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44295|issue:45230", + "issue:44492|issue:46328", + "issue:44297|issue:45230", + "issue:44704|issue:44908", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:43976|issue:46218", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:31795|issue:33453", + "issue:43576|issue:44273", + "issue:44315|issue:44829", + "issue:43519|issue:46655", + "issue:44360|issue:45468", + "issue:44273|issue:44908", + "issue:45468|issue:46894", + "issue:45230|issue:45310", + "issue:44655|issue:44861", + "issue:43519|issue:44485", + "issue:43576|issue:46218", + "issue:43519|issue:44908" + ], + "prepared_review_unit_hash": "257d8464f74cd5ca33c3b7de0bee042c4ba4975dcca9da90399d6a6044cfa4cb", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 39670, + "estimated_input_tokens": 9918, + "estimated_eval_tokens": 20092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of Transformers v5 regressions and model-specific bugs, with only one clear duplicate-like pair around import/lazy-loading slowdown. No pull requests are present in the cluster.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 44273 is the clearest specific report in the import/lazy-loading subcluster and is the best duplicate target among the provided issues.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the strongest overall umbrella for the cluster because it captures the broad post-upgrade import/load breakage theme, even though it is still broader than the individual bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3.5-related problems: one is a model/support bug, the other is missing tensor-parallel metadata." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Different code paths and symptoms: Qwen3-VL 2D box output issue vs PretrainedConfig type-checking regression." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time performance from wildcard imports is not the same bug as missing GenerationMixin cache during chunked prefill." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but not the same concrete bug: bbox output error vs infinite loop on one image." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "CLI/environment command breakage is unrelated to checkpoint reload incompatibility after v5 upgrade." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Specific Llama4 pad_token_id regression is not the same as a generic 'v5 breaks many models' complaint." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "Specific StableLm pad_token_idx regression is not the same underlying issue as the broad v5 breakage report." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 compatibility complaints, but they point to different concrete failures and code paths." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A typo in cache strategies is unrelated to a broad version-breakage report." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad v5 complaints, but one is specifically about import errors after upgrade and the other is a generic statement." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API guidance is unrelated to the cache-strategy typo." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import performance regression is different from a specific missing model import/module error." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Wrong-model warning and Liger Kernel application under model_init are unrelated issues." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support request and weight-conversion-script question are different feature areas." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": true, + "reason": "Both describe the same lazy-loading/import slowdown problem; one reports the symptom and the other names the underlying lazy-loading failure." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test flakiness from missing seeds is unrelated to inverse_sqrt scheduler kwargs handling." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env CLI and scheduler kwargs are unrelated bugs." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support gating for Nemotron is unrelated to Qwen3-VL bbox output errors." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python-version compatibility and Qwen3-VL hallucination/bbox errors are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs handling is unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Specific position_ids buffer access bug is not the same as a generic bug report." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache strategy typo and SequenceFeatureExtractor.pad inefficiency are different issues." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save_pretrained mismatch is unrelated to a generic bug report." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs passthrough and scheduler kwargs are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Wildcard-import slowdown is not the same as a HybridCache import failure." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import performance regression is unrelated to a testing_utils bug." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python-version compatibility is not the same as a generic 'many models break' complaint." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python-version compatibility is unrelated to a missing Qwen3_5ForCausalLM module import." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL bbox hallucination and Qwen3.5 Moe from_pretrained failure are different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "PatchTSMixer post_init validation and tied-weight-key AttributeError are different code paths." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion about inputs/cache is unrelated to tokenizer loading regression." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env CLI and lazy-loading/import slowdown are different bugs." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application under model_init is unrelated to flash_attention_3 degenerate training." + }, + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Qwen3VL timestamp calculation and SmolVLM video padding are different processor bugs." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer/ReLU discussion and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Generic bug report and Qwen3.5 Moe from_pretrained regression are not duplicates." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained failure is unrelated to tied-weight-key AttributeError." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Timestamp calculation in Qwen3VL Processor is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Env command breakage and missing Qwen3_5ForCausalLM import are unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp logic is unrelated to inverse_sqrt scheduler kwargs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The chosen umbrella issue and the lazy-loading/import duplicate pair are both supported by the titles and soft-pair evidence, and the note that there are no PRs is correct." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 25, + "review_unit_index": 60, + "review_unit_count": 60, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-503", + "nodes": [ + "issue:28348", + "issue:29870", + "issue:30333", + "issue:31515", + "issue:31795", + "issue:34689", + "issue:36296", + "issue:37934", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43618", + "issue:43668", + "issue:43704", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44263", + "issue:44351", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44938", + "issue:44998", + "issue:45250", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46144", + "issue:46172", + "issue:46218" + ], + "soft_pairs": [ + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:43704|issue:44485", + "issue:46018|issue:46172", + "issue:44485|issue:45468", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:45850|issue:46218", + "issue:43519|issue:44623", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:43519|issue:43704", + "issue:43298|issue:46095", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:28348|issue:45250", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:31515|issue:38617", + "issue:43572|issue:44861", + "issue:43525|issue:44861", + "issue:44230|issue:44623", + "issue:28348|issue:43352", + "issue:29870|issue:31515", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:43298|issue:44230", + "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": "6432d40ee442d1f271ca9a99550eff4c416d4b24fdcc27d6f288ac417db9b82c", + "component_budget": { + "node_count": 503, + "item_count": 503, + "soft_pair_count": 1851, + "serialized_chars": 960087, + "estimated_input_tokens": 240022, + "estimated_eval_tokens": 480300 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 46, + "serialized_chars": 40596, + "estimated_input_tokens": 10149, + "estimated_eval_tokens": 20554 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a loose similarity cluster, not a true duplicate group. The strongest common thread is broad Transformers 5.x regression/compatibility complaints, but the items span unrelated model bugs, import errors, docs, and support requests, so I would not merge them as duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:45850 is the broadest umbrella (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the shared regression theme, though the cluster is too heterogeneous for a strict duplicate canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the most representative high-level regression report in the set, but it is still too vague to serve as a true dedupe anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-* performance regression and multi-thread VRAM leak are different failure modes." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion is not the same bug as a model-loading breakage." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in dataloader threads and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Shared-expert gating logic and DeepSeek-V4 weight transposition are different code paths." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "RoPE implementation discussion and Gemma-4 audio positional encoding are not the same issue." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API guidance and a VRAM leak bug do not share the same root cause." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Local-files-only network requests and Qwen3VL timestamp math are unrelated." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A broad v5 regression report is not the same as a specific Qwen3_5 import failure." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp calculation and missing save_pretrained files are different processor bugs." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split output handling and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch is a different bug from audio positional encoding." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Timestamp calculation and multi-thread VRAM leak do not match." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are different requests." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Both are import failures, but for different missing symbols/modules and not one concrete bug." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support for NemotronH and Qwen3_5 import failure are unrelated." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5, but one is a general behavior change and the other a specific decoder-type bug." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "Specific GPT2 flash-attention support and a generic Flash Attention 2.0 request are not the same concrete issue." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs handling and the unrelated 'Unemployment' issue are clearly different." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Cache-strategy typo and the unrelated 'Unemployment' issue are not duplicates." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIP attentions regression and the unrelated 'Unemployment' issue are different." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "The unrelated 'Unemployment' issue is not the same as the broad v5 breakage report." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 load failure and the unrelated 'Unemployment' issue do not match." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key crash and chat-template tool-call crash are different bugs." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained loading and a missing import symbol are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "Missing config pad token index and tied-weight key crash are different problems." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing Llama4Config pad_token_id and tied-weight key crash are not the same issue." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 support request and missing processor.save_pretrained files are unrelated." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but they target different models and code paths." + }, + { + "left": "issue:29870", + "right": "issue:31515", + "accept": false, + "reason": "BART generate() mask bug and slow checkpoint loading are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting on failure and slow from_pretrained loading are different concerns." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading speed and tensor-parallel training bug are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading performance and documentation clarity are not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 inference support and a Qwen3.5 MoE from_pretrained error are different issues." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and FP8 support are separate feature requests." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are different bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A config type-hint issue is not the same as a tokenizer regex warning." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Unused ignore_value and missing kwargs in loss calculation are different code paths." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-function kwargs handling and RoFormer attention implementation are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2former ignore_value and tokenizer regex warnings are not duplicates." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2former ignore_value handling and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Model bug and a typo in quantization-operation naming are different." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Type-hint correctness and a typo fix are not the same issue." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs propagation and ModernBERT type hints are unrelated." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4 normalization semantics and ModernBERT type hints are different topics." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2former ignore_value handling and ModernBERT config typing 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": 2, + "component_count": 25, + "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": "56160a4ca2bae4cf4f36cc3eef09302bd1586bf40f028fe05bda7618f1b672d5", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are not duplicates: one is about tokenizer truncation behavior, one requests ReduceLROnPlateau-like scheduler behavior, and one asks for Wav2Vec2 word time offsets. They are separate feature requests in different components.", + "confidence": 0.98, + "canonical_issue_reason": "issue:10519 is the most concrete and self-contained feature request, with a clearly stated desired behavior and scope.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:10519 is the best representative of the cluster because it is the most specific and actionable issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different topics and code paths: tokenizer truncation control vs Wav2Vec2 word timing metadata." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Unrelated feature requests: truncation direction option vs learning-rate scheduler behavior." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Separate requests in different subsystems: ReduceLROnPlateau-like scheduling vs Wav2Vec2 alignment/time offsets." + } + ] + }, + "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": 25, + "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": "2a9bd14cb250ded5b293b7203536ad019b7351fe6129598f3d9b589d62920175", + "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": "Only one issue appears to be about a model-loading failure; the other two are unrelated topics (Tapas tokenization and private-model auth error messaging). No soft pair should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 16618 is the clearest representative of the model-loading problem ('Can't load the model for bert-base-uncased') and is more general than the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "16618 best captures the underlying user-visible failure around loading a Hugging Face model; 15980 is specifically about an authentication/error-message edge case, and 13244 is about tokenizer behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different problems: Tapas tokenization mismatch vs. model loading failure for bert-base-uncased." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Related only at a high level to model download/load errors, but 15980 is specifically about private model download without login and error messaging, not the same concrete bug." + } + ] + }, + "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": 25, + "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": "72a87f88f84f9289c8b321cd757fa3edb42240cd2b6bb14565d1e263ea7a3d38", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate community-contribution feature requests in different areas (Flash Attention 2 vs GGUF support). They share a generic framing but not the same underlying change, so they should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Not the same bug/change: one asks for Flash Attention 2 support for more architectures, the other asks for GGUF support for more architectures. They are unrelated feature requests with no shared targets/files or concrete common code path." + } + ] + }, + "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": 25, + "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": "d5c1638b212a772c3447e004d9d13e31e36ab3904e1011d2ae8b29b41303c810", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15974, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15974, + "estimated_input_tokens": 3994, + "estimated_eval_tokens": 8244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Mostly not duplicates: several issues are separate docs-translation requests for different languages. The only clear duplicate groups are the identical placeholder-template issues and the identical Thai pair; the rest should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43834 is the best anchor among the duplicate-like items because it is the earliest of the identical placeholder-template translation issues and sits at the center of that exact-duplicate subgroup.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43834 is the most representative issue in the cluster, but only for the repeated placeholder-template translation-request subgroup, not for the language-specific requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different language requests (Turkish vs generic template); same theme but not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language requests; these are separate translation tickets, not one bug/change." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish request vs generic placeholder template; related topic, but not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different target languages and separate translation work items." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish and Spanish translation requests are distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request is separate from the generic placeholder-template issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language translation requests; not the same underlying ticket." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish vs Romanian translation requests are separate." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Spanish request vs generic template issue; same subsystem, different issue." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request and generic template issue are not the same underlying task." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian request vs generic placeholder-template issue; related, but distinct." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian request is separate from the repeated generic template issues." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Romanian translation issue vs unrelated Thai-titled issue." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation issue vs unrelated Thai-titled issue." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Same placeholder title and near-identical body; looks like an exact duplicate of the same translation request." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same placeholder-template issue; the bodies/titles are effectively the same." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Very similar body text, but the titles are unrelated and this does not clearly look like the same issue." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Same as above: high textual similarity, but not enough evidence that it is the same underlying issue." + }, + { + "left": "issue:43834", + "right": "issue:46931", + "accept": false, + "reason": "Different Thai-language title and not clearly the same request as the placeholder template." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Identical placeholder-template translation issue; this is a clear duplicate." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Text similarity is high, but the titles point to different issues and the match is not clear enough." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "High similarity alone is not enough; the issue titles are unrelated." + }, + { + "left": "issue:44365", + "right": "issue:46931", + "accept": false, + "reason": "Different Thai-titled issue; not clearly the same underlying task." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Very similar bodies, but different titles and no clear evidence they are the same issue." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Similar content, but not enough to treat as the same issue." + }, + { + "left": "issue:44370", + "right": "issue:46931", + "accept": false, + "reason": "Different Thai-titled issue; not the same request." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Same Thai title/body, so this is an exact duplicate pair." + }, + { + "left": "issue:46222", + "right": "issue:46931", + "accept": false, + "reason": "The titles differ and the evidence does not show a clear one-to-one duplicate." + }, + { + "left": "issue:46226", + "right": "issue:46931", + "accept": false, + "reason": "Different titles and no strong proof they are the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The summary correctly separates the language-specific translation requests from the clear duplicate subgroups, and the accepted pairs are supported by nearly/fully identical titles and bodies. No obvious overstatement beyond the evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 25, + "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-36718-19", + "nodes": [ + "issue:36718", + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653" + ], + "soft_pairs": [ + "issue:36718|issue:41115", + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:42738|issue:46653", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue: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": "05774a1c79042f3e14b8d00c26f65df4a79a30feb7b840f6b042ed74b82e4b6d", + "component_budget": { + "node_count": 19, + "item_count": 19, + "soft_pair_count": 40, + "serialized_chars": 24676, + "estimated_input_tokens": 6169, + "estimated_eval_tokens": 12594 + }, + "budget": { + "node_count": 19, + "item_count": 19, + "soft_pair_count": 40, + "serialized_chars": 24676, + "estimated_input_tokens": 6169, + "estimated_eval_tokens": 12594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are all about adding or supporting different models/features, but they are not duplicates of one another; the cluster is a loose thematic group, not a single underlying request.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits: the items request distinct models or capabilities (e.g. RoMa, EuroBert, RF-DETR, Qwen3-TTS/ASR, Voxtral, etc.) rather than the same bug/change.", + "canonical_pr_reason": null, + "best_issue_reason": "None of the issues is a strong global canonical. If a representative had to be picked, issue 43671 is a concrete model-support request, but it is not a duplicate of the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36718", + "right": "issue:41115", + "accept": false, + "reason": "Different model families: RoMa keypoint matcher vs MiniCPM3 architecture." + }, + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert config addition and DEIMv2 are unrelated model requests." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config work is unrelated to Qwen3-TTS support." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert and RF-DETR are different models with different integration needs." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are separate detector requests, not the same change." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio are different model domains." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR detection support is unrelated to the weight-sparse transformer request." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are different model integrations." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are separate model requests." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated models." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT support and circuitgpt weight-sparse transformer are different changes." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio target different architectures and tasks." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are distinct model support requests." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture and Qwen3-ASR are not the same model or code path." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 support is unrelated to the OpenAI sparse transformer request." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are separate model integrations." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture work is unrelated to Voxtral TTS support." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are different models and tasks." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different model requests." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT models and Qwen3-TTS are unrelated." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and BERT-like models with RoPE are only broadly related by theme, not by the same underlying change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS are unrelated changes." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "BERT-like model support and the unrelated person-only issue are not duplicates." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are related by branding but different modalities and code paths." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support and the sparse transformer request are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations, not one bug." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS support and DeepSeek-OCR2 are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are different speech/audio models." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and circuitgpt are different features." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model domains." + }, + { + "left": "issue:43909", + "right": "issue:46645", + "accept": false, + "reason": "LFM2.5 Audio support and SWA+sink for Granite are not the same change." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "OpenAI sparse transformer support and EUPE are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are distinct speech-model support requests." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated models." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different tasks and code paths." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are separate vision-language model requests." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT models and the unrelated person issue are not duplicates." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "Voxtral TTS support and the unrelated person issue are not duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 25, + "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": "fce130c229484c5968d6420ea9c16206fce5ab6b7b8eda936c3615c24d88aa30", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one is a docs request about adding Pruna, and the other is a Vision RoPE refactor. The soft similarity is superficial only.", + "confidence": 0.99, + "canonical_issue_reason": "Issue 46443 is the better canonical representative because it is the substantive technical item; issue 38740 is a separate docs request and not part of the same change/problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46443 is the clearest standalone issue in the cluster. Issue 38740 is unrelated and should not be used as the representative problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different domains and intents: documentation/vendor listing vs. a model architecture refactor. No shared code path or underlying bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 25, + "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": "9fbe694375bc10af878582b80d7ec087fd4301715ac7e221203919df32a8921f", + "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": "These two issues are not duplicates: one is a vague request about block decoding, while the other is a specific feature/support request for DeepSeek v3.2. They do not appear to describe the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "#43037 is the more specific and actionable issue, but it is not a duplicate of #41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "The titles and contexts are unrelated: 'Set Block Decoding' is a generic decoding-control request, while 'DeepSeek v3.2 support' is a model/version support request. No clear shared bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 25, + "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": "ac8c3f5871fdad618f17c5bd20c449c6b67fbee5c68d78320684656029bced2f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one asks for a `from_pretrained()` flag for `TiledFusedLogitsLoss`/FLCE, while the other requests integrating `TiledMLP` for lower memory use. They share only the broad \u201cTiled\u201d naming and target different components and changes.", + "confidence": 0.97, + "canonical_issue_reason": "No true canonical duplicate here; the issues describe different features in different code paths, so neither should be collapsed into the other.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a representative issue for the cluster, #41306 is slightly more concrete and implementation-specific, but it should still remain separate from #41826.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41306", + "right": "issue:41826", + "accept": false, + "reason": "Reject: the titles and problem statements concern different modules and different requested changes (`FLCE` flag in `from_pretrained()` vs. integrating `TiledMLP` for memory reduction). Similar naming is insufficient to treat them as the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 25, + "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": "843a994792c93835f354b09d35fd506e2568a93d599e915d4f5debc3ef3dfbdd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are related to Trainer customization, but they describe different problems: one is about `remove_unused_column`, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "Issue #43598 is the more concrete and narrowly scoped report, centered on a specific Trainer behavior (`remove_unused_column`) and its customizability.", + "canonical_pr_reason": null, + "best_issue_reason": "#43598 is the best issue anchor because it names the exact Trainer mechanism involved and is easier to map to a concrete code change or discussion.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They share a broad Trainer/customization theme, but one concerns unused-column removal and the other concerns FLOPS estimation for custom trainers; these are different underlying bugs." + } + ] + }, + "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": 25, + "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": "5df20f030f06c2a9755d026fd9248bdfdb69b8b743cec48cb04c99c4a137249b", + "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": "The two issues are related only at a very broad level (Qwen model support in Auto classes) but describe different requests: one asks for registration of Qwen3-Omni in AutoModel/AutoModelForConditionalGeneration, while the other asks for AutoModelForSequenceClassification support for Qwen3.5. They are not the same underlying bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "Issue #43684 is the broader and more general request, focused on AutoModel registration for a specific Qwen3-Omni model, so it is the better canonical representative of this small cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43684 is the best issue candidate because it is more general and closer to an AutoModel registration gap, whereas #44405 is a separate task for sequence-classification support on a different Qwen variant.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families/variants and different functionality: model registration for Qwen3-Omni vs sequence-classification support for Qwen3.5. They share the same ecosystem but not the same concrete fix." + } + ] + }, + "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": 25, + "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": "7b27d96e4216afe1eed2b8f25fb2d64999c8d02ba91745032d9c86942ef4cc5d", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Four closed issues with very little content; one has a distinct title (\u201cFett\u201d) and three share the same generic title (\u201chuggingface \u201d). There isn\u2019t enough concrete information to confidently merge any pair as the same underlying bug.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 43845 is the best representative of the repeated \u2018huggingface \u2019 titles and is the most central item in the apparent mini-cluster, but the available text is too sparse to prove it is a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the best issue to anchor the cluster because it matches the repeated title pattern and is one of the earliest of the similar issues, but the evidence is too thin for a stronger claim.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Only weak title similarity; \u2018Fett\u2019 and \u2018huggingface \u2019 do not indicate the same underlying bug." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Only weak title similarity; no shared details or explicit linkage show the same issue." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Only weak title similarity; the distinct title makes a duplicate relationship unlikely." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "The titles match, but there is no body detail or linkage confirming the same concrete problem, so this is too thin to accept." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same generic title, but no substantive description or references to establish the same underlying bug." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Same generic title, but the available evidence is insufficient to confirm a duplicate relationship." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "4af827ef5cc19a9b6c2b637ccca4a5d639648bec3654a0f67931e6f6cffff697", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are low-information, mostly unrelated issue reports: three identical spam placeholders, one documentation suggestion, and one branded phrase. There is no evidence they describe the same underlying bug or change, so none of the soft pairs should be merged as duplicates.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44371 is the earliest of the identical \"\" entries, so it is the most reasonable canonical representative for the spam-like subset.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44371 is the best overall representative because it matches the repeated placeholder spam pattern and is the earliest of the identical titles.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are placeholder spam issues, but there is no concrete shared bug or change description to justify treating them as duplicates." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same generic spam title, but no substantive content indicates the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam placeholder text and the other is a documentation enhancement request; they are clearly different artifacts." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "A spam placeholder and a distinct phrase about \"Google LLC\" are not the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation enhancement versus spam placeholder; no common concrete problem." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Documentation enhancement versus spam placeholder; unrelated." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders, but the available information is too generic to confirm the same underlying report beyond superficial similarity." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and a distinct branded phrase; no indication of the same bug or request." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder and a distinct branded phrase; unrelated." + } + ] + }, + "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": 25, + "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": "a31fd907d278bd37b5386c7b5044a15e1528d1e31f63215411c5e9027bd0920d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues appear to be duplicates of the same report, with identical titles and near-identical body size/structure. Issue #45145 is the better canonical choice because it has an inbound reference and was created earlier.", + "confidence": 0.98, + "canonical_issue_reason": "#45145 is slightly better as the canonical issue: it was created earlier and already has an inbound reference, suggesting it is the more established thread.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best representative issue for deduplication for the same reasons: earlier timestamp and more linkage/activity than #45151.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Accept: the issues have the same title and highly similar body characteristics, indicating the same underlying report rather than merely related topics." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the issues share the same title, have close body lengths, and the earlier issue (#45145) also has an inbound reference, so treating #45145 as canonical is reasonable. The wording is mostly conservative; the only slight caution is that \u201cbody size/structure\u201d is stronger than the packet explicitly proves, but the duplicate assessment itself is supported by the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 25, + "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": 25, + "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": "2fd8dd3a4ce2bf6f160770247a357d9a52e52cb6d27e557c3e31971a8fac2c11", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues do not look like duplicates of the same bug or change; they appear to be separate person-name/date entries with only a shared surname.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a clear canonical duplicate target; if forced, 45762 is only a marginal representative, not because it matches 45763.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Different names and different dates; no evidence of the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 25, + "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": "c6ae58f6e3917949458a07d8459aa5d44251b817f5794deb0b81f0d51a46c684", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are bare title-only stubs with different personal-name/date strings and no shared technical content, so they do not look like duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good canonical target; there is no evidence of the same underlying bug or request, only unrelated name/date titles with no body.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Reject: the titles refer to different people and dates, with no body, references, or technical overlap indicating 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": 25, + "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": "ced237c45ae21303ebba322476c1b09c3f88d19b7eca17b51c2cfa469b025691", + "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 head, while the other asks for FP8 training support. They do not describe the same bug or change.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue for this pair; they target different functionality and code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "45810 is the more self-contained and specific request, but it is not a true duplicate of 46295.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different feature areas: model-head addition for token classification versus FP8 training support. They are not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 25, + "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": "2334846e7b90faee21adaf922a7ade3774e74930ed9eaf0abf610e435e80a9ac", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The items split into two separate exact-duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is about a financial dataset bundle and is not the same underlying request as the example/service issues, so it should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "45872 is a clean exact-duplicate representative of the more specific DataForge financial time series + scraping service request; it is the clearest standalone issue among the items.", + "canonical_pr_reason": null, + "best_issue_reason": "45872 is the strongest issue to keep as the representative because it is an exact duplicate pair with 45873 and has the most concrete, differentiated title.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Dataset bundle vs example application; related domain, but not the same issue or change." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Dataset bundle vs example application; similar financial theme, but different artifact type and intent." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Same title and same body length; looks like an exact duplicate issue." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both are financial DataForge ideas, but one is an example and the other adds a scraping service; not the same concrete request." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Example issue vs scraping-service request; same broad area, different underlying ask." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Same broad financial topic, but one is an example and the other is a different service-oriented request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Same broad topic, but not the same underlying issue; one is an example, the other is a scraping-service request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Same title and same body length; looks like an exact duplicate issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded in the packet. The two accepted soft edges are supported by identical titles, identical body lengths, and jaccard 1.0, while the false edges stay conservative by distinguishing the dataset-bundle item from the example/service requests. The representative choice of 45872 is also reasonable." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 25, + "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": "9337c6ed8b3a6109cab9d1876ea3548c95c075e8a112826b467d6b885707c50d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1637, + "estimated_input_tokens": 410, + "estimated_eval_tokens": 1076 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1637, + "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 around linear attention, but they describe different failure modes: one is wrong-shape corruption during early initialization, the other is a KeyError for `linear_attention` in Qwen3.5. I would keep them separate absent stronger evidence they share the exact same root cause.", + "confidence": 0.73, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it is the more general description of the static-cache/linear-attention corruption, while 46441 is a specific model-specific crash report.", + "canonical_pr_reason": null, + "best_issue_reason": "46439 is the best issue candidate because it names the underlying mechanism and failure class more directly than the narrower Qwen3.5-specific report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Reject: both mention static cache and linear attention, but one reports shape corruption in early initialization and the other reports a `KeyError` on `linear_attention`; these look like related symptoms, not clearly the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 25, + "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": "2d912bbda9b8672e316bec841fbb20eb470403f079b49ed4ec7281ef14dcda02", + "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 near-exact duplicates: same title, same body length, and the same tokenizer bug description. The open issue is the better canonical target.", + "confidence": 0.99, + "canonical_issue_reason": "Issue #46552 is the better canonical issue because it is still open and appears to be the active report for the same exact bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #46552 is the best issue to keep as the primary tracker: it has the same problem statement as #46551 and is open with inbound references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "Same title and same underlying Wav2Vec2CTCTokenizer bug report; these are duplicate issues for the same defect." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The packet supports the core conclusion: both issues have the same title and identical body length, and the soft-similarity signal is maximal. Treating #46552 as the canonical open tracker is reasonable, and the duplicate-pair verdict is appropriately conservative." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 22, + "component_count": 25, + "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": "ff738f307664558bac3f47a38ea47d2b5cdb5ea31c4a1b9dd9d6939d9832305b", + "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 describe different training failures: one is a tensor-shape mismatch during object detection training, the other is an out-of-memory condition after several epochs in RT-DETR v2. They are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "issue:46657 is the better anchor because it reports a concrete tensor size mismatch bug; issue:46692 is a separate memory-exhaustion report with a different symptom and likely different root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46657 is the more specific and self-contained bug report, while issue:46692 is about OOM during training and does not match the tensor mismatch failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46657", + "right": "issue:46692", + "accept": false, + "reason": "Different underlying problems: tensor size mismatch vs. out-of-memory after epochs. Similar training context, but not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 23, + "component_count": 25, + "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": "c4f8af730f86dc338d589c865b46ad095648a1b135b8dc0f923812fa3fde022f", + "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 issues are about different failures in different loading paths: ZeRO-3 zero.Init partitioning of a composite submodule vs 4-bit auto device mapping wrongly spilling to CPU. They should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "If a single representative is needed, issue 46822 is the better anchor because it describes a specific ZeRO-3 partitioning bug with a clear failure mode and reproduction context; issue 46823 is a separate 4-bit dispatch problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46822 is the best overall issue candidate for this cluster because it is more self-contained and points to a concrete code path (zero.Init / composite submodule partitioning).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46822", + "right": "issue:46823", + "accept": false, + "reason": "Different underlying bugs and code paths: ZeRO-3 partitioning/OOM during model load versus 4-bit device_map/max_memory CPU dispatch ValueError. Shared theme is model loading, but not the same defect." + } + ] + }, + "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": 25, + "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-46906-5", + "nodes": [ + "issue:46906", + "issue:46940", + "issue:46948", + "issue:46969", + "issue:46985" + ], + "soft_pairs": [ + "issue:46906|issue:46940", + "issue:46906|issue:46948", + "issue:46906|issue:46969", + "issue:46906|issue:46985", + "issue:46940|issue:46948", + "issue:46940|issue:46969", + "issue:46940|issue:46985", + "issue:46948|issue:46969", + "issue:46948|issue:46985", + "issue:46969|issue:46985" + ], + "prepared_review_unit_hash": "fd17e0271905c4434cfcda0effc42b67b7654f1d9ebb1bda03a1a6f44b24ada8", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6439, + "estimated_input_tokens": 1610, + "estimated_eval_tokens": 3476 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6439, + "estimated_input_tokens": 1610, + "estimated_eval_tokens": 3476 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are highly similar daily triage issues, but they appear to be separate reports for different dates (2026-06-26 through 2026-06-30), not duplicates of one underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate exists here; the items are date-stamped triage reports for different days. If one must be chosen as an anchor, issue 46906 is the earliest and most referenced.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46906 is the best representative only as an anchor because it is the earliest and has the most inbound references/discussion, but it should not be treated as a duplicate of the later daily triage issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46906", + "right": "issue:46940", + "accept": false, + "reason": "Same template and subsystem, but the titles/body indicate different daily triage reports on different dates, so they are separate issues rather than duplicates." + }, + { + "left": "issue:46906", + "right": "issue:46948", + "accept": false, + "reason": "Date-separated triage entries; similarity is format-driven, not evidence of the same underlying bug." + }, + { + "left": "issue:46906", + "right": "issue:46969", + "accept": false, + "reason": "These are distinct daily triage reports for different days, not one shared issue." + }, + { + "left": "issue:46906", + "right": "issue:46985", + "accept": false, + "reason": "Different report date and separate triage context; not the same underlying issue." + }, + { + "left": "issue:46940", + "right": "issue:46948", + "accept": false, + "reason": "Both are daily triage reports, but for different dates, so they should not be merged." + }, + { + "left": "issue:46940", + "right": "issue:46969", + "accept": false, + "reason": "The overlap is only in wording and subsystem; the reports are for different days." + }, + { + "left": "issue:46940", + "right": "issue:46985", + "accept": false, + "reason": "Separate date-stamped triage issues, not duplicate bug reports." + }, + { + "left": "issue:46948", + "right": "issue:46969", + "accept": false, + "reason": "Different daily triage entries; no shared concrete bug/change is indicated." + }, + { + "left": "issue:46948", + "right": "issue:46985", + "accept": false, + "reason": "These are consecutive-day triage reports, but the evidence points to distinct issues." + }, + { + "left": "issue:46969", + "right": "issue:46985", + "accept": false, + "reason": "Same format and area, but different dates and no concrete indication of 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": 25, + "component_count": 25, + "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": "ada375511251ec3ea7bdb00d4eb660249dacd71fbc19edde65e47ffcac916be6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1490, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1490, + "estimated_input_tokens": 373, + "estimated_eval_tokens": 1002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are the same feature request: identical titles, very similar bodies, and no evidence of distinct scope. The open issue is the better canonical tracking issue; the closed one is a duplicate or older copy.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #46984 is open, so it is the better long-lived tracker for the same RWKV7 support request. #46983 appears to be a duplicate with no distinguishing details.", + "canonical_pr_reason": null, + "best_issue_reason": "#46984 is the best issue overall because it is open and can continue collecting progress and discussion for the same request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46983", + "right": "issue:46984", + "accept": true, + "reason": "Same title and near-identical request for RWKV7 model support; no signs of different scope, so this is a duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusion is well grounded in the packet: both items have the same title, the bodies are reported as highly similar via soft similarity, and there is no evidence of different scope or separate targets. Choosing the open issue as the canonical tracker is a conservative and reasonable call." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-39", + "nodes": [ + "pull_request:40546", + "pull_request:43451", + "pull_request:44517", + "pull_request:45862", + "pull_request:46180", + "pull_request:46196", + "pull_request:46266", + "pull_request:46425", + "pull_request:46658", + "pull_request:46707", + "pull_request:46836", + "pull_request:46891", + "pull_request:46895", + "pull_request:46902", + "pull_request:46903", + "pull_request:46907", + "pull_request:46908", + "pull_request:46912", + "pull_request:46920", + "pull_request:46926", + "pull_request:46927", + "pull_request:46928", + "pull_request:46941", + "pull_request:46949", + "pull_request:46950", + "pull_request:46951", + "pull_request:46954", + "pull_request:46957", + "pull_request:46958", + "pull_request:46961", + "pull_request:46965", + "pull_request:46970", + "pull_request:46971", + "pull_request:46974", + "pull_request:46977", + "pull_request:46986" + ], + "soft_pairs": [ + "pull_request:46895|pull_request:46958", + "pull_request:46903|pull_request:46958", + "pull_request:46950|pull_request:46951", + "pull_request:46950|pull_request:46970", + "pull_request:46927|pull_request:46950", + "pull_request:46927|pull_request:46951", + "pull_request:46970|pull_request:46986", + "pull_request:46970|pull_request:46971", + "pull_request:46951|pull_request:46971", + "pull_request:46949|pull_request:46971", + "pull_request:46951|pull_request:46970", + "pull_request:46927|pull_request:46986", + "pull_request:46908|pull_request:46928", + "pull_request:46907|pull_request:46926", + "pull_request:46902|pull_request:46965", + "pull_request:46907|pull_request:46928", + "pull_request:46951|pull_request:46986", + "pull_request:46926|pull_request:46928", + "pull_request:46941|pull_request:46986", + "pull_request:46926|pull_request:46941", + "pull_request:46949|pull_request:46986", + "pull_request:46941|pull_request:46970", + "pull_request:46907|pull_request:46941", + "pull_request:46928|pull_request:46941", + "pull_request:46180|pull_request:46658", + "pull_request:46658|pull_request:46903", + "pull_request:46928|pull_request:46951", + "pull_request:46907|pull_request:46927", + "pull_request:46954|pull_request:46961", + "pull_request:46954|pull_request:46977", + "pull_request:46908|pull_request:46951", + "pull_request:46926|pull_request:46927", + "pull_request:46908|pull_request:46941", + "pull_request:46891|pull_request:46954", + "pull_request:43451|pull_request:46954", + "pull_request:46954|pull_request:46957", + "pull_request:46903|pull_request:46970", + "pull_request:46958|pull_request:46961", + "pull_request:46958|pull_request:46977", + "pull_request:46658|pull_request:46970", + "pull_request:46961|pull_request:46974", + "pull_request:46958|pull_request:46974", + "pull_request:46974|pull_request:46977", + "pull_request:43451|pull_request:46920", + "pull_request:46920|pull_request:46957", + "pull_request:46836|pull_request:46920", + "pull_request:46891|pull_request:46920", + "pull_request:46957|pull_request:46958", + "pull_request:46658|pull_request:46920", + "pull_request:46891|pull_request:46958", + "pull_request:46891|pull_request:46974", + "pull_request:46970|pull_request:46974", + "pull_request:45862|pull_request:46425", + "pull_request:46957|pull_request:46965", + "pull_request:43451|pull_request:46707", + "pull_request:46707|pull_request:46836", + "pull_request:46266|pull_request:46707", + "pull_request:46707|pull_request:46961", + "pull_request:46707|pull_request:46957", + "pull_request:46707|pull_request:46891", + "pull_request:40546|pull_request:43451", + "pull_request:46912|pull_request:46977", + "pull_request:46196|pull_request:46965", + "pull_request:46836|pull_request:46965", + "pull_request:46425|pull_request:46965", + "pull_request:46196|pull_request:46977", + "pull_request:44517|pull_request:46425", + "pull_request:46196|pull_request:46425", + "pull_request:43451|pull_request:46425", + "pull_request:46425|pull_request:46961", + "pull_request:43451|pull_request:46912", + "pull_request:46425|pull_request:46836" + ], + "prepared_review_unit_hash": "f9325fad47da095af63226fcfa4ee6b620ff8889778852b4d3937e949f0a43b0", + "component_budget": { + "node_count": 39, + "item_count": 39, + "soft_pair_count": 113, + "serialized_chars": 131300, + "estimated_input_tokens": 32825, + "estimated_eval_tokens": 65906 + }, + "budget": { + "node_count": 36, + "item_count": 36, + "soft_pair_count": 72, + "serialized_chars": 102500, + "estimated_input_tokens": 25625, + "estimated_eval_tokens": 51506 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: new-model PRs, docs-only updates, CI/workflow changes, and several isolated bugfixes for different models. The only clear duplicate-like pair is the two CohereASR loss-shift fixes; the rest are unrelated.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for the whole cluster: the items span unrelated features and bugfixes across many subsystems, so there is no true duplicate center.", + "best_issue_reason": null, + "best_pr_reason": "PR 46895 is the cleanest narrowly scoped representative of the only real duplicate subcluster: it directly fixes the CohereASR training-loss double-shift with an explicit issue link.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46895", + "right": "pull_request:46958", + "accept": true, + "reason": "Same underlying CohereASR loss double-shift bug on the same code path; the two PRs are alternative implementations of the same fix." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46958", + "accept": false, + "reason": "They overlap on CohereASR, but 46903 is a broader multi-fix PR (also PPFormulaNet) with a different implementation shape, so it is not a clean duplicate." + }, + { + "left": "pull_request:46950", + "right": "pull_request:46951", + "accept": false, + "reason": "Different models and failures: Mamba2 OOM vs PEFT integration load_error." + }, + { + "left": "pull_request:46950", + "right": "pull_request:46970", + "accept": false, + "reason": "Unrelated models and bugs: Mamba2 OOM vs MusicGen-Melody generation mismatch." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46950", + "accept": false, + "reason": "Different model families and failure modes; no shared bug or change." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46951", + "accept": false, + "reason": "GLM46V fix vs PEFT integration fix; unrelated code paths." + }, + { + "left": "pull_request:46970", + "right": "pull_request:46986", + "accept": false, + "reason": "Different models and issues: MusicGen-Melody generation logic vs GLM-OCR model construction." + }, + { + "left": "pull_request:46970", + "right": "pull_request:46971", + "accept": false, + "reason": "MusicGen-Melody output mismatch and Mamba2 OOM are unrelated." + }, + { + "left": "pull_request:46951", + "right": "pull_request:46971", + "accept": false, + "reason": "PEFT integration vs Mamba2 generation OOM; different bugs." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46971", + "accept": false, + "reason": "Whisper/AudioFlamingo cache fix vs Mamba2 OOM; unrelated." + }, + { + "left": "pull_request:46951", + "right": "pull_request:46970", + "accept": false, + "reason": "Different subsystems and failure classes." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46986", + "accept": false, + "reason": "Unrelated model fixes." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46928", + "accept": false, + "reason": "DeepSeekV3.2 load_error test fix vs BLIP-2 tokenizer mapping; unrelated." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46926", + "accept": false, + "reason": "EdgeTam import/config test fix vs MiniMax-M3-VL load_error; unrelated." + }, + { + "left": "pull_request:46902", + "right": "pull_request:46965", + "accept": false, + "reason": "Offloading/save-state bugfix vs DeepSeek V4 cache handling; unrelated." + }, + { + "left": "pull_request:46180", + "right": "pull_request:46658", + "accept": false, + "reason": "Two new-model additions in different domains; not the same bug or change." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46965", + "accept": false, + "reason": "Finegrained FP8 generation optimization vs DeepSeek V4 use_cache handling; unrelated." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46928", + "accept": false, + "reason": "Same broad infra style of test fixes, but different models and failures." + }, + { + "left": "pull_request:46951", + "right": "pull_request:46986", + "accept": false, + "reason": "PEFT integration vs GLM-OCR model init; unrelated." + }, + { + "left": "pull_request:46941", + "right": "pull_request:46986", + "accept": false, + "reason": "DAC output mismatch vs GLM-OCR construction bug; unrelated." + }, + { + "left": "pull_request:46926", + "right": "pull_request:46941", + "accept": false, + "reason": "MiniMax-M3-VL load_error vs DAC tensor mismatch; unrelated." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46941", + "accept": false, + "reason": "Different model families and failure modes." + }, + { + "left": "pull_request:46928", + "right": "pull_request:46941", + "accept": false, + "reason": "BLIP-2 tokenizer mapping vs DAC codec bug; unrelated." + }, + { + "left": "pull_request:46180", + "right": "pull_request:46658", + "accept": false, + "reason": "Repeated pair: still unrelated new-model additions." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46903", + "accept": false, + "reason": "Step-3.7-Flash new-model work vs CohereASR/PPFormulaNet loss fixes; unrelated." + }, + { + "left": "pull_request:46928", + "right": "pull_request:46951", + "accept": false, + "reason": "Tokenizer mapping tweak vs PEFT integration bugfix; unrelated." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46927", + "accept": false, + "reason": "Different model fixes with no shared code path." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46961", + "accept": false, + "reason": "Pytest-v9 compatibility fix vs Serge normalization workflow; unrelated." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46977", + "accept": false, + "reason": "Pytest-v9 fix vs Mllama argument deprecation; unrelated." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46961", + "accept": false, + "reason": "CohereASR loss fix vs workflow normalization; unrelated." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46977", + "accept": false, + "reason": "CohereASR loss fix vs Mllama arg-name cleanup; unrelated." + }, + { + "left": "pull_request:46961", + "right": "pull_request:46974", + "accept": false, + "reason": "Workflow cleanup vs MusicGen typo fix; unrelated." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46974", + "accept": false, + "reason": "CohereASR training-loss fix vs MusicGen generate typo; unrelated." + }, + { + "left": "pull_request:46974", + "right": "pull_request:46977", + "accept": false, + "reason": "MusicGen generate typo vs Mllama deprecation cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46920", + "accept": false, + "reason": "Molmo2 model add vs Voxtral-realtime audio API adaptation; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46920", + "accept": false, + "reason": "Unlimited OCR model add vs Voxtral-realtime API adaptation; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46920", + "accept": false, + "reason": "PEFT CI workflow test branch vs Voxtral-realtime API migration; unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46958", + "accept": false, + "reason": "Finegrained FP8 generation optimization vs CohereASR loss fix; unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46965", + "accept": false, + "reason": "Finegrained FP8 generation optimization vs DeepSeek V4 cache handling; unrelated." + }, + { + "left": "pull_request:46961", + "right": "pull_request:46977", + "accept": false, + "reason": "Repeated pair: unrelated workflow cleanup vs Mllama cleanup." + }, + { + "left": "pull_request:40546", + "right": "pull_request:43451", + "accept": false, + "reason": "VibeVoice model add vs Molmo2 model add; different models." + }, + { + "left": "pull_request:46912", + "right": "pull_request:46977", + "accept": false, + "reason": "RoPE tests/rotary logic work vs Mllama argument cleanup; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46965", + "accept": false, + "reason": "Wav2vec2 SDPA dispatch vs DeepSeek V4 cache semantics; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46965", + "accept": false, + "reason": "Unlimited OCR model add vs DeepSeek V4 cache fix; unrelated." + }, + { + "left": "pull_request:46425", + "right": "pull_request:46965", + "accept": false, + "reason": "RADIO model support vs DeepSeek V4 cache fix; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46977", + "accept": false, + "reason": "Wav2vec2 SDPA dispatch vs Mllama deprecation cleanup; unrelated." + }, + { + "left": "pull_request:44517", + "right": "pull_request:46425", + "accept": false, + "reason": "Qwen3-TTS model add vs RADIO model support; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46425", + "accept": false, + "reason": "Wav2vec2 SDPA dispatch vs RADIO model support; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is heterogeneous, and the only clearly duplicate-like pair is the CohereASR loss-double-shift fix pair (46895/46958). The conservative false verdicts for the other pairs are supported by the diffs and titles, with no obvious overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 35, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-39", + "nodes": [ + "pull_request:43451", + "pull_request:46196", + "pull_request:46425", + "pull_request:46509", + "pull_request:46658", + "pull_request:46836", + "pull_request:46891", + "pull_request:46907", + "pull_request:46908", + "pull_request:46912", + "pull_request:46917", + "pull_request:46926", + "pull_request:46927", + "pull_request:46949", + "pull_request:46950", + "pull_request:46951", + "pull_request:46957", + "pull_request:46961", + "pull_request:46968", + "pull_request:46970", + "pull_request:46971", + "pull_request:46977" + ], + "soft_pairs": [ + "pull_request:46425|pull_request:46891", + "pull_request:46912|pull_request:46957", + "pull_request:43451|pull_request:46196", + "pull_request:46891|pull_request:46912", + "pull_request:46836|pull_request:46912", + "pull_request:46196|pull_request:46961", + "pull_request:46425|pull_request:46509", + "pull_request:46425|pull_request:46917", + "pull_request:46950|pull_request:46971", + "pull_request:46949|pull_request:46950", + "pull_request:46949|pull_request:46951", + "pull_request:46927|pull_request:46949", + "pull_request:46927|pull_request:46971", + "pull_request:46949|pull_request:46970", + "pull_request:46927|pull_request:46970", + "pull_request:46908|pull_request:46926", + "pull_request:46907|pull_request:46908", + "pull_request:43451|pull_request:46836", + "pull_request:46658|pull_request:46836", + "pull_request:43451|pull_request:46658", + "pull_request:46891|pull_request:46957", + "pull_request:43451|pull_request:46891", + "pull_request:46658|pull_request:46957", + "pull_request:46658|pull_request:46891", + "pull_request:46836|pull_request:46957", + "pull_request:46836|pull_request:46891", + "pull_request:46961|pull_request:46977", + "pull_request:46891|pull_request:46968", + "pull_request:46891|pull_request:46961", + "pull_request:46957|pull_request:46968", + "pull_request:43451|pull_request:46957", + "pull_request:43451|pull_request:46968", + "pull_request:46836|pull_request:46968", + "pull_request:46957|pull_request:46961", + "pull_request:46957|pull_request:46977", + "pull_request:46891|pull_request:46977", + "pull_request:46658|pull_request:46968", + "pull_request:43451|pull_request:46961", + "pull_request:43451|pull_request:46977", + "pull_request:46836|pull_request:46961", + "pull_request:46836|pull_request:46977" + ], + "prepared_review_unit_hash": "6d0dbf280bc56fa88add1f8f0cdec435984b153ede268dbcf5bebcf1dc9a04f0", + "component_budget": { + "node_count": 39, + "item_count": 39, + "soft_pair_count": 113, + "serialized_chars": 131300, + "estimated_input_tokens": 32825, + "estimated_eval_tokens": 65906 + }, + "budget": { + "node_count": 22, + "item_count": 22, + "soft_pair_count": 41, + "serialized_chars": 61370, + "estimated_input_tokens": 15343, + "estimated_eval_tokens": 30942 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster: mostly unrelated model-addition PRs plus separate one-off fixes, docs, CI, and test-only changes. I do not see any soft pair that is clearly the same concrete bug/change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "43451 is the broadest, earliest model-addition PR and the closest hub by similarity, but the cluster is too mixed for it to be a true duplicate anchor.", + "best_issue_reason": null, + "best_pr_reason": "43451 is the least-bad representative because it has the largest scope and highest discussion activity, but it still does not match the rest of the cluster as a duplicate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46425", + "right": "pull_request:46891", + "accept": false, + "reason": "Different features: RADIO model support vs PEFT CI workflow changes." + }, + { + "left": "pull_request:46912", + "right": "pull_request:46957", + "accept": false, + "reason": "Unrelated changes: RoPE test coverage vs FP8/generation optimization work." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46196", + "accept": false, + "reason": "Different model families and different changes; model addition vs wav2vec2 SDPA dispatch." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46912", + "accept": false, + "reason": "CI/workflow change vs RoPE tests; not the same bug or code path." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46912", + "accept": false, + "reason": "Unlimited OCR model addition is unrelated to RoPE tests." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46961", + "accept": false, + "reason": "Wav2vec2 attention dispatch vs Serge branch-normalization workflow; unrelated." + }, + { + "left": "pull_request:46425", + "right": "pull_request:46509", + "accept": false, + "reason": "Different model additions (RADIO vs Nemotron Omni), no shared bug/change." + }, + { + "left": "pull_request:46425", + "right": "pull_request:46917", + "accept": false, + "reason": "Model support addition vs deprecation cleanup; unrelated." + }, + { + "left": "pull_request:46950", + "right": "pull_request:46971", + "accept": false, + "reason": "Same model family and symptom, but one changes model code and the other only test loading args; not mergeable as one PR." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46950", + "accept": false, + "reason": "Whisper/audioflamingo cache fix vs Mamba2 OOM fix; unrelated models and code paths." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46951", + "accept": false, + "reason": "Whisper/audioflamingo fix vs PEFT loading fix; unrelated." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46949", + "accept": false, + "reason": "GLM46V input-id fix vs Whisper/audioflamingo cache fix; unrelated." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46971", + "accept": false, + "reason": "GLM46V model fix vs Mamba2 test loading tweak; not the same issue." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46970", + "accept": false, + "reason": "Different models and failures: Whisper/audioflamingo vs MusicGen Melody output mismatch." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46970", + "accept": false, + "reason": "GLM46V vs MusicGen Melody; unrelated." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46926", + "accept": false, + "reason": "Different models and different failure modes; DeepSeek V3.2 load error vs MiniMax-M3-VL OOM test tweak." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46908", + "accept": false, + "reason": "Separate model-specific test fixes for different architectures." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46836", + "accept": false, + "reason": "Both are model additions, but for different models and independent code changes." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46836", + "accept": false, + "reason": "Step-3.7 Flash implementation vs Unlimited OCR addition; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46658", + "accept": false, + "reason": "Molmo2 addition vs Step-3.7 Flash build/conversion work; different model families." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46957", + "accept": false, + "reason": "PEFT CI workflow changes vs FP8/generation optimization; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46891", + "accept": false, + "reason": "Model addition vs PEFT CI workflow; unrelated." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46957", + "accept": false, + "reason": "Step-3.7 Flash model-conversion work vs FP8/generation optimization; unrelated." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46891", + "accept": false, + "reason": "Step-3.7 Flash model work vs PEFT CI workflow; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46957", + "accept": false, + "reason": "Unlimited OCR addition vs FP8/generation optimization; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46891", + "accept": false, + "reason": "Unlimited OCR model addition vs PEFT CI workflow; unrelated." + }, + { + "left": "pull_request:46961", + "right": "pull_request:46977", + "accept": false, + "reason": "Branch-normalization workflow vs Mllama argument deprecation; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46968", + "accept": false, + "reason": "PEFT CI workflow vs docs autolink cleanup; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46961", + "accept": false, + "reason": "PEFT CI workflow vs branch-normalization workflow; different automation tasks." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46968", + "accept": false, + "reason": "FP8/generation code vs docs autolink cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46957", + "accept": false, + "reason": "Molmo2 model addition vs FP8/generation optimization; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46968", + "accept": false, + "reason": "Molmo2 model addition vs docs cleanup; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46968", + "accept": false, + "reason": "Unlimited OCR model addition vs docs cleanup; unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46961", + "accept": false, + "reason": "FP8/generation optimization vs branch-normalization workflow; unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46977", + "accept": false, + "reason": "FP8/generation optimization vs Mllama arg deprecation; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46977", + "accept": false, + "reason": "PEFT CI workflow vs Mllama arg-name cleanup; unrelated." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46968", + "accept": false, + "reason": "Step-3.7 Flash model work vs docs cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46961", + "accept": false, + "reason": "Molmo2 model addition vs branch-normalization workflow; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46977", + "accept": false, + "reason": "Molmo2 model addition vs Mllama arg cleanup; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46961", + "accept": false, + "reason": "Unlimited OCR model addition vs branch-normalization workflow; unrelated." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46977", + "accept": false, + "reason": "Unlimited OCR model addition vs Mllama arg cleanup; 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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "997a7c81311f5231255a739db0b60380965b84a69f284a089ec11f4bdd3dbb81", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They are not duplicates and should not be clustered together for merge triage.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the items address different code paths, different models, and different user-facing capabilities.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:41224 is the better representative only in the sense that it is a tighter, more self-contained feature addition; however, it is not a duplicate of the TimesFM PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models and different functionality. PR 40755 adds TimesFM covariate forecasting across TimesFM docs/code/tests, while PR 41224 adds DINOv3 image-classification support. They do not fix the same bug or implement 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": 3, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41855-13", + "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", + "pull_request:46267" + ], + "soft_pairs": [ + "pull_request:41855|pull_request:43823", + "pull_request:41855|pull_request:43424", + "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:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "323bd23d01e6cc2ca125db73e1e51b687abaf813b90a512018005243863911c7", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34782, + "estimated_input_tokens": 8696, + "estimated_eval_tokens": 17648 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34782, + "estimated_input_tokens": 8696, + "estimated_eval_tokens": 17648 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only one pair looks like a true duplicate-style overlap: the two `fixed_cross_entropy` PRs that both fix ignored cross-entropy kwargs. The rest are broad-similarity matches across unrelated models, docs, or infrastructure changes and should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the clearest representative of the only real overlap: it directly fixes `fixed_cross_entropy` to forward supported kwargs and explicitly tracks issue 43240.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the best canonical PR because it is the more explicit, issue-linked version of the same loss-function fix and is the only duplicate-worthy change in this set.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Different bugs/features: Mistral tokenizer missing methods vs adding MobileLLM model support." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated domains: tokenizer methods vs executorch export dynamic-shape testing." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both are processor/model-loading related only at a very broad level; they fix different models and different issues." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs fix is unrelated to adding a new model package." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different code paths: cross-entropy kwargs vs Mistral4 test/model changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch export tests are unrelated to Mistral4 model/test fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Dynamic-shape export testing is unrelated to MobileLLM model addition." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New model addition and config-resolution bug fix are distinct changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "Different models and different problems: MobileLLM support vs Falcon output-interface refactor." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor is unrelated to a new-model scaffolding skill PR." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output-collection refactor and Mistral4 test fixes do not target the same bug." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes are unrelated to CLIP-like `layrnorm` conversion mapping." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 testing and chat response parsing updates are different subsystems and issues." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 tests are unrelated to adding the TDT loss kernel." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test/model work does not match the `cls.config` resolution-order fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "A model test PR and a model-creation skill PR are not the same underlying change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "New model support and the skill/documentation scaffolding PR are not duplicates." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs fix the same concrete bug in `fixed_cross_entropy`: forwarding supported kwargs to `nn.functional.cross_entropy` for the same loss path." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet. The only accepted overlap, PR 43251 vs PR 43254, is supported by the shared file, shared issue target 43240, and the near-identical `fixed_cross_entropy` kwargs-forwarding fix. The remaining pairwise verdicts stay conservative: they are rejected based on clearly different subsystems/models or only broad soft similarity, without overstating the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "e50f5e32747a82db1c53ad55c057f2cfd0483ce6092c6016d2a5a41a5848d921", + "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": "These are related GitHub Actions maintenance PRs, but they are not the same underlying change: one is a Node 24 compatibility bump for core GitHub Actions, while the other upgrades docker/build-push-action and other workflows to newer releases. Similar area, different concrete fixes.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #42976 is the broader GitHub Actions version-bump PR and better represents the shared workflow-maintenance theme, while #42975 is a narrower compatibility-specific upgrade.", + "best_issue_reason": null, + "best_pr_reason": "PR #42976 is the better cluster anchor because it spans more workflows and is framed as the general latest-version upgrade, making it the most representative single PR here.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Both are workflow dependency upgrades, but they target different action families and different problems: Node 24 compatibility for actions/checkout/cache/upload-artifact vs. docker/build-push-action and other latest-version bumps. They are not the same concrete change 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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "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": "f9bf30b2a4a9854a7a5f296b04092acbded7cb28d06778fd050d023acdd55271", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13041, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13041, + "estimated_input_tokens": 3261, + "estimated_eval_tokens": 6778 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are separate model-addition changes, not duplicates. The shared files are mostly boilerplate docs/auto-registration plumbing, not the same underlying implementation or bug fix.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "45490 is the most central representative if one PR must be chosen: it sits between the others on shared infra/docs touchpoints, but the set is still unrelated overall.", + "best_issue_reason": null, + "best_pr_reason": "45490 is the best cluster anchor only in the sense of centrality and shared plumbing; it is not a true duplicate of the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "ViT NEPA and Molmo are different model additions; overlap is limited to generic docs/auto-registry scaffolding." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model families and code paths; shared model-registration/docs files are boilerplate, not a common change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated architectures; shared files are generic model-addition plumbing." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "CTSM and Step3p7 target different model stacks and modalities; shared auto/docs edits do not indicate the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-7", + "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": "9a569dd0b2396bfb6108857074bb5b55949dc1f822fcd76cf1821cef8e9815e2", + "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": "This cluster is mostly a set of unrelated PRs that all happen to have generic or similar titles. The only strong duplicate-like pair is the two YOLOS fixes, which address the same device-mismatch issue in the same file and target the same tracking issue.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46886 is the best canonical PR because it has the clearest, specific title, explicit issue target, and a focused fix for the YOLOS device-mismatch bug.", + "best_issue_reason": null, + "best_pr_reason": "PR 46886 is the strongest representative of the cluster\u2019s one real duplicate pair; the others are either test-only changes or different subsystems and should not be grouped with it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Both are generic test-fix PRs, but they touch different model test files and appear to fix unrelated test issues." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper test cleanup and FP8 dequant changes are unrelated bugs and code paths." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Whisper test changes are unrelated to generation-config watermarking and generation logic." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46922", + "accept": false, + "reason": "Whisper tokenizer/test changes are unrelated to the YOLOS device-mismatch fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "Model test fixes do not indicate the same underlying bug; one is CLIP/ViT tests, the other is CI/integration setup." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Different subsystems and different failure modes: vision model tests vs FP8 loading/dequant logic." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Test PR for CLIP/ViT is unrelated to generation config and Whisper code." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46922", + "accept": false, + "reason": "The CLIP/ViT test edits do not match the YOLOS runtime/device-mismatch change." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8 quantization/loading changes are unrelated to Phimoe integration-test CI fixes." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "These address different code paths: FP8 dequantization vs generation config validation." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46922", + "accept": false, + "reason": "No common bug or code path; FP8 padding/loading changes are unrelated to YOLOS." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation watermarking validation is unrelated to Phimoe CI/device-memory test setup." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46922", + "accept": false, + "reason": "Generation config validation fixes and YOLOS device handling are separate changes." + }, + { + "left": "pull_request:46539", + "right": "pull_request:46922", + "accept": false, + "reason": "CI memory/device-map test harness changes are not the same bug as the YOLOS model forward/device fix." + }, + { + "left": "pull_request:46886", + "right": "pull_request:46922", + "accept": true, + "reason": "Both PRs target the same YOLOS device-mismatch issue, share the same file and issue target, and the second is a concrete refinement of the same fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: the non-YOLOS PRs are plausibly unrelated despite superficial title similarity, and the only clearly mergeable duplicate-like pair is the two YOLOS fixes sharing the same file and issue target. The soft-edge rejections are conservative, and the one accepted pair looks legitimately combinable 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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "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": "7f95f34d31d74addebacc909990c023192cd474ed9a0efe626f8314b5dc4106c", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6347, + "estimated_input_tokens": 1587, + "estimated_eval_tokens": 3430 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6347, + "estimated_input_tokens": 1587, + "estimated_eval_tokens": 3430 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated synthetic \u201c[don\u2019t merge]\u201d PRs touching different areas (setup/deps, CI/workflows, and a security-gate workflow). They share some broad repository/testing context, but not the same underlying bug or change, so none should be clustered as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#43532 is the broadest/most central artifact in the set by file spread and activity, but it still targets a different change set than the other PRs.", + "best_issue_reason": null, + "best_pr_reason": "#43532 is the best representative only in the sense of being the most central and active PR here; it is not a true duplicate of #43488 or #46747.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:46747", + "accept": false, + "reason": "Different changes: setup/dependency/test-bot checks versus a PR security-gate workflow tweak. Shared repo/testing context is too broad to treat as the same fix." + }, + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "#43488 is a setup/dependency/version test-bot PR, while #43532 changes CI workflows and helper utilities. They overlap only on a Vit test file, not the same code-path problem." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46747", + "accept": false, + "reason": "Both touch workflows, but they modify different workflows for different purposes. They would not plausibly merge into one PR 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": 9, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "354c5f34f26384c3695ced6baa09e572fcbef94b108ad4947c6474295c9aa2c9", + "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 PRs are related in theme and scaffolding, but they add different pipelines for different segmentation tasks, so they should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43613 is the stronger representative: it adds the Promptable Visual Segmentation pipeline with broader model coverage and test support, making it the more complete change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43613 is the best standalone PR in this cluster because it is more comprehensive and better covers the intended code path and integrations.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different underlying features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They share pipeline/docs scaffolding, but target different tasks and backends, so they are not 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": 10, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "2a1ef1769f78e69bec1e3e8a40b004b9ca6b86907bc0427f1149f9289b7216c2", + "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 touch `trainer.py`, but they address different trainer extension points: one fixes gradient-accumulation loss scaling, the other refactors signature-column inspection for PEFT/custom column filtering. They are related to customization in Trainer, but not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete representative of the cluster because it fixes a specific training-step bug in loss scaling and introduces the corresponding hook/property in one focused change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the best overall representative here: it is a self-contained bugfix with a clear behavioral change, whereas PR #43663 is a separate refactor for column-signature handling.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Both modify `trainer.py`, but they change different code paths and solve different problems: GA loss scaling vs. signature column filtering. They would not plausibly be merged as 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": 11, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "1596ff57e1eedcccf97d0096e418874cb2e49231d721e2f8a131f6ce6c4dab7a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch `src/transformers/cli/serve.py`, but they fix different `TypeAdapter` annotation/runtime issues: one is a broader typing/postponed-annotations cleanup, the other specifically guards `TypeAdapter` for missing `pydantic`.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43656 is the better cluster representative because it has more activity and an explicit linked issue, and its fix is the broader CLI `TypeAdapter` NameError report.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the strongest standalone fix because it targets a concrete runtime failure when `pydantic` is absent, with a minimal, focused change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Reject: these are not the same underlying bug. #43656 adds postponed-annotation typing changes to address a CLI NameError, while #43842 specifically makes `TypeAdapter` a TYPE_CHECKING-only import to avoid failures when `pydantic` is not installed." + } + ] + }, + "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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44018|pull_request:44071", + "pull_request:44059|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44059|pull_request:44071", + "pull_request:44056|pull_request:44101", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44068", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:43997|pull_request:44018", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43997|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "1685ef7a2234eb5ede85f1681c1bc275d6f44928442fb8ad74b4e4c4a8c69868", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a batch of model-specific output-tracing refactors under the shared tracker #43979, not one duplicate bug. Only PRs that touch the same model/code path and the same tracing refactor are duplicates; cross-model pairs should stay separate.", + "confidence": 0.9, + "canonical_issue_reason": "No standalone issue artifact is present in the cluster; the common tracker referenced by many PRs is #43979.", + "canonical_pr_reason": "PR #44066 is the strongest canonical PR for the GPT-J/CodeGen subset: it has the explicit tracker, broader scope, and the clearest overlap with the other GPT-J tracing PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR #44066 is the best representative PR in this cluster because it is explicit, substantive, and closest to the other true duplicates in the GPT-J family.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths; same issue target is just a shared tracking issue." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model families; only the output-tracing theme is shared." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor on the same file; these are plausibly mergeable into one PR." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Different backbones and implementations; same tracker does not make them duplicates." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen and EfficientNet are different model refactors." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa V2 are unrelated model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J vs MGP-STR; same general refactor theme only." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors in the same file and look mergeable together." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different models." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model-specific refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated despite the shared tracing pattern." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different models; same tracker is not enough." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are different code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model-specific refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are unrelated model implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are not the same underlying change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different models." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model families." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are unrelated code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are different model refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBERT are unrelated models." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different model-specific changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are separate model refactors." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-2 and GPT-Neo are different models." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J/CodeGen are unrelated changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are different model-specific refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBERT are not the same underlying change." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated model refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBERT are different models." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo are separate code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa V2 are unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44074", + "accept": false, + "reason": "SqueezeBert and TextNet are different model-specific changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa V2 and TextNet are unrelated code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 and SpeechT5 are different model families." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 and VITS are unrelated implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing refactor family; overlapping file and change intent make them mergeable." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J tracing refactors on the same code path, so they look like duplicates." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different model code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet/ResNet are different model refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated model implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 and SpeechT5 are different underlying changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 and VITS are unrelated model-specific refactors." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it correctly frames the cluster as mostly model-specific output-tracing refactors under shared tracker #43979, and the accepted soft-edge pairs are all within the same model/code path with plausible single-PR mergeability. No pair is overclaimed across unrelated model families." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 35, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "6296e5f7064674c4ad2e4ba1580d861792b7fae146fdc87f5ea3f44981e73edc", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17127, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs are thematically related by a shared refactor pattern around standardized output tracing/capture, but they touch different model implementations and code paths, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:44129 is the best single anchor for the cluster because it has the clearest issue linkage and a concrete end-to-end output-tracing refactor, but it still represents only one model-specific instance of the broader pattern.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:44129 is the strongest representative PR: it is issue-linked, has a focused scope, and shows the standardized output-capture change most explicitly.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized-output refactors, but they target different models (FocalNet vs ImageGPT) and different internal execution paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Similar decorator-based refactor, but these are separate model implementations (VisualBert vs TextNet) with distinct forward logic." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Same general output-tracing theme, but MPNet and VisualBert are unrelated code paths and not plausibly one merged fix." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "Both are refactors for output capture, but they modify different models and should not be merged as one duplicate change." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated model families; the similarity is only at the refactor pattern level." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "Both adopt standardized output tracing, but they affect different model stacks (SqueezeBert vs SpeechT5) and different concrete layers." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44028", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "757e63c33ca9ba911bf249c8631ec4e466c640d541f707d81ec1624f400bc842", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40178, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40178, + "estimated_input_tokens": 10045, + "estimated_eval_tokens": 20346 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of related-but-distinct output-tracing refactors across many different Transformer models. They share a naming pattern and some helper plumbing, but they are not duplicates of the same bug or change.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the best representative because it is the broadest of the group, touching two closely related Mamba-family files and adding the new output-capture plumbing; it best captures the cluster\u2019s theme, though it is not a duplicate of the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the strongest cluster representative for the same reason: it has the widest scope and most central refactor shape among these output-tracing PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths (timm_backbone vs upernet); same refactor style, not the same change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Unrelated architectures (timm_backbone vs vision_text_dual_encoder); too broad a similarity to treat as one fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different models and behavior surfaces (timm_backbone vs univnet); not one concrete bug or mergeable PR." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Both are tracing refactors, but they target different model stacks and code paths; not the same underlying change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "timm_backbone and rwkv are unrelated implementations; similarity is only the output-tracing pattern." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "MobileNetV1 and UnivNet are unrelated models; shared helper imports do not make this one duplicate." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model code paths (MobileNetV1 classification vs UperNet segmentation); not mergeable as one PR." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "MobileNetV1 and Depth Anything fix different architectures and output APIs; only the refactor theme matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "MobileNetV1 and VisionEncoderDecoder are different code paths; not the same concrete bug/change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different wrappers and modalities (image classifier vs speech encoder-decoder); too broad to merge together." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Vision-text dual encoder and UnivNet are unrelated models; same helper style, different fixes." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and forward paths; no shared concrete bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Vision-text dual encoder and Mamba-family refactors are separate model stacks, so this is not a duplicate." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different tasks and architectures (vision-text vs depth estimation); not the same underlying change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Unrelated model code paths (audio generation vs segmentation); same refactor motif only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and implementations; too broad to consider one fix." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Unrelated models and output contracts; not a single bug or patch." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet segmentation and RWKV language modeling are unrelated code paths; not a duplicate." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model stacks (segmentation vs Mamba language models); same theme, not same change." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Mamba-family output tracing and SuperPoint feature extraction are different implementations; not one concrete fix." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different code paths and model classes (Mamba vs CodeGen); the overlap is only in refactor style." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "timm_backbone and CodeGen are unrelated models; same output-tracing pattern is not enough for duplicate status." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen are different model families with distinct forward paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "UperNet segmentation and CodeGen generation are unrelated; not the same underlying bug/change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder and CodeGen have different code paths and outputs; only the refactor theme matches." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Depth Anything and VisionEncoderDecoder are different model families; not mergeable as one PR." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Depth estimation and speech encoder-decoder are unrelated changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models and output semantics (depth estimation vs SuperPoint); not the same fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Depth Anything and RWKV are unrelated code paths; only the output-tracing refactor pattern overlaps." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Depth estimation and DPR retrieval models are different implementations; not a duplicate." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "VisionEncoderDecoder and RWKV are unrelated model stacks; same helper refactor, different changes." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "VisionEncoderDecoder and SpeechEncoderDecoder are similar wrapper names but different modalities and code paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model types and outputs; not the same concrete bug or mergeable change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "VisionEncoderDecoder and DPR address different components and APIs; not duplicates." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "SpeechEncoderDecoder and SuperPoint are unrelated models; similarity is only superficial." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Speech encoder-decoder and RWKV are different forward paths and output contracts." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Different architectures and behaviors (speech seq2seq vs retrieval encoder); not one bug fix." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint feature extraction and RWKV language modeling are unrelated code paths." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "SuperPoint and DPR are different model families; not the same underlying change." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are unrelated implementations; same refactor motif does not imply duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "2d3fbe7891b78329a12f27bdeba09b30b5bd479c29904fe286ccfa42649bff43", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated model-addition PRs that just share the same Transformers integration pattern and some auto-registration files. PenguinVL and Voxtral TTS implement different models, so they should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45401 is the more canonical standalone change: it has an explicit issue target, a narrower and well-scoped Voxtral TTS implementation, and clear intent tied to a specific model release.", + "best_issue_reason": null, + "best_pr_reason": "PR #45401 is the better representative PR for this cluster because it is tightly scoped, linked to a tracking issue, and clearly centered on one concrete model integration.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different underlying features: PenguinVL and Voxtral TTS are separate model implementations with different codepaths, tests, and docs. Shared auto-registration and docs scaffolding are not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44965-2", + "nodes": [ + "pull_request:44965", + "pull_request:46845" + ], + "soft_pairs": [ + "pull_request:44965|pull_request:46845" + ], + "prepared_review_unit_hash": "c56d0c6abf629bed27687fc470d8fa944f394d4978541e9085f46eef3b61b973", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4219, + "estimated_input_tokens": 1055, + "estimated_eval_tokens": 2366 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4219, + "estimated_input_tokens": 1055, + "estimated_eval_tokens": 2366 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs, but they do not look like the same underlying change. One is a draft-conversion/documentation workflow experiment; the other is a CI branch/workflow tweak with an unrelated placeholder code edit.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No issue is present. Of the two PRs, #46845 is the better representative only because it has a clearer CI-oriented scope and a more coherent diff; #44965 looks like a separate workflow experiment with no clear problem statement.", + "best_issue_reason": null, + "best_pr_reason": "PR #46845 is the strongest standalone candidate in the cluster, but only marginally; it is more explicit about a CI/runner experiment than #44965. Still, it should not be grouped with #44965 as a duplicate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44965", + "right": "pull_request:46845", + "accept": false, + "reason": "Different changes: #44965 rewrites a docs workflow into a draft-conversion workflow, while #46845 changes CI branch routing and includes an unrelated placeholder edit in model code. They do not fix the same concrete problem and would not plausibly 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": 17, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "e0a5d2287d07b32937993cbf5eaaec6c047bdc596335b0cf55214a479c2b928d", + "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 PRs are related by the broad theme of fusion, but they target different changes: #44979 adds a new Module Fusion API and tests, while #45363 integrates n-to-1 kernel fusion through KernelConfig across different integration points. They are not the same underlying change and should not be deduplicated.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "#44979 is the more foundational, self-contained API addition (new `module_fusion.py` plus tests), so it is the better anchor for this cluster, even though #45363 is a separate integration PR.", + "best_issue_reason": null, + "best_pr_reason": "#44979 is the clearest representative because it introduces the core module-fusion abstraction rather than a specific downstream KernelConfig use case.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Both are fusion-related, but they implement different concrete features in different code paths: one adds a Module Fusion API, the other wires n-to-1 fusion through KernelConfig. They are not plausibly mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "e2d48929dd036853cd6f6f7d729cce6d3048a78363fb4a142d9b0b4ce3881d27", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs patch the same AttributeError in `_patch_mistral_regex` with the same code change; the later PR adds tests and is the stronger canonical candidate.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45317 is the better canonical PR because it contains the same fix plus test coverage, making it the more complete implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR #45317 is the best overall PR candidate: same underlying bug, same affected file and issue target, and it includes tests verifying the fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError fix in the same function and file, with effectively identical patch intent; the second PR appears to be an expanded version with tests." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "bdab1c16b4c19e920a55fe430161fe22235ee3ec9509194363ff43bdf5e4b18f", + "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": "These are related NemotronH PRs but not duplicates: one is a refactor to make NemotronH inherit GraniteMoeHybrid and adjust conversion mappings, while the other introduces the proper NemotronH plus dense/sparse variants, docs, and tests. They overlap in files, but they target different changes and would not plausibly be merged as one fix.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the more complete and canonical model-architecture change: it adds the new NemotronH dense/sparse model split, documentation, auto-mappings, and tests, whereas PR 45115 is a narrower refactor/conversion-mapping adjustment.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 is the best representative PR for this cluster because it describes the larger intended feature work and touches the full set of user-facing and model-definition changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Shared NemotronH files alone are not enough: 45115 is a refactor/conversion-mapping patch, while 45569 is a broader model-split implementation with docs and tests. Different concrete changes, so not a duplicate pair." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "ca3931663cc2f3b28f07668d44e4fac69c5c66e132259b67f21b108a5bdfb2ce", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6091, + "estimated_input_tokens": 1523, + "estimated_eval_tokens": 3302 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6091, + "estimated_input_tokens": 1523, + "estimated_eval_tokens": 3302 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related but not duplicates: both introduce heterogeneous configuration docs and config utilities, but 45332 also adds the full modeling/runtime pieces (cache, masking, skip, supported models, mxfp4), while 45333 is configuration-only.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "45332 is the broader umbrella change: it includes the heterogeneous config work plus the modeling/runtime integration that 45333 does not.", + "best_issue_reason": null, + "best_pr_reason": "45332 is the best representative for the cluster because it is the more complete end-to-end heterogeneous model support PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Shared docs and config helpers are not enough here; 45332 adds substantial modeling, cache, and masking changes that make it a different, larger feature rather than the same mergeable PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "99c7b366702fd8e2159db3abcd512512241a134ce4c3107c6f06a768ca7f22b5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are typing-focused PRs, but they address different changes: PR 45415 is a broad type-checking sweep/configuration update across many files, while PR 45425 adds shared typing protocols and typing support around modeling_utils/integrations. They are related in theme but not duplicates of the same code change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the better canonical representative because it introduces reusable typing abstractions in core typing helpers and then applies them to specific integration/modeling code, which is a more coherent single change than the broad repo-wide type-checking sweep in 45415.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best overall fit for the cluster because it is the more focused, self-contained typing change. PR 45415 is broader and mostly a repository-wide static typing pass, making it less representative of a single underlying fix/change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Both are about typing, but 45415 is a large type-checking pass/config update across many files, while 45425 adds specific typing protocols and targeted modeling/integration support. They do not appear to fix the same concrete problem or be mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "6a75c4b755f4192839f7fdcdd46ad53ac4f96a591f75e34ce5130a7b6330d4d7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests concern expert-parallel/distributed model loading, but they fix different failure modes: PR 45548 targets accelerate launch with DeepSpeed ZeRO-3 and config serialization, while PR 45662 targets FSDP2/rank-0 broadcast overwriting experts during load.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR exists here; if one must be chosen as the cluster representative, PR 45662 is the stronger canonical candidate because it addresses a specific concrete load-path overwrite bug with a larger, more focused fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 45662 is the better representative for the cluster because its bug description is more concrete and its code changes are centered on the actual parameter-loading path, whereas PR 45548 is a separate accelerate/ZeRO-3 compatibility fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Do not merge: they address different distributed-loading bugs and different code paths. 45548 is about EP + DeepSpeed ZeRO-3 loading via accelerate launch and config serialization; 45662 is about FSDP2 expert tensors being overwritten by rank-0 broadcast during parameter loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45630-3", + "nodes": [ + "pull_request:45630", + "pull_request:46831", + "pull_request:46978" + ], + "soft_pairs": [ + "pull_request:45630|pull_request:46978", + "pull_request:46831|pull_request:46978" + ], + "prepared_review_unit_hash": "7a8b888ba53cd4ea88654c8119d139b531928582c56b5fdeeaa19ab6ddb39f93", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 9370, + "estimated_input_tokens": 2343, + "estimated_eval_tokens": 4942 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 9370, + "estimated_input_tokens": 2343, + "estimated_eval_tokens": 4942 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "PR 46978 is the broad upstream fix for accelerate hook firing across linear-attention models; PR 46831 is a narrower Qwen-specific application of the same underlying hook problem, while PR 45630 is unrelated model addition work.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46978 is the best canonical PR because it generalizes the accelerate-hook fix across affected models and subsumes the Qwen-specific hook workaround.", + "best_issue_reason": null, + "best_pr_reason": "PR 46978 is the strongest representative of the cluster since it targets the core device_map/offload hook bug directly and applies it broadly rather than only to one model family.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45630", + "right": "pull_request:46978", + "accept": false, + "reason": "Kimi2-6 model integration is unrelated to accelerate hook firing or linear-attention offload behavior." + }, + { + "left": "pull_request:46831", + "right": "pull_request:46978", + "accept": true, + "reason": "Both address the same underlying bug: accelerate hooks not firing when Qwen linear-attention code accesses child module weights directly under CPU offload/device_map." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45630 is clearly a model-addition PR with no evidence of the hook/offload bug, and 46978 is the broader accelerate-hook fix that also touches Qwen files. The 46831\u219446978 pairing is conservative enough because both patches address the same device_map/CPU-offload hook-firing issue, with 46978 broadening the fix. No evident overstatement." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "7781730a063d57a95443096d2217e47e2170e9b497a020c65a1e8b6fbd2c0e5e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist-related captured_info/logging collision problem around `testing_utils.py`; one is a narrower helper fix, the other includes the helper fix plus CI artifact aggregation to handle per-worker files.", + "confidence": 0.8, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the broader, more complete fix: it makes the log output xdist-safe and also updates CI/reporting to collect multiple `captured_info` files, so it best represents the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the best single PR to keep because it covers the core bug and the downstream CI log-preservation behavior, while 45639 is just the narrower helper-side portion.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "They target the same underlying xdist collision in captured_info handling, share the same issue target and files, and 45645 is effectively an expanded version of the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "b9e4e68a2e2a13d8dc3ebc36225a54f32a1746c9ea49c16659aab820871e0f57", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7234, + "estimated_input_tokens": 1809, + "estimated_eval_tokens": 3874 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7234, + "estimated_input_tokens": 1809, + "estimated_eval_tokens": 3874 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated linter-driven PRs overlap on model files but fix different code paths: TRF003 adjusts forward/return_dict behavior, while TRF018 refactors weight initialization. They should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the changes target different modeling concerns (forward-signature handling vs weight initialization) and are not the same underlying fix.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a strong global duplicate representative because they address different rules and different code paths; the overlap is only broad subsystem/file scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Reject: PR 45720 implements TRF003 forward/return_dict tuple handling across model forwards, while PR 46259 implements TRF018 weight-initialization cleanup. Shared filenames are incidental; these are not the same bug/change and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "40be774674dcfd128e5189812d09598da14f0fac616f89fcc6ab8f27f3da5693", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open WIP PRs, but they target different changes: one is a CI/infra + model/config tweak bundle, the other is a test migration adding kernel/model tests. They are not the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45932 is the broader, more substantive change set in the cluster, but it is not a duplicate of #46100; the two PRs address different code paths and purposes.", + "best_issue_reason": null, + "best_pr_reason": "PR #45932 is the better representative only because it is the more general PR, but it should not be merged with #46100 as a duplicate.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different goals and edits: #45932 changes CI secrets and multiple model/config files, while #46100 adds kernel tests and test migration files. Shared test-related context is not enough to treat them as the same fix/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "fa0f4a8cd54fdefc3e1ac2c7cd88acc3e6893768312863b348eb884b4ab79642", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: one PR is a focused EP/trainer integration, the other is a broad distributed-branch base touching many unrelated areas. Shared filenames alone are not enough to treat them as the same change.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 46126 is a targeted EP/trainer feature change, while 46269 is a much broader distributed base branch with additional FSDP, CI, modeling, and integration work.", + "best_issue_reason": null, + "best_pr_reason": "46126 is the better representative if one PR must be chosen because it has a single concrete purpose and tighter scope; 46269 is too umbrella-like to serve as the same-change representative.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Shared distributed files do not imply duplication here. 46126 is EP/trainer integration around tensor parallel routing; 46269 is a broad distributed branch base with many unrelated edits, so they are not the same underlying fix/change and should 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": 29, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "ef2ac52c548ecf85c89cf474aba7b941d4730b56db230e084cc88dc0b58d4abc", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26671, + "estimated_input_tokens": 6668, + "estimated_eval_tokens": 13592 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 31, + "serialized_chars": 26671, + "estimated_input_tokens": 6668, + "estimated_eval_tokens": 13592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly unrelated workflow-maintenance PRs. The only close-looking set is the pr-ci-security-gate.yml edits, but they change different behaviors (logic, permissions, comments, and secrets handling) rather than the same underlying fix, so I would not merge any of the soft edges.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46554 is the best anchor for the only substantive mini-cluster: it makes the actual pr-ci-security-gate behavior change (checkout + workflow-file checks), while the other similar PRs are distinct follow-ups or metadata/comment tweaks.", + "best_issue_reason": null, + "best_pr_reason": "PR 46554 is the strongest overall representative in this set because it changes runtime security-gate logic, not just YAML permissions, comments, or unrelated workflow wiring.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different problems: self-comment CI stdin handling vs pr-ci-security-gate security logic." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows; one updates comment-pipe handling, the other changes PR CI gate permissions/approval flow." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows: stdin piping for a script vs a grep/printf tweak in build-ci image logic." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "SSH-runner hardening is unrelated to benchmark_v2 dispatch behavior." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "SSH-runner hardening is unrelated to anti-slop permission placement." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "SSH-runner hardening is unrelated to assign-reviewers checkout ref changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and different security concerns; not the same concrete change." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and different fixes; one hardens SSH runner secrets, the other adjusts PR CI gate permissions." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Anti-slop permissions and benchmark_v2 workflow_dispatch are unrelated." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Unrelated workflow edits: anti-slop permissions vs assign-reviewers checkout ref." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Unrelated workflow edits: anti-slop permissions vs build-ci image tagging shell change." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different purposes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Unrelated workflow edits: reviewer assignment vs benchmark trigger conditions." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Unrelated workflow edits: reviewer assignment vs build-ci image tagging shell change." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different purposes." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Benchmark trigger logic and build-ci image tagging are different code paths and different issues." + }, + { + "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 workflows; build-ci shell safety tweak is unrelated to PR CI security-gate logic." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Build-ci image tagging and PR CI gate permissions/approval are distinct changes." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same workflow file, but different fixes: 46554 adds checkout plus file-change checks; 46558 adjusts permissions and approval gating." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "46554 changes security-gate behavior; 46563 only adds explanatory notes about the reusable workflow." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46716", + "accept": false, + "reason": "46554 adds gate checks; 46716 only changes secrets passing to the reusable workflow." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Permissions/approval gating changes are not the same as comment-only notes about the reusable workflow." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46716", + "accept": false, + "reason": "Different changes in the same file: permissions/approval gating vs secrets inheritance." + }, + { + "left": "pull_request:46563", + "right": "pull_request:46716", + "accept": false, + "reason": "Comments about workflow safety are not the same change as altering secrets passing." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows 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": 30, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": 31, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46281-2", + "nodes": [ + "pull_request:46281", + "pull_request:46890" + ], + "soft_pairs": [ + "pull_request:46281|pull_request:46890" + ], + "prepared_review_unit_hash": "b7d582f59348856da747f45a18005d2bb49b426ba2df8a206f6d941b967d1498", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4280, + "estimated_input_tokens": 1070, + "estimated_eval_tokens": 2396 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4280, + "estimated_input_tokens": 1070, + "estimated_eval_tokens": 2396 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are both security hardening around path traversal, but they target different loaders and different inputs: one validates shard filenames from a checkpoint index (`weight_map`), the other validates `transformers_weights` in config. They are related but not the same change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: these are distinct fixes in different code paths, so neither cleanly subsumes the other.", + "best_issue_reason": null, + "best_pr_reason": "PR #46890 is the better cluster representative only in the sense that it is a clear, self-contained hardening change; however, it is still not a duplicate of #46281.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46281", + "right": "pull_request:46890", + "accept": false, + "reason": "Different attack surfaces and code paths: shard file resolution from an index vs explicit `transformers_weights` config handling. Same broad security theme, but not the same concrete bug or mergeable 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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "85cc4b9ce142026519acbb0eb60a903715426b1b3fb6a543d08d3004d00b1e3e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5959, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5959, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one adds multimodal continuous batching support, the other adds the HunYuan VL model and docs. No duplicate or mergeable overlap is evident.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46417 is the better standalone anchor: it is a coherent model-addition PR with docs/tests, while PR 46376 is a draft, broader infrastructure change and not related to the same code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different features and code paths: continuous batching/multimodal runtime changes versus adding a new HunYuan VL model package and documentation. They are 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": 33, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46574-2", + "nodes": [ + "pull_request:46574", + "pull_request:46580" + ], + "soft_pairs": [ + "pull_request:46574|pull_request:46580" + ], + "prepared_review_unit_hash": "030455be56c860aadaf2f2db78eef6d19271f87accb148fbc6e38489e7afff72", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated pull requests: one fixes CodeLlama decode stripping leading spaces, the other fixes DeepSeek-Coder/DeepSeek-LLM tokenizer class selection. They do not look like duplicates and should not be clustered together as the same change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46574 is the better canonical representative only because it is a self-contained, concrete tokenizer behavior fix with a clear code-path and expected decode regression to verify.", + "best_issue_reason": null, + "best_pr_reason": "PR #46574 is the strongest standalone PR in this set: it fixes a specific user-visible decoding bug in one code path and is narrowly scoped with an explicit regression target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46574", + "right": "pull_request:46580", + "accept": false, + "reason": "Different bugs and different code paths: CodeLlama decode stripping vs. model-type/tokenizer-class misclassification for DeepSeek checkpoints. They are not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 34, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46603-2", + "nodes": [ + "pull_request:46603", + "pull_request:46604" + ], + "soft_pairs": [ + "pull_request:46603|pull_request:46604" + ], + "prepared_review_unit_hash": "b99bf71d816db838d7c3581e7889318b7e2dc4827e51abf2352f00e010a3255a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5496, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5496, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two open PRs are related to Mistral tokenization infrastructure, but they address different changes: #46603 refactors/moves the existing MistralConverter into an integrations package, while #46604 adds tekken tokenizer detection, conversion, and save utilities. They overlap in files because they touch the same subsystem, not because they fix the same underlying problem.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46604 is the broader, user-facing feature change in the Mistral tokenizer stack; PR #46603 is a packaging/refactor move of MistralConverter. They are adjacent but not the same change.", + "best_issue_reason": null, + "best_pr_reason": "PR #46604 is the best representative of the cluster because it introduces the substantive new tokenizer behavior, whereas #46603 is a structural relocation of existing code.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46603", + "right": "pull_request:46604", + "accept": false, + "reason": "Different concrete changes: one moves MistralConverter into an integrations package, the other adds tekken tokenizer detection/conversion/save utilities. Shared Mistral files do not imply the same underlying bug or mergeable PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 35, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46933-2", + "nodes": [ + "pull_request:46933", + "pull_request:46942" + ], + "soft_pairs": [ + "pull_request:46933|pull_request:46942" + ], + "prepared_review_unit_hash": "a5ce36c911e14dcecb4138cf6ca14d429c21d974700b2c91acf9df99b7384b75", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4372, + "estimated_input_tokens": 1093, + "estimated_eval_tokens": 2442 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4372, + "estimated_input_tokens": 1093, + "estimated_eval_tokens": 2442 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs fix the same general quantized-weight dtype-casting pattern, but they target different model implementations and code paths (Gemma4 vs TimesFM 2.5), so they are not duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46933 is the stronger representative of the cluster because it has an explicit issue link and a clear, narrowly scoped fix, but it is still not a duplicate of PR 46942.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46933", + "right": "pull_request:46942", + "accept": false, + "reason": "Same broad dtype-cast/quantization theme, but they patch different model families and different classes of modules; they do not appear to fix the same concrete bug or be mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/manifest.json b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..9be09ed94401457b6b57b2f0556a04523d8301fa --- /dev/null +++ b/snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260701t060031z", + "artifacts": { + "hybrid": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-07-01T06:12:29Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260701T060031Z", + "variant": "hybrid" +} diff --git a/snapshots/20260701T060031Z/manifest.json b/snapshots/20260701T060031Z/manifest.json index 0406f0299fead186dddaf7cfe37884045141bf59..49d43b7243c491b6fcf4e8c6471a3adc23f4d42f 100644 --- a/snapshots/20260701T060031Z/manifest.json +++ b/snapshots/20260701T060031Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 680 }, "extracted_at": "2026-07-01T06:00:31Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260701t060031z", + "runs": { + "hybrid-model-20260701t060031z": { + "analysis_id": "hybrid-model-20260701t060031z", + "artifacts": { + "hybrid": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260701T060031Z/analysis-runs/hybrid-model-20260701t060031z/manifest.json", + "model": null, + "published_at": "2026-07-01T06:12:29Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260701T060031Z", "watermark": {