diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index d3cb3d9de91e395554c20f24d8ae9f1c5ac9ef4f..74e7fd0732101fc56955e399da96a22210c55f0c 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": "20260511T060028Z", - "generated_at": "2026-05-11T06:10:38Z", + "snapshot_id": "20260511T120023Z", + "generated_at": "2026-05-11T12:10:24Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1683,281 +1683,6 @@ ] } ] - }, - { - "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": [], @@ -2024,15 +1749,6 @@ "target_issue_number": 43979, "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." }, - { - "cluster_id": "cluster-44018-2", - "canonical_pr_number": 44068, - "duplicate_pr_numbers": [ - 44018 - ], - "target_issue_number": null, - "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-45081-3", "canonical_pr_number": 45317, diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 0a117c28221c648b54bdcc7fb0913cea2be883ae..331811966ca4443c0b6ab85aabacee0a44001d0a 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": "20260511T060028Z", - "generated_at": "2026-05-11T06:10:38Z", + "snapshot_id": "20260511T120023Z", + "generated_at": "2026-05-11T12:10:24Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -26,7 +26,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "74713450dc2b8aaf7469bc63c3acfb7f0f54ea42f6b3b3648b2b36ad13a119f6", + "prepared_review_unit_hash": "7d84f1fbd8936e94fa9a48cec974049e4b6fa203f37cfeb08274c64d8e2ab352", "component_budget": { "node_count": 3, "item_count": 3, @@ -48,24 +48,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues do not look like duplicates: one is about Tapas tokenization, one about a private-model auth/download error, and one about loading bert-base-uncased. The cluster appears heterogeneous, so no soft pair should be merged.", - "confidence": 0.34, - "canonical_issue_reason": "If a single anchor is needed, 16618 is the broadest issue title ('Can't load the model...'), but it is not actually the same bug as the other two.", + "summary": "This is not a duplicate cluster: the three issues concern different problems (Tapas tokenization mismatch, private-model auth/error messaging, and inability to load bert-base-uncased). The soft similarities are too broad to justify merging.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "16618 is the most generic and central-looking issue of the set, but the titles point to different failure modes, so this cluster should likely be split rather than deduplicated.", + "best_issue_reason": "#16618 is the broadest and most user-visible model-loading failure and has the most discussion, so it is the best representative only if a single issue must be kept; however, it is not the same bug as the other two.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tokenization mismatch with TensorFlow code is unrelated to a model-loading failure; not the same underlying bug." + "reason": "Tokenization discrepancy in Tapas vs TensorFlow code is a different bug from a generic bert-base-uncased loading failure." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model login/error-message behavior is a different auth/download problem from a general 'can't load bert-base-uncased' report." + "reason": "Private-model login/error-message handling is an auth/download UX issue, not the same underlying model-loading defect." } ] }, @@ -189,8 +189,8 @@ "issue:30277|issue:36331", "issue:34928|issue:36010", "issue:43039|issue:44805", - "issue:43901|issue:44512", "issue:41762|issue:43596", + "issue:43901|issue:44512", "issue:39290|issue:41720", "issue:44841|issue:45823", "issue:43039|issue:44560", @@ -209,130 +209,130 @@ "issue:43541|issue:43828", "issue:36010|issue:39290" ], - "prepared_review_unit_hash": "35309c9b5e1456f8d78933c21fc35941728f65ddf1f50df83641d75e92a25942", + "prepared_review_unit_hash": "78fbaba6baf415ac7a86665ff31acdbace7bfa3ee928980fd9f9aaca8fd423c9", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 69, - "serialized_chars": 48318, - "estimated_input_tokens": 12080, - "estimated_eval_tokens": 24416 + "serialized_chars": 48321, + "estimated_input_tokens": 12081, + "estimated_eval_tokens": 24418 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are only loosely related by language-model/runtime errors and are not true duplicates. The only clear duplicate pair is the DeepSpeed ZeRO-3/BERT index-out-of-bounds report; the rest differ by model, subsystem, or failure mode.", - "confidence": 0.92, - "canonical_issue_reason": "issue:43638 is the broader, clearer representative of the accepted ZeRO-3/BERT index-out-of-bounds bug, so it works best as the canonical issue.", + "summary": "This cluster is mostly heterogeneous: several unrelated issue families are mixed together by loose text similarity. The only clear near-duplicate I found is the ZeRO-3/Bert index-out-of-bounds pair; the rest are different bugs, docs cleanups, or model-specific failures.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43596 is the clearest representative of the ZeRO-3/Bert size-0 index bug and is nearly identical to 43638.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43638 best represents the duplicate pair because it describes the general non-pretrained BertModel + DeepSpeed ZeRO-3 failure without narrowing the reproduction too much.", + "best_issue_reason": "issue:43596 has the cleanest, most specific reproduction for the only strong duplicate family in the cluster, so it is the best issue anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different Qwen3.5 failures: config attribute error vs streaming chat/completions runtime issue." + "reason": "Different Qwen3.5 problems: config attribute lookup vs streaming chat/completions failure." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different problems: AMD GPU model runtime failure vs MPS histogram op not implemented." + "reason": "Different hardware/backend failures: AMD GPU loading vs MPS histogram not implemented." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and symptoms: Mistral-3 dtype mismatch vs GPT-OSS on AMD GPUs." + "reason": "Different model families and failure modes: Mistral-3 dtype mismatch vs GPT-OSS-on-AMD." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are dtype/hardware-related, but they affect different models and code paths." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same bug as GPT-OSS on AMD GPUs." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model/runtime issues; not the same concrete bug." + "reason": "Different model/backend issues: AMD GPU runtime vs GraniteMoEHybrid invalid method call." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model families and failure modes, so not duplicates." + "reason": "Different model-specific runtime errors; no concrete shared code-path." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch is unrelated to ModernBERT ONNX export." + "reason": "Both are Deepspeed/model-loading issues, but one is collective mismatch and the other is ONNX export; not the same bug." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Both are model/export related, but the errors and affected paths are different." + "reason": "Different model/export failure vs Qwen3Moe FP8 failure." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export issue vs auto device mapping CUDA assert are unrelated." + "reason": "ONNX export error and Qwen3 auto device-map cuda assert are unrelated failures." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanups for v5 removals, but they concern different removed references." + "reason": "Both are v5 docs cleanups, but they point to different removed docs items and are not the same concrete change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export bug vs import error for GenerationMixin; different layers." + "reason": "ONNX export bug is unrelated to GenerationMixin import relocation." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask-shape index error during generation is not the same as Gemma4 PLE device mismatch." + "reason": "Generic mask-shape mismatch vs Gemma4 device mismatch; different symptoms and code paths." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Same error family superficially, but the models and reproductions differ." + "reason": "Both mention index/mask style errors, but one is ZeRO-3/Bert init and the other is a generic mask-length mismatch." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Both are index errors, but one is ZeRO-3/BERT initialization and the other is a mask-shape mismatch." + "reason": "Same broad error class, but different concrete failure: Bert ZeRO-3 init vs token mask shape mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different symptoms and scenarios: mask mismatch vs GPT-OSS tensor parallel issue." + "reason": "Mask shape mismatch is unrelated to GPT-OSS MoE TP-only failure." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Different failure modes in different training stacks." + "reason": "Different tensor-shape failure vs Qwen3.5 embedding unpacking with TRL." }, { "left": "issue:34634", @@ -344,319 +344,313 @@ "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but the actual collectives mismatch and BERT index error are different bugs." + "reason": "Both involve Deepspeed, but collective mismatch and Bert zero-size index errors are different bugs." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same DeepSpeed ZeRO-3/BERT index-out-of-bounds bug; titles describe the same underlying crash." + "reason": "Near-identical ZeRO-3/Bert index-out-of-bounds reports with the same underlying init failure." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross_entropy call issue is unrelated to GPT-OSS tensor-parallel failure." + "reason": "Liger Kernel cross_entropy dispatch and GPT-OSS TP-only failure are different issues." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Detokenization bug and LoRA save/load mismatch are different problems." + "reason": "Detokenization regression is unrelated to LoRA merge/save/load output mismatch." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different model/runtime issues; not the same underlying defect." + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA save/reload mismatch are unrelated." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Checkpointing/FSDP tensor-size mismatch is different from the mask-shape index error." + "reason": "Activation-checkpointing/FSDP recompute mismatch is not the same as the mask-shape index error." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed is shared, but one is a weight-dimension error and the other is a collective mismatch." + "reason": "DeepSpeed training weight-dimension error and collective mismatch are different failures." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3/BERT index error is unrelated to GPT-OSS tensor-parallel failure." + "reason": "Bert ZeRO-3 init failure is unrelated to GPT-OSS MoE tensor-parallel issues." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation checkpointing FSDP issue is unrelated to CustomTrainer.compute_loss signature change." + "reason": "FSDP checkpointing shape mismatch is unrelated to CustomTrainer.compute_loss signature breakage." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different model families and different error classes." + "reason": "Different model/backend areas: Bert ZeRO-3 init vs GPT-OSS MoE tensor parallel." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different model-specific runtime bugs." + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are unrelated." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3TextConfig missing field with vLLM is unrelated to Mistral tokenizer regex patch crash." + "reason": "Gemma3TextConfig sliding_window_pattern and Mistral tokenizer regex patch are different model/tokenizer bugs." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Both concern Gemma4, but one is a device-mapping forward bug and the other is an expensive reverse-embedding lookup." + "reason": "Gemma4 device-map mismatch during forward and expensive reverse embedding lookup are different code paths." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch version incompatibility is unrelated to Qwen-3-8B-NVFP4 runtime failure." + "reason": "Torch-version incompatibility is unrelated to Qwen-3-8B-NVFP4 runtime failure." }, { "left": "issue:43828", - "right": "issue:44457", - "accept": false, - "reason": "dtype autocast mismatch is unrelated to LoRA merge/save output divergence." - }, - { - "left": "issue:34928", - "right": "issue:41720", + "right": "issue:45161", "accept": false, - "reason": "FSDP checkpointing mismatch is not the same as Qwen3 auto device mapping CUDA assert." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as GPT-OSS MoE TP-only failure." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "Qwen3.5 embedding unpacking bug differs from GPT-OSS tensor-parallel failure." + "reason": "Qwen3.5 embedding unpacking with TRL and GPT-OSS TP-only failure are different bugs." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Mixtral grouped_mm tracing error is unrelated to Mistral-3 image preprocessor dtype mismatch." + "reason": "Mixtral grouped_mm tracing failure is unrelated to Mistral-3 image-preprocessor dtype mismatch." }, { "left": "issue:43716", "right": "issue:45685", "accept": false, - "reason": "Different models and different backend-specific failures." + "reason": "Different model/backend failures: Mistral-3 dtype mismatch vs MPS histogram not implemented." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Different dtype/runtime issues on different models." + "reason": "Autocast dtype mismatch and Qwen-3-8B-NVFP4 runtime error are unrelated." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "GraniteMoEHybrid invalid method is unrelated to the MPS histogram not implemented error." + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the MPS Int histogram limitation." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different model-specific dtype/runtime problems." + "reason": "Mistral-3 dtype mismatch and Qwen-3-8B-NVFP4 runtime error are different issues." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "NVFP4 runtime failure is unrelated to Qwen3.5 embedding unpacking." + "reason": "Different Qwen-related failures with different triggers and code paths." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger Kernel cross_entropy issue is unrelated to Qwen3.5 embedding unpacking." + "reason": "Liger Kernel cross_entropy dispatch and Qwen3.5 embedding unpacking are unrelated." }, { "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Gemma3/Gemma3 ZeRO-3 index error is similar in shape but a different model and failure path." + "reason": "Gemma3 ZeRO-3 loading failure and Bert ZeRO-3 loading failure share a backend, but not the same concrete bug." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs issues for removed v5 features, but they target different references." + "reason": "Both are docs updates for v5 removals, but they cover different removed items and pages." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3-vl-embedding StopIteration." + "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3-vl StopIteration." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Different backend/model combinations and different runtime errors." + "reason": "Phi autocast dtype mismatch is unrelated to the MPS histogram backend limitation." }, { "left": "issue:44918", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 embedding issue is unrelated to MPS histogram op failure." + "reason": "Qwen3.5 embedding/TRL issue is unrelated to MPS Int histogram support." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Different dtype-related bugs in different subsystems." + "reason": "Liger Kernel cross_entropy dispatch is unrelated to Phi-tiny-MoE autocast dtype mismatch." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch is unrelated to trainer compute_loss API breakage." + "reason": "DeepSpeed collective mismatch and Trainer compute_loss signature mismatch are unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Activation checkpointing/FSDP bug is unrelated to the GenerationMixin import error." + "reason": "Activation checkpointing/FSDP mismatch is unrelated to GenerationMixin import error." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross_entropy call issue is unrelated to mask-shape indexing error." + "reason": "Liger Kernel cross_entropy call and mask-shape index error are unrelated." }, { - "left": "issue:43901", - "right": "issue:44512", + "left": "issue:41762", + "right": "issue:43596", "accept": false, - "reason": "Different docs regressions: text classification docs vs run command docs." + "reason": "Both are ZeRO-3 loading issues, but Gemma3 and Bert fail differently." }, { - "left": "issue:41762", - "right": "issue:43596", + "left": "issue:43901", + "right": "issue:44512", "accept": false, - "reason": "Both involve DeepSpeed ZeRO-3 and index errors, but the affected models and root causes differ." + "reason": "TextClassificationPipeline return_all_scores docs and run-command docs are distinct documentation fixes." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3TextConfig/vLLM issue is unrelated to Qwen3 auto device mapping CUDA assert." + "reason": "Gemma3 sliding_window_pattern and Qwen3 auto device map assert are different bugs." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral is unrelated to Gemma4 device-mapping forward mismatch." + "reason": "Voxtral processor failure is unrelated to Gemma4 device mismatch." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy issue is unrelated to Qwen3-vl-embedding StopIteration." + "reason": "Liger Kernel cross_entropy call is unrelated to Qwen3-vl video StopIteration." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted checkpoint and NemotronH loading bug are not the same issue." + "reason": "Different model-loading problems: corrupted Ovis2 checkpoint vs NemotronH checkpoint compatibility." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Unit-test model load failure differs from a corrupted checkpoint report." + "reason": "GLM-4.7-Flash test loading failure and Ovis2 checkpoint corruption are unrelated." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Gemma3TextConfig attribute bug is unrelated to Qwen3-MoE sliding_window issue." + "reason": "Gemma3 sliding_window_pattern and Qwen3-MoE sliding_window are different model config bugs." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask/tensor shape mismatch is a different bug from the ZeRO-3/BERT index error." + "reason": "Mask-index mismatch and Bert ZeRO-3 zero-size index error are different concrete failures." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version incompatibility is unrelated to the pydantic PretrainedConfig field regression." + "reason": "Torch version incompatibility and pydantic PretrainedConfig field breakage are unrelated." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "ZeRO-3/BERT index error is unrelated to torch/transformers version incompatibility." + "reason": "ZeRO-3/Bert init failure is unrelated to torch/transformers version incompatibility." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper load failure is unrelated to the pydantic PretrainedConfig regression." + "reason": "Whisper loading failure and pydantic PretrainedConfig breakage are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Qwen3.5 chat crash and streaming inference failure are different symptoms." + "reason": "Qwen3.5 chat crash and streaming chat/completions failure are different bugs." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "lm_head weight tying bug is unrelated to Ovis2 checkpoint corruption." + "reason": "Mistral3 lm_head tying bug and Ovis2 checkpoint corruption are unrelated." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Different model/config issues; not the same bug." + "reason": "Gemma3 sliding-window config and Gemma4 PLE device mismatch are different model issues." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 streaming failure is unrelated to the MPS histogram backend error." + "reason": "Streaming inference failure and MPS histogram backend limitation are unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Mixtral torch.dynamo tracing failure is unrelated to the MPS histogram issue." + "reason": "Torch Dynamo grouped_mm tracing failure is unrelated to MPS Int histogram support." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Different models and different numerical/runtime failures." + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different bugs." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "GenerationMixin import error is unrelated to Gemma3TextConfig missing attribute." + "reason": "GenerationMixin import error is unrelated to Gemma3 sliding_window_pattern." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall. The summary stays conservative, and the only accepted duplicate pair is well-supported by the very close ZeRO-3/BERT index-out-of-bounds titles. I did not see any soft-edge verdict that overstates the evidence based on the packet." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43596 and issue:43638, is well supported by their near-identical ZeRO-3/Bert index-out-of-bounds reports. The other soft-edge rejections are phrased cautiously and do not overclaim duplicate status. No PR-pair mergeability concerns apply here because this cluster contains issues, not pull requests." }, "evaluator_used": true, "retried": false, @@ -762,14 +756,14 @@ "issue:44292|issue:45507", "issue:39692|issue:43054" ], - "prepared_review_unit_hash": "f97f38e2cb29fcb674c8580b43832c1c7e21297134d41898c96578dab3c7314d", + "prepared_review_unit_hash": "cdd922facfff498c8bb0063da0fef50144165f456446b01473460b9066578b83", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -784,222 +778,222 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated Transformers bug reports. The only clear duplicate subcluster is the Transformers v5 non-persistent-buffer regression, where three issues describe the same corruption bug from slightly different angles.", - "confidence": 0.74, - "canonical_issue_reason": "issue:43950 is the best canonical issue because it most precisely states the shared regression (`from_pretrained()` corrupting non-persistent buffers) and sits at the center of the clear duplicate subcluster.", + "summary": "Mostly unrelated issues were grouped by weak lexical similarity. The only clear duplicate cluster is the Transformers v5 non-persistent buffer corruption regression (`43644` / `43950` / `44534`); the rest should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": "`43950` is the best canonical issue for the real duplicate set because it states the regression most explicitly (`from_pretrained` corrupts non-persistent buffers) and cleanly subsumes `43644` and `44534`.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the strongest representative issue in this cluster: it is specific, regression-focused, and matches the confirmed duplicate theme better than the broader rephrasings.", + "best_issue_reason": "`43950` is the strongest representative of the cluster\u2019s actual shared bug: a v5 regression corrupting `register_buffer(persistent=False)` state during load.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Different failures: a processor/model load problem versus a compile-time template-node TypeError." + "reason": "Different problems: a model-specific processor/load failure versus a compiler error about non-template nodes." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Both mention version/model incompatibility, but one is a torch-transformers compatibility report and the other is a Gemma4 device-mismatch forward bug." + "reason": "Unrelated failures: version compatibility between torch/transformers versus a Gemma4 device-map forward mismatch." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are index-out-of-bounds style errors, but the model paths and triggering conditions differ; not clearly the same bug." + "reason": "Both are IndexErrors under ZeRO-3, but the reported symptoms and code paths differ too much to treat as the same bug." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "One is a vLLM/PaddleOCR-VL loading failure; the other is GGUF support for gpt-oss. Same area, different bug." + "reason": "One is a vLLM loading failure for PaddleOCR-VL; the other is a request for GGUF support for gpt-oss." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Unrelated regressions: missing pad_token_idx in StableLmConfig versus a pydantic PretrainedConfig field breakage." + "reason": "A missing config field after v5 update is not the same as a pydantic model-field regression." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "One is an int4 quantization memory/OOM regression; the other is a tied-weights issue in translation models." + "reason": "Quantization/OOM behavior and tied-weights handling are different code paths and different failure modes." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both involve MoE/GPT-style models, but the reported runtime failures and code paths differ." + "reason": "Torch Dynamo grouped_mm tracing error is distinct from a tensor-parallel support issue in GPT-OSS MoE." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "An autocast dtype-mismatch error and a chat/completions streaming failure are not the same concrete bug." + "reason": "Autocast dtype mismatch and streaming chat/completions failure are not the same concrete bug." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Different regressions: init_empty_weights argument handling versus a pydantic model field breakage." + "reason": "Unexpected `_is_hf_initialized` argument during `init_empty_weights` is unrelated to the pydantic `PretrainedConfig` field regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GPT-OSS GGUF support and Qwen3.5 streaming chat failures are unrelated." + "reason": "GGUF architecture support and Qwen3.5 streaming inference are different feature areas and failures." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code breakage and torch/transformers incompatibility are too broad and do not point to one shared defect." + "reason": "Remote-code breakage is a different issue than a torch/transformers version incompatibility." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and PretrainedConfig/pydantic regression are different problems." + "reason": "A sliding-window bug in Qwen3-MoE is unrelated to the v5 pydantic config-field problem." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different code paths: torch dynamo grouped_mm tracing versus missing Float8 storage support." + "reason": "These are different runtime errors in different subsystems; only superficial similarity exists." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index errors under ZeRO-3, but the reported shapes/models and likely root causes differ." + "reason": "Both mention index errors with ZeRO-3, but one is a mask-shape mismatch while the other is an empty-tensor load failure." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both concern SigLIP2 quality, but one is a general embedding-quality complaint and the other is an AutoModel/pipeline misbehavior report." + "reason": "Both mention SigLIP2, but one reports poor embedding quality while the other reports nonsensical AutoModel/pipeline outputs; not clearly the same bug." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorBoard-without-TensorFlow backend error is unrelated to torch/transformers version incompatibility." + "reason": "A missing TensorFlow backend check and a torch/transformers incompatibility are unrelated." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "WikiText-2 perplexity on gpt-oss and a GLM5 issue are different model-level reports." + "reason": "High perplexity on GPT-OSS and a GLM5 issue are not the same underlying problem." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Both involve model/input mismatches, but one is an image preprocessor dtype issue and the other is an embeddings/unpacking issue with trl SFT." + "reason": "Image preprocessor/model dtype mismatch is unrelated to Qwen3.5 input embedding unpacking in TRL SFT trainer." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape runtime error and Qwen3.5 embedding unpacking are not the same bug." + "reason": "FSDP `weight must be 2-D` and TRL embedding unpacking failures are different code paths and different bugs." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both are about custom-code loading, but one is a leaf-directory cache collision and the other is wrong module loading after save_pretrained; not enough to merge as one bug." + "reason": "Both involve custom-code loading, but one is a cache-path collision for local leaf directories and the other is wrong-module loading after save/load; too distinct to merge." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Tied lm_head weights in Mistral3 and tied weights on translation models are related conceptually but not the same concrete issue." + "reason": "Both mention tied weights, but they target different model families and don\u2019t clearly describe the same concrete regression." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Notebook/custom-model initialization and flash-attn fallback checks are unrelated regressions." + "reason": "Notebook custom-model initialization and flash-attn fallback selection are unrelated." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache linear_attention failure and tied-weights translation-model regression are different subsystems and bugs." + "reason": "PagedAttentionCache crashing on `linear_attention` is unrelated to tied-weights handling." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "DSA indexer ReLU discussion and docs mentioning removed run command are unrelated." + "reason": "A DSA-indexer discussion and a docs typo are not the same underlying change or bug." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and flash-attn fallback gating are different issues." + "reason": "FineGrainedFP8Config failure and flash-attn fallback blocking are different quantization/attention issues." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior and tied-weights handling are not the same bug." + "reason": "Quantization offloading behavior is not the same as tied-weights regressions." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading and tied weights in translation models are unrelated." + "reason": "NemotronH checkpoint loading failure is unrelated to tied-weights issues." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A cache-strategy typo and a removed CLI command in docs are not the same underlying change." + "reason": "A cache-strategy typo and a removed run-command doc issue are just documentation-level differences." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 SFT input-embedding unpacking and GraniteMoEHybrid invalid method calls are different failures." + "reason": "TRL SFT embedding unpacking and an invalid model method call are different bugs." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Incorrect detokenization and Qwen-3-8B-NVFP4 runtime errors do not share a concrete code path." + "reason": "Incorrect detokenization and an NVFP4 runtime error do not point to the same fix." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same regression: both report Transformers v5 corrupting non-persistent buffers; only wording differs." + "reason": "Both describe the same v5 regression where non-persistent buffers get corrupted/junk during loading." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Same underlying bug in the v5 non-persistent-buffer handling regression; 43950 just states the API and corruption more explicitly." + "reason": "Clear duplicate: both report the same non-persistent buffer corruption regression in Transformers 5.x." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Different model/runtime failures: Qwen-3-8B-NVFP4 versus GraniteMoEHybrid invalid method calls." + "reason": "A Qwen-3-8B-NVFP4 runtime error is not the same as GraniteMoEHybrid calling an invalid method." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but one is a docs example mismatch/quantization failure and the other is text-embedding quality regression; not the same issue." + "reason": "SigLIP2 docs-example errors and poor SigLIP2 text embeddings are related only by model name, not by the same concrete bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall: the packet supports the claim that most items are unrelated and that 43644/43950/44534 form the clear duplicate subcluster around non-persistent-buffer corruption in Transformers v5. The conservative false verdicts are reasonable from the titles alone, and no pair is overstated as mergeable beyond what the packet supports." + "feedback": "The conclusions are grounded in the packet. The canonical/best issue choice is supported by the titles and descriptions, and the only accepted soft-edge pairs are the clear non-persistent-buffer regression duplicates (`43644`, `43950`, `44534`). The remaining pair rejections are phrased conservatively and do not overclaim duplicate status." }, "evaluator_used": true, "retried": false, @@ -1104,20 +1098,20 @@ "issue:43404|issue:43950", "issue:43742|issue:45399" ], - "prepared_review_unit_hash": "9865c39becefa3a005dae70336e39b858cb830ab5d19f3c65998f061e5b7f169", + "prepared_review_unit_hash": "687b549a922000cb982914a9a7330d21a686f0bfbfda5dc64f387063d4daf7b8", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 34, - "serialized_chars": 37387, + "serialized_chars": 37385, "estimated_input_tokens": 9347, "estimated_eval_tokens": 18950 }, @@ -1126,225 +1120,225 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mixed bag of unrelated model-loading, quantization, device, and docs regressions. The only clear duplicate pair is the two v5 docs issues about removed pipeline tasks; most other soft matches are too broad or only share a subsystem/theme.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44509 is the clearest representative of the one true duplicate group: it precisely states the lingering v5 docs references to removed pipeline tasks, whereas issue:43827 is the same cleanup in a more generic form.", + "summary": "Mostly heterogeneous cluster of Transformers issues. The only clearly duplicate soft edge is the docs cleanup pair; the SigLIP2 pair is thematic but not strong enough to merge as a duplicate from the titles/evidence shown.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45020 is still the best umbrella representative: it captures the broad v5-era remote_code/loading regression theme that overlaps with several nearby reports, even though it is not a true duplicate of most items.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44509 is the strongest single issue here because it is specific, actionable, and directly matches its duplicate counterpart; the rest of the items are mostly unrelated one-off regressions rather than a cohesive bug cluster.", + "best_issue_reason": "issue:45020 is the best global issue candidate because it is the broadest, most central report in the cluster and best represents the common version-related loading breakage theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different models and different failure modes: autocast dtype mismatch vs Qwen3.5 flash-attn/position_ids crash." + "reason": "Different bugs and code paths: autocast dtype mismatch vs Flash Attention/position_ids handling." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Unrelated errors: a custom training loss signature mismatch vs a ZeRO-3 loading IndexError." + "reason": "Trainer API kwarg mismatch vs ZeRO-3 loading index error; unrelated failures." }, { "left": "issue:41628", "right": "issue:45800", "accept": false, - "reason": "Only a broad version-compatibility theme is shared; not the same concrete import bug." + "reason": "Missing import/export for AutoImageProcessor vs torch/transformers version incompatibility." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Corrupted checkpoint report is not the same as generic loading noise." + "reason": "Corrupted checkpoint vs generic loading noise; not the same defect." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Both are compatibility-type errors, but they involve different quantization/storage paths and are not the same bug." + "reason": "bitsandbytes Int8Params constructor incompatibility vs missing Float8 storage object; different backends." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality issue vs tokenizer behavior change; too different in symptom and scope." + "reason": "SigLIP2 embedding quality regression vs tokenization changes; different underlying behavior." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error message for Voxtral is unrelated to Molmo's missing tied-weights attribute." + "reason": "AutoTokenizer error for Voxtral vs missing all_tied_weights_keys on Molmo." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both involve SigLIP2, but one is zero probabilities and the other is embedding quality; not clearly the same defect." + "reason": "Both mention SigLIP2, but zero probabilities and degraded embeddings are not clearly the same concrete bug." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Flash-attn packed-sequence crash and CPU-offload rotary embedding incompatibility are different code paths." + "reason": "Both involve Qwen3.5, but one is a Flash Attention crash from 3D position_ids and the other is CPU-offload incompatibility in rotary embedding." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and different failures." + "reason": "Different model-loading failures for different models and code paths." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Both mention memory, but one is ineffective gradient checkpointing and the other int4 CUDA reserved-memory growth." + "reason": "Gradient checkpointing/LoRA issue vs int4 quantization memory regression; not the same bug." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Likely related SigLIP2 problems, but the evidence is too indirect to treat them as the same bug." + "reason": "Only a thematic SigLIP2 match; the titles point to a generic quality complaint vs a specific HF-vs-JAX implementation discrepancy, so duplicate status is not established." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise is not the same as a specific model load failure." + "reason": "Generic loading noise vs inability to load a specific model; not duplicates." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both involve remote-code loading, but one is missing config fields and the other is broader recent-version breakage." + "reason": "Missing fields in AutoConfig with trust_remote_code is a specific config-return bug, not the same concrete failure as the broader remote_code breakage report." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Tokenizer detokenization regression is unrelated to the flash_attn import-utils KeyError." + "reason": "Detokenization regression vs flash_attn import KeyError; unrelated subsystems." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "DeepSeek tokenization bug and AMD GPU support issue are unrelated." + "reason": "Detokenization regression vs AMD GPU support failure; different problems." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Generic loading noise is not the same as a specific model failing to load." + "reason": "Loading noise vs model-specific loading failure; not duplicates." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Both are training-memory complaints, but not the same mechanism or fix." + "reason": "LoRA checkpointing behavior vs MoE training memory usage; related area, different bug." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model families and different loading/runtime errors." + "reason": "Both loading-related, but one is a model load failure and the other is a missing attribute during weight tying." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and Qwen3MoE FP8 failure are unrelated." + "reason": "Gemma3 ZeRO-3 load crash vs Qwen3 MoE LoRA compatibility; different models and code paths." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only load error is not the same as the GLM5 issue." + "reason": "Qwen3VL weight_only load error vs GLM5 issue; unrelated." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Version-compatibility warning vs LoRA compatibility breakage; related theme, different bug." + "reason": "Both mention Qwen3 MoE/LoRA, but one is a cross-version incompatibility and the other is a warning; not the same concrete bug." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Detokenization regression vs specific model loading failure; different defects." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram kernel limitation and Gemma4 device mismatch are different device issues." + "reason": "MPS histogram op limitation vs Gemma4 device mismatch during forward; separate failures." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Importing a missing model class is unrelated to bitsandbytes Int8Params incompatibility." + "reason": "Missing model class import vs bitsandbytes Int8Params constructor error; unrelated." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code breakage and Gemma4 device-mapping mismatch are different problems." + "reason": "Remote_code loading regressions vs Gemma4 PLE device placement bug; different code paths." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer mapping logic failure is not the same as remote-code model loading regression." + "reason": "Tokenizer mapping helper failure vs remote_code loading breakage; not the same bug." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer load failure and flash-attn fallback gating are unrelated." + "reason": "Tokenizer load failure for a specific model vs flash-attn fallback gating; unrelated." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch and PLE device mismatch are different runtime errors." - }, - { - "left": "issue:43824", - "right": "issue:43872", - "accept": false, - "reason": "Import error and bitsandbytes storage/parameter incompatibility are unrelated." + "reason": "Autocast dtype mismatch in Phi MoE vs Gemma4 device mismatch; different issues." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both are version regressions, but one is remote-code loading and the other is a Kimi-K2.5 tokenizer regression." + "reason": "Remote_code loading regression vs Kimi-K2.5 tokenizer regression; different model and mechanism." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral regex patch crash and flash-attn2 fallback gating are unrelated." + "reason": "Mistral regex patch crash vs flash-attn fallback gating; unrelated." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both report stale docs still referencing pipeline tasks removed in v5; same documentation cleanup." + "reason": "Both are the same docs cleanup problem: outdated pipeline-task references removed in v5." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Tied lm_head weights bug and non-persistent buffer corruption are different state/load bugs." + "reason": "Tied-weight bug in Mistral3 vs non-persistent buffer corruption in from_pretrained; different loading semantics." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM key error is unrelated to flash-attn fallback checks." + "reason": "MobileLLM load key error vs flash-attn fallback checks; different failures." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative on the soft matches. The only accepted duplicate pair, issue:43827 \u2194 issue:44509, is well supported by the titles and descriptions, and the rejected pairs are mostly appropriately treated as same-subsystem-but-different-bug rather than duplicates." + "accept": false, + "feedback": "The overall cluster summary is reasonable and conservative, but one soft-edge verdict is too strong for the evidence available: `issue:43054` vs `issue:43493` is only a plausible thematic match (both SigLIP2-related), not clearly the same bug/PR from the titles and packet alone. The `issue:43827` vs `issue:44509` duplicate call looks grounded, but because at least one accepted pair overstates the evidence, I can\u2019t accept the full evaluation." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1445,14 +1439,14 @@ "issue:44560|issue:44918", "issue:43950|issue:45325" ], - "prepared_review_unit_hash": "d74958cfb31320a7475bf33735a9d76fcdf1d5aad296a17b491be1a8a00adc43", + "prepared_review_unit_hash": "48dbe281c8815820c0efc3885d2ccccb22d19f01ca4f7fa23c3599ebbd1a884d", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -1467,210 +1461,210 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issue reports with a few likely duplicate pairs. I only found three soft edges that look like true duplicates: the custom-model-init regression, the `_is_hf_initialized` loading failure, and the repeated mask/tensor length IndexError.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43646 is the cleanest, broadest statement of the custom model initialization regression, while issue:43645 reads like a narrower Jupyter-specific instance of the same bug.", + "summary": "This cluster is mostly a noisy mix of unrelated Transformers model-loading/training regressions. Clear duplicates exist for the `_is_hf_initialized` init-empty-weights bug, the mask-shape IndexError, the custom-model initialization regression, and likely a couple of Qwen ZeRO-3/video regressions.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here: it is open, spans recent Transformers versions, and captures the recurring remote-code / model-loading regression theme better than the narrower model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best representative issue overall because it is concise, general, and easy to search/triage compared with the more environment-specific duplicate.", + "best_issue_reason": "Issue 45020 is the best representative because it is the most general and still-open issue among many specific loading failures, so it can anchor follow-up duplicates and related reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Different audio issues: one is a Whisper model run failure, the other is a feature-extraction function bug." + "reason": "Both are audio-related, but one is a generic Whisper failure and the other is a specific fbank helper issue; not enough evidence they are the same bug." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Different problems: a Qwen3MoE FineGrainedFP8Config failure versus a broad remote_code regression." + "reason": "Both mention model failures, but one is a Qwen3MoE FP8/config problem and the other is a broad remote_code regression; different code paths." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Both are model-loading complaints, but they concern different models and different failure modes." + "reason": "Different models and failure modes: Qwen3 MoE weight conversion with Accelerate/DeepSpeed vs a corrupted Ovis2 checkpoint." }, { "left": "issue:43638", "right": "issue:45823", "accept": false, - "reason": "Different code paths: DeepSpeed ZeRO-3 training index error vs Gemma4 device-map forward mismatch." + "reason": "Both involve loading/training, but one is a ZeRO-3/BERT index error and the other is a Gemma4 device-mismatch forward bug." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same underlying `_is_hf_initialized`/empty-weights loading regression; one report surfaces through bitsandbytes, the other through init_empty_weights." + "reason": "Same underlying `_is_hf_initialized` / `init_empty_weights` constructor regression; the second report is the concrete TypeError variant of the first." }, { "left": "issue:43940", "right": "issue:45313", - "accept": false, - "reason": "Both mention ZeRO-3 weight loading, but they target different Qwen variants and the reports point to different missing-weight symptoms." + "accept": true, + "reason": "Both describe the same DeepSpeed ZeRO-3 weight-loading failure on Qwen3-family models, producing missing parameters during load." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Unrelated: FP8/Qwen3MoE failure versus bitsandbytes `_is_hf_initialized` constructor incompatibility." + "reason": "FP8/Qwen3MoE issues are unrelated to the bitsandbytes `_is_hf_initialized` constructor bug." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading problems, but for different models and different error behavior." + "reason": "Both are tokenizer/loading complaints, but they concern different models and different problems (bad error message vs actual tokenizer load failure)." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different checkpoints/models and different loading failures." + "reason": "Different model families and different load failures; no shared concrete bug is evident." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Different bugs: weight conversion under accelerate+deepspeed vs gradient averaging under GAS." + "reason": "Both involve training/loading with distributed setups, but one is weight conversion under DeepSpeed while the other is gradient averaging with GAS." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different errors and code paths: Trainer signature mismatch vs tensor mask-shape IndexError." + "reason": "Different regressions: a custom compute_loss signature change vs a tensor/mask length mismatch IndexError." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "Unrelated: Qwen3.5 RLHF/SDPA numerical instability versus flash-attention aux-state AttributeError." + "reason": "Qwen3.5 SDPA/position_id issues are unrelated to the flash_attention `s_aux=None` crash." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "Different subsystems: model loading constructor kwargs vs CLIPTokenizer max length handling." + "reason": "The bitsandbytes constructor regression is unrelated to CLIPTokenizer's `model_max_length` behavior." }, { "left": "issue:41093", "right": "issue:44805", "accept": true, - "reason": "Same mask/tensor length mismatch IndexError pattern; these look like duplicate reports of the same bug." + "reason": "Same generic mask-shape IndexError; the later issue is the same failure class with a different reproduction." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are loading failures, but for different tokenizers/models and different underlying causes." + "reason": "Both are model-loading complaints, but they target different tokenizers/models and do not share a concrete code path." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Different issues: Qwen3.5 PagedAttentionCache crash versus NemotronH checkpoint loading." + "reason": "PagedAttentionCache's `linear_attention` crash is unrelated to NemotronH checkpoint loading." }, { "left": "issue:44479", "right": "issue:44560", - "accept": false, - "reason": "Same broad Qwen3.5 area, but one is a video-input regression and the other is a different embedding/video error." + "accept": true, + "reason": "Both describe the same Qwen VL video-input regression in Transformers 5.3.0; the second is a specific Qwen3-VL manifestation." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Completely different domains: DeepSpeed training RuntimeError vs image processor segmentation-map handling." + "reason": "DeepSpeed training weight-shape error and void segmentation-map processing are unrelated." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Unrelated: GGUF gpt-oss architecture support versus Phi MoE autocast dtype mismatch." + "reason": "GGUF gpt-oss architecture support and a Phi MoE autocast dtype mismatch are different problems." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4 issues, but one is FSDP2 shared_kv_states logic and the other is a device-map mismatch." + "reason": "Both are Gemma4-related, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Both involve audio generation/inference, but they are different bugs with different triggers and symptoms." + "reason": "AudioFlamingo token leakage between tracks and Whisper empty-transcription generation failure are different audio bugs." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different Qwen3/GGUF loading and sliding-window bugs." + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window behavior are unrelated." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Different bugs: non-persistent buffer corruption in from_pretrained vs Kosmos2.5 OCR index error." + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to Kosmos2.5 OCR index errors." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Both touch Qwen3.5 attention, but the reported failures are different code paths and not clearly one bug." + "reason": "Different model families and different failure modes: SDPA torch.compile crash vs flash-attention illegal memory access from 3D position_ids." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Unrelated: Qwen3-vl video StopIteration versus an MPS histogram kernel limitation." + "reason": "Qwen video regression and an MPS histogram dtype limitation are unrelated." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Different issues: missing class import versus meta-device model loading regressions." + "reason": "Importing a missing class and meta-device loading regressions are different issues." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same Transformers 5.0.0 custom-model-initialization regression; the notebook mention is just a narrower environment-specific variant." + "reason": "Same custom model initialization regression; the notebook reproduction is just a context detail." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Different model/feature areas and different runtime errors." + "reason": "GGUF architecture support is unrelated to the Float8 storage lookup error." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Different failures: Qwen3MoE FP8 config issue versus meta-device loading regression." + "reason": "Qwen3MoE FP8/config failure and meta-device loading regression do not share the same bug." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "No clear overlap: 3d_parallel.py run failure versus GGUF gpt-oss architecture support." + "reason": "A 3d_parallel.py runtime failure and gpt-oss GGUF support are not the same issue." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Both reference Qwen3.5, but one is a video input StopIteration and the other is an embedding unpacking/trainer issue." + "reason": "Qwen3-VL video StopIteration and TRL input-embedding unpacking are different code paths." }, { "left": "issue:43950", "right": "issue:45325", "accept": false, - "reason": "Different bugs: non-persistent buffer corruption versus Qwen2.5-VL rope-index scaling." + "reason": "Buffer corruption in from_pretrained is unrelated to the Qwen2.5-VL rope-index scaling bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet, and the three accepted soft edges are conservative and supported by the issue titles/body context. The rejected pairs also read as cautious, with no obvious overstatement. Canonical issue selection between 43645/43646 is reasonable." + "feedback": "Grounded overall: the cluster is mostly heterogeneous, and the rejected pairs are appropriately cautious. The accepted edges are limited to closely matching failure classes and stay within the evidence in the packet." }, "evaluator_used": true, "retried": false, @@ -1776,20 +1770,20 @@ "issue:36010|issue:39404", "issue:45276|issue:45335" ], - "prepared_review_unit_hash": "efa3b4472fd22cba3cdadc42919ad6e08b161af97d06adfcc7dff8a4bd311367", + "prepared_review_unit_hash": "00a67d79b91120d25edaad19053fd17c3fc2ee2061eca997f4b69171c5f4bf24", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37781, + "serialized_chars": 37782, "estimated_input_tokens": 9446, "estimated_eval_tokens": 19148 }, @@ -1798,219 +1792,219 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of model-loading, tokenizer, config, and memory/regression reports. From titles alone, only a small number of soft edges look like the same underlying bug; most candidates should stay rejected.", - "confidence": 0.48, - "canonical_issue_reason": "No issue cleanly covers the whole cluster. 44488 is the best broad representative because it captures the recurring user-facing model/tokenizer loading regression theme, but it does not unify the unrelated sub-bugs.", + "summary": "Mostly a heterogeneous Transformers issue set with a few true duplicate pairs around the same regressions (non-persistent buffers junk, SigLIP2 bad outputs, and resize_token_embeddings propagation). Most other soft pairs are only thematically similar, not the same underlying bug.", + "confidence": 0.59, + "canonical_issue_reason": "Best duplicate anchor in the set: a concrete Transformers v5 regression with a clear symptom and a close near-duplicate (44534), making it the cleanest issue to represent a subcluster.", "canonical_pr_reason": null, - "best_issue_reason": "44488 is the most umbrella-like, user-facing regression report in the set and best matches the repeated loading-failure pattern.", + "best_issue_reason": "Best overall issue to represent the cluster's broad regression theme: it is the most general, actionable Transformers v5 bug here and has stronger centrality than the model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Different failures and code paths: tokenizer-mapping lookup vs Gemma4 PLE device mismatch." + "reason": "Different failures: tokenizer-mapping logic vs Gemma4 PLE device mismatch." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Unrelated bugs: GenerationMixin import error vs Qwen3Moe FP8 load failure." + "reason": "ImportError for GenerationMixin is unrelated to Qwen3MoE FP8 loading." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 inference regression surface: both report broken outputs for google/siglip2-base-patch16-224." + "reason": "Same SigLIP2 base model/output regression; both report nonsensical/zero results with AutoModel/pipeline." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "Different code paths and symptoms: `_is_hf_initialized` flag regression vs Kosmos2.5 OCR index error." + "reason": "_is_hf_initialized/deepspeed-v5 issue is unrelated to Kosmos2.5 long-input index errors." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Both are indexing failures, but the model/symptom details do not indicate the same bug." + "reason": "Different code paths and symptoms: video StopIteration vs mask/tensor shape mismatch." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both mention config loading, but the titles do not establish the same concrete validation bug." + "reason": "Both involve config validation, but they are different model-specific loading failures." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Different problems: ZeRO-3/BertModel initialization vs tokenizer-mapping logic." + "reason": "Zero3 Bert initialization error is unrelated to add-new-model-like tokenizer mapping failure." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Only a generic error theme overlaps; the concrete failures look unrelated." + "reason": "Non-pretrained Bert Zero3 index error is a different bug from TOKENIZER_MAPPING_NAMES handling." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping bug vs mask/tensor shape mismatch; not the same underlying issue." + "reason": "Tokenizer-mapping bug and mask-shape IndexError are not the same underlying problem." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both involve loading/memory, but the concrete failures and fixes differ." + "reason": "Dequantized offload load failure is different from mmap-related OOM mitigation." }, { "left": "issue:43644", "right": "issue:44534", - "accept": false, - "reason": "The titles are very similar, but the packet only gives titles/release context, which is not enough to confirm an exact duplicate." + "accept": true, + "reason": "Same non-persistent buffer junk regression in Transformers v5, just reported with slightly different wording." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but one is bad inference output and the other is a docs/example plus quantization failure." + "reason": "Same model family, but doc-example errors and runtime zero-probabilities are different bugs." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different models and failures: GLM5 vs Qwen3.5 flash-attention generation crash." + "reason": "GLM5 and Qwen3.5 flash-attention crashes are different model-specific issues." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizer code paths and model families." + "reason": "Tokenizer pre-tokenizer override regression is unrelated to Granite token-id corruption." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different serialization/loading paths: ZeRO-3 weight loading vs save_pretrained visual-key serialization." + "reason": "ZeRO-3 weight loading and save_pretrained visual-key regression are distinct." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but the visible evidence is too weak to prove the same root cause." + "reason": "Both are BERT loading failures, but the models and failure details are too different to treat as one bug." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different model families and different loading failures." + "reason": "Qwen3 MoE hf4/hf5 compatibility is unrelated to BitNet packed-weight loading." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3-related, but the concrete problems differ: auto device mapping vs FP8 config loading." + "reason": "Auto device mapping CUDA assert and FineGrainedFP8Config failure are different issues." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Training resume step-count logic is unrelated to SGD argument handling." + "reason": "Checkpoint resume max_steps bug is unrelated to SGD args being ignored." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different checkpoints and different loading symptoms." + "reason": "GLM-4.7-Flash unit-test loading and Qwen3-VL weight-shape mismatch are distinct model-loading bugs." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Qwen3 MoE LoRA loading issue vs Gemma4 resize_token_embeddings propagation bug." + "reason": "Qwen3 MoE loading compatibility is unrelated to resize_token_embeddings propagation." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Qwen3Moe FP8 config failure vs mmap/OOM on Strix Halo are different bugs." + "reason": "Qwen3MoE FP8 failure is unrelated to mmap OOM on Strix Halo." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different model families and different loading code paths." + "reason": "MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7-Flash test loading." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both touch SigLIP2, but the visible reports point to different symptoms and contexts." + "reason": "SigLIP2 doc-example issues are not the same as the zero/nonsensical output regression." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility and Qwen3-VL shape mismatch are unrelated." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Different memory/offload issues; the titles do not show the same bug." + "reason": "Training memory inefficiency and quantization offload behavior are different problems." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading failures, but the model-specific reports do not establish the same code-path bug." + "reason": "Both are tokenizer-loading regressions, but for different models and likely different root causes." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "The titles are near-identical, but there is not enough visible evidence here to confirm an exact root-cause duplicate." + "reason": "Non-persistent buffer corruption and Jupyter custom-model initialization are separate regressions." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Separate crashes in different components: chat flow vs transformers serve/Gemma4Processor." - }, - { - "left": "issue:43856", - "right": "issue:44387", - "accept": false, - "reason": "Both involve memory, but one is training inefficiency and the other is an int4 reserved-memory OOM regression." + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both involve flash attention, but the concrete failures are different." + "reason": "Flash-attention packed-sequence bug and s_aux=None AttributeError are different code paths." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Unrelated bugs: Qwen3 MoE LoRA compatibility vs resize_token_embeddings propagation." + "reason": "Qwen3 MoE loading issue is unrelated to resize_token_embeddings not updating output embeddings." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin and Whisper return_language pipeline regression are unrelated." + "reason": "GenerationMixin import failure is unrelated to Whisper return_language pipeline behavior." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Same resize_token_embeddings propagation bug: both report embeddings not being updated after resize, with 45335 narrowing the affected target to decoder.embed_tokens." + "reason": "Same resize_token_embeddings propagation bug affecting model-specific embedding/output modules." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is broadly grounded, but at least one soft-edge accept overstates the evidence. In particular, the 43644\u219443645 verdict reads like an exact/root-cause duplicate claim, yet the packet only gives titles and release context, not enough to confirm the same non-persistent-buffer bug. The same caution applies to some other accepted pairs where the reasoning is stronger than the visible evidence. The cluster is heterogeneous and the conservative call is to reject." + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The three accepted soft pairs are plausible duplicate/subduplicate matches based on shared symptoms and model/feature context, and the rejected pairs do not overclaim sameness. The overall cluster description as heterogeneous with a few true duplicate regressions is well supported." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2114,289 +2108,253 @@ "issue:44589|issue:44756", "issue:44683|issue:45823" ], - "prepared_review_unit_hash": "5c18d1474230ee7f60b5f8b114d8db2797a0d340d05cc2e7ad197cb5afe97b93", + "prepared_review_unit_hash": "605e66edac7bb0768e9fd35c7bd084f550d000fad4ef100fed29d96c13797efd", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38193, - "estimated_input_tokens": 9549, - "estimated_eval_tokens": 19354 + "serialized_chars": 38197, + "estimated_input_tokens": 9550, + "estimated_eval_tokens": 19356 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "One clear duplicate pair was found: the two resize_token_embeddings reports are the same regression, with one being a narrower variant. The rest look like related-but-distinct model/loading issues, so they should stay separate.", - "confidence": 0.76, - "canonical_issue_reason": "issue:45276 is the better canonical because it states the broader resize_token_embeddings regression and includes both output_embeddings and embed_tokens_per_layer, while 45292 is a narrower subset focused only on output_embeddings.", + "summary": "Only one duplicate pair is well supported in this packet: the Gemma4 `resize_token_embeddings` regression (`45276` \u2194 `45292`). The nearby Transformers 5.0.0 custom-model initialization issues are not supported here as a duplicate pair and should not be claimed as such.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 45276 is the best canonical issue because it gives the broader Gemma4 `resize_token_embeddings` failure, explicitly covering both `embed_tokens_per_layer` and `output_embeddings`, while 45292 is the narrower variant.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45276 is the strongest cluster anchor: it is the most complete description of the shared resize_token_embeddings failure mode and subsumes the narrower companion report.", + "best_issue_reason": "Issue 45276 is the strongest overall issue candidate for the supported duplicate pair because it describes the broader symptom set and subsumes 45292.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Different failure modes: save/from_pretrained extra_state handling vs init_empty_weights/_is_hf_initialized loading error." + "reason": "Both are save/load failures, but they describe different bugs: `extra_state` handling vs unexpected `_is_hf_initialized` in `init_empty_weights`." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both involve lm_head tying, but they are different model classes and likely separate fixes rather than one shared bug." + "reason": "Both mention `lm_head` tying, but they affect different model families and different code paths." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Unrelated problems: Qwen3 device-mapping CUDA assert vs AudioFlamingo3 batched inference token leakage." + "reason": "Unrelated failures: Qwen3 auto device-mapping CUDA assert vs AudioFlamingo3 batched inference token leak." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both are load failures, but the reported model/tokenizer issues are different and not clearly the same root cause." + "reason": "Similar loading complaints, but different checkpoints with no clear shared root cause." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "One is torch 2.10 compatibility, the other is meta-device loading breakage on transformers 5.x." + "reason": "Torch version incompatibility and meta-device loading regression are different issues." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Documentation/example errors for SigLIP2 are unrelated to GLM5 model behavior." + "reason": "Different areas entirely: SigLIP2 documentation errors vs a GLM5 model issue." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing/LoRA behavior is unrelated to a T5 apex RMSNorm memory leak." + "reason": "Different bugs: PEFT LoRA gradient checkpointing behavior vs a T5 RMSNorm memory leak." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both mention quantization and memory, but one is offloading behavior and the other is int4 reserved-memory/OOM regression." + "reason": "Both are quantization/memory themed, but one is offloading behavior and the other is an int4 reserved-memory/OOM regression." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to Gemma4 FSDP2 shared_kv_states reconstruction." + "reason": "Different models and bugs: sparse layer creation vs Gemma4 FSDP2 shared-kv-state rebuilding." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Different multimodal/config issues: Gemma4 FSDP2 KeyError vs Qwen 3.5/3.6 text_config unwrapping." + "reason": "Different composite-model failures: per-layer shared-kv-state bug vs missing `text_config` unwrap." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen/Gemma model bugs, but the code paths differ completely: Gemma4 shared_kv_states vs Qwen3.5Moe rotary embedding offload." + "reason": "Different Qwen3.5-MoE problems: attention state rebuild vs CPU offload incompatibility in rotary embedding." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected init_empty_weights argument error is unrelated to Strix Halo mmap OOM." + "reason": "Both involve loading/runtime failures, but the concrete bugs differ: unexpected init argument vs mmap OOM on Strix Halo." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Audio track token leakage is unrelated to Gemma4 PLE device mismatch." + "reason": "Different device-related bugs: embedding/token leak between tracks vs Gemma4 PLE device mismatch under auto device mapping." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization memory regression is different from disabling mmap for Strix Halo OOM." + "reason": "Different regressions: int4 memory growth/OOM vs mmap-related OOM on Strix Halo." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Model-loading unit test failure for GLM-4.7-Flash is unrelated to Qwen3.5 save_pretrained checkpoint corruption." + "reason": "Both are model-loading/saving issues, but they concern different models and failure modes." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different quantization/loading failures for different models and formats (NVFP4 vs MXFP4 fallback)." + "reason": "Both are quantized model loading issues, but they target different models/quantization paths." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is PEFT/LoRA ineffectiveness and the other is eval-mode restriction." + "reason": "Same broad gradient-checkpointing area, but one is ineffective under PEFT LoRA and the other is an eval-mode restriction." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage is unrelated to mmap/OOM on Strix Halo." + "reason": "Meta-device loading regression and Strix Halo mmap OOM are different underlying bugs." }, { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "Qwen3_5 recurrent-state reset bug is unrelated to GLM5." + "reason": "Different models and symptoms: Qwen3_5 recurrent-state reset vs generic GLM5 issue." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same resize_token_embeddings regression; 45276 is the broader report and 45292 is a narrower duplicate about output_embeddings." + "reason": "Same `resize_token_embeddings` regression; 45276 is the broader Gemma4 report and 45292 is the narrower duplicate focused on `output_embeddings`." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve save_pretrained/from_pretrained, but one is file permission propagation and the other is wrong custom-module loading." + "reason": "Related save/load area, but different problems: read-only permission propagation vs wrong custom module loaded after `save_pretrained`." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior is unrelated to RLHF NaN explosion from 3D position_ids and SDPA fallback." + "reason": "Unrelated: eval-mode gradient checkpointing restriction vs RLHF NaN / SDPA fallback instability." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "extra_state save/load handling is unrelated to mmap/OOM on Strix Halo." + "reason": "Different save/load and memory issues with no shared root cause." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to resume_from_checkpoint max_steps calculation." + "reason": "Both mention training/custom behavior, but they are different bugs: custom-model init in notebooks vs `resume_from_checkpoint` max-steps calculation." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression in Qwen VL models is unrelated to Granite speech config loading strictness." - }, - { - "left": "issue:44589", - "right": "issue:44756", - "accept": false, - "reason": "Float8 storage lookup failure is unrelated to mmap/OOM." - }, - { - "left": "issue:44683", - "right": "issue:45823", - "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is unrelated to Gemma4 forward device mismatch." + "reason": "Different model families and failures: video input regression vs strict config loading of `granite_speech`." }, { - "left": "issue:39404", - "right": "issue:45701", + "left": "issue:45588", + "right": "issue:45663", "accept": false, - "reason": "Whisper return_language pipeline regression is unrelated to tokenization changes across transformers versions." + "reason": "Different attention-path bugs: `s_aux=None` flash-attention crash vs Gemma4 shared-kv-state rebuild." }, { "left": "issue:45292", "right": "issue:45335", "accept": false, - "reason": "Same resize_token_embeddings family, but 45335 is about t5gemma decoder.embed_tokens and is not the same report as 45292." + "reason": "Both are `resize_token_embeddings` complaints, but they affect different model components and do not clearly share one bug." }, { - "left": "issue:45588", - "right": "issue:45663", + "left": "issue:43645", + "right": "issue:44164", "accept": false, - "reason": "Flash-attention s_aux=None crash is unrelated to Gemma4 FSDP2 shared_kv_states." + "reason": "Notebook/custom-model initialization regression and `extra_state` save/load handling are different bugs." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Different checkpoint-loading failures for different model families and loading stacks." - }, - { - "left": "issue:45684", - "right": "issue:45698", - "accept": false, - "reason": "Duplicate of the earlier save/load pair? no: the reported problems differ materially, so keep separate." - }, - { - "left": "issue:45704", - "right": "issue:45663", - "accept": false, - "reason": "T5 apex memory leak is unrelated to Gemma4 training KeyError." - }, - { - "left": "issue:45759", - "right": "issue:45823", - "accept": false, - "reason": "Qwen multimodal config unwrapping is unrelated to Gemma4 PLE device mismatch." - }, - { - "left": "issue:45701", - "right": "issue:45859", - "accept": false, - "reason": "Tokenizer/version behavior change is unrelated to Qwen3_5Moe rotary embedding CPU-offload support." + "reason": "Different loading failures for different checkpoints and execution paths." }, { - "left": "issue:45276", - "right": "issue:45335", + "left": "issue:44402", + "right": "issue:44451", "accept": false, - "reason": "Related resize_token_embeddings area, but 45335 is a different model-specific path (t5gemma decoder.embed_tokens) and not the same bug." + "reason": "Both are model-loading failures, but the titles indicate different checkpoints and no shared root cause." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch for Phi-tiny-MoE is unrelated to meta-device loading failures." + "reason": "Different regressions: autocast dtype mismatch on Phi-tiny-MoE vs meta-device loading breakage." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version compatibility and autocast dtype mismatch are separate issues." + "reason": "Torch 2.10 compatibility is unrelated to the specific autocast dtype mismatch bug." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Different regressions: Whisper pipeline return_language vs tokenization changes." + "reason": "Whisper pipeline `return_language` regression is unrelated to tokenization changes." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage object error is unrelated to mmap-induced OOM." + "reason": "Different low-level failures: missing Float8 storage object vs mmap/OOM behavior." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Different subsystems and root causes; no shared concrete code-path problem." + "reason": "Different issues: compiled flex_attention on torch>=2.9 vs Gemma4 device mismatch during forward." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The analysis is grounded in the packet and stays appropriately conservative. The accepted pair, issue:45276 / issue:45292, is well supported by the titles and high similarity score, and the broader/narrower framing is reasonable. The remaining soft-edge rejections avoid overclaiming duplicate status and the summary does not overstate the evidence." + "accept": false, + "feedback": "The Gemma4 resize_token_embeddings pair looks grounded, but the summary overreaches on the Transformers 5.0.0 custom-model initialization claim: the packet does not provide a supported duplicate pair for that bug, and the listed soft pair for nearby issues was rejected. Please keep the summary limited to what the packet actually supports and avoid asserting an additional true duplicate without evidence." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2498,232 +2456,232 @@ "issue:43716|issue:45084", "issue:44206|issue:44220" ], - "prepared_review_unit_hash": "6c9e8b10757d00e505d18c81f4ee28157713f606ee2ec5b73f7482b41763de8f", + "prepared_review_unit_hash": "3dec5884f03e9d20aa7f1914a837c8fd580a25685b8377d87b926622b61eedf0", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37645, - "estimated_input_tokens": 9412, - "estimated_eval_tokens": 19080 + "serialized_chars": 37642, + "estimated_input_tokens": 9411, + "estimated_eval_tokens": 19078 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose similarity group rather than a true duplicate set. The only relatively tight subgroup is around Transformers v5 model-loading/initialization regressions, and issue 43611 is the best canonical representative; none of the soft edges look like exact duplicates.", - "confidence": 0.71, - "canonical_issue_reason": "43611 is the clearest concrete regression in the small v5 model-loading/init subgroup and has the most direct, actionable failure description; related issues nearby are thematically similar but not the same bug.", + "summary": "This cluster is a loose collection of Transformers v5 regression issues, mostly around model/tokenizer loading and initialization, but the soft pairs generally point to different concrete failures rather than the same underlying bug.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43611 is the broadest and most representative loading-regression report in the set, centered on v5 breakage in model loading with a clear generic symptom.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the most representative issue for triage: specific, reproducible, and centered on a concrete loading path rather than a broad symptom or docs-only mismatch.", + "best_issue_reason": "issue:43611 is a strong anchor for the cluster because it describes a reusable v5 model-loading regression and is closer to the recurring theme than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43931", "right": "issue:44863", "accept": false, - "reason": "Both are model-loading failures, but they involve different architectures and failure modes (shape mismatch vs checkpoint load support), so not the same bug." + "reason": "Both are loading failures, but they concern different models and different failure modes (shape mismatch vs checkpoint incompatibility)." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "One is an isolated reproduction of an older issue; the other is a Qwen3.5 cache crash. Same general area, different concrete code paths." + "reason": "An isolated reproduction of an earlier issue is unrelated to the Qwen3.5 PagedAttentionCache linear_attention crash." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both are v5 initialization/loading regressions, but one is about base_model_prefix loading and the other about custom model init in notebooks; different failures." + "reason": "Both involve v5 model initialization/loading regressions, but one is base_model_prefix loading and the other is notebook custom-class initialization; not the same bug." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch incompatibility and Gemma4 device mismatch are unrelated concrete problems; shared runtime context is not enough." + "reason": "Torch version compatibility and Gemma4 device-map forward mismatch are different problems." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "These are different tokenizer/cache failures with different triggers and stack paths, not one underlying bug." + "reason": "Different subsystems and symptoms: cache group-type handling vs template compilation error." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Version incompatibility versus image preprocessor/model dtype mismatch are distinct issues, despite both involving runtime execution." + "reason": "A framework version incompatibility is not the same as a multimodal dtype mismatch." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression and gpt-oss MXFP4 fallback behavior are unrelated model/pipeline issues." + "reason": "Whisper pipeline return_language regression and MXFP4 quantization fallback are unrelated." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs mismatches, but they concern different removed/changed pipeline docs and are not the same documentation bug." + "reason": "Both are docs issues, but they cover different pipeline/task deprecations and are not the same change." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export export failure with torch_compilable_check is unrelated to Perceiver interpolation failure; different code paths and symptoms." + "reason": "torch.export with torch_compilable_check is unrelated to Perceiver interpolation/resolution failures." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen/VL loading regressions, but one is a Transformers v5 breakage and the other is a specific checkpoint shape mismatch." + "reason": "Both involve Qwen3-VL loading, but one is a dev0 regression on MoE models and the other is a specific weight-shape mismatch; not clearly the same underlying bug." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression for Qwen VL models and a generic mask shape IndexError are not the same underlying bug." + "reason": "Video input regression for Qwen VL models is unrelated to a generic mask-shape IndexError." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "SigLIP2 bad outputs and ModelOutput key handling are different problems; one is inference quality, the other is output container behavior." + "reason": "A model producing bad outputs is different from ModelOutput key assignment behavior." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline behavior and CamemBERT masked LM predictions are unrelated model-specific regressions." + "reason": "Whisper pipeline regression and CamemBERT masked-LM prediction changes are different code paths and symptoms." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Storage lookup failure and template-node compilation error are different low-level failures with no shared code path shown." + "reason": "Float8 storage lookup failure and template compilation error are unrelated." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "High perplexity on gpt-oss and ineffective gradient checkpointing with LoRA are unrelated training/evaluation problems." + "reason": "Perplexity quality on a model is not the same as gradient checkpointing being ineffective with LoRA." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch version incompatibility and Qwen3.5 embedding unpacking in TRL SFT trainer are different issues." + "reason": "Torch compatibility and Qwen3.5 embedding unpacking with TRL are different bugs." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer/model loading regressions for different checkpoints; no evidence they share the same bug." + "reason": "CamemBERT prediction regression and tokenizer loading failure for EMBEDDIA/est-roberta do not match." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "Extra_state save/load handling and composite config text_config unwrapping are separate serialization/loading defects." + "reason": "save/from_pretrained extra_state handling and AutoModelForCausalLM config unwrapping are separate serialization/config issues." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer method-name conflict and gguf_file loading breakage are unrelated loader issues." + "reason": "T5Tokenizer method conflict and gguf_file loading failures are different tokenizer/model-loading bugs." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "Deepspeed zero3 initialization IndexError and compiled flex_attention torch>=2.9 failure are different runtime failures." + "reason": "Deepspeed zero3/BertModel indexing and compiled flex_attention on newer torch versions are unrelated." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "MoE training memory usage and disabling mmap on Strix Halo are not the same bug; one is a performance issue, the other an OOM workaround." + "reason": "Both mention memory, but one is Qwen3 MoE training inefficiency and the other is mmap OOM avoidance on Strix Halo." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer NoneType broadcast error and Whisper return_language regression are unrelated failures." + "reason": "Trainer GPU broadcast TypeError and Whisper return_language regression are unrelated." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve loading/tokenizer issues, but they affect different models and different code paths." + "reason": "Tokenizer loading for est-roberta and Mistral regex patching fail in different tokenizer code paths." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "XLNet CPU arange placement and Qwen3_5Moe CPU offload incompatibility are distinct device-placement bugs." + "reason": "CPU device placement bug in XLNet and CPU-offload incompatibility in Qwen3_5Moe are different model-specific issues." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "Serialization extra_state handling and Qwen VL video regression are unrelated problems." + "reason": "Serialization extra_state handling and Qwen VL video-input regression are unrelated." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "AppleSilicon allocator warmup TypeError and _is_hf_initialized regression are different low-level issues." + "reason": "Apple Silicon allocator warmup TypeError and _is_hf_initialized flag breakage are separate regressions." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading failure for a specific model and Mistral regex patch crash are separate bugs." + "reason": "Loading cjvt/sleng-bert and Mistral regex patch crashes are unrelated tokenizer issues." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer mapping registration failure and flex_attention compilation failure are unrelated." + "reason": "TOKENIZER_MAPPING_NAMES add-new-model-like failure and compiled flex_attention failure are unrelated." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity on a model and runtime post-processor update behavior are not the same defect." + "reason": "High perplexity on a model is not the same as a runtime special-token/post-processor update bug." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are model loading reports, but they concern different models and different failure modes." + "reason": "A tokenizer loading failure for one model is not the same as a multimodal config-loading failure for another." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both sit in the v5 model-loading/init area, but one is base_model_prefix loading and the other is non-persistent buffer initialization; not the same bug." + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization; not the same underlying bug." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor dtype mismatch and template-node compilation errors are unrelated." + "reason": "Multimodal dtype mismatch and template compilation error do not share a concrete code path." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both are audio feature-extraction related, but the reported failure modes are too different and the second is too vague to confirm the same bug." + "reason": "Unsupported center argument in feature extraction and _torch_extract_fbank_features issues are related in area but not evidently the same bug." } ] }, @@ -2832,248 +2790,251 @@ "issue:44062|issue:44987", "issue:36032|issue:43742" ], - "prepared_review_unit_hash": "d1e0d6ff181f356a063d79502975dce56f3972e191d2c24e534d97387cda9dd9", + "prepared_review_unit_hash": "7f02c409a302593a65e150ced39e40f3c9dfb3ab9353453310fdc1d4018d75b6", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 35, - "serialized_chars": 37697, - "estimated_input_tokens": 9425, - "estimated_eval_tokens": 19106 + "serialized_chars": 37696, + "estimated_input_tokens": 9424, + "estimated_eval_tokens": 19104 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated bugs across different models/code paths. The only strong duplicate-like pair is the Qwen3.5 save_pretrained regression (45216/45357); the rest should stay separate.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45216 is the broadest description of the Qwen3.5 save_pretrained regression and cleanly subsumes the more specific follow-up in 45357.", + "summary": "These items are mostly distinct regression/compatibility reports across different model families and APIs. A few are thematically related (especially around Transformers v5 model-loading and initialization), but none of the soft pairs look like the same underlying bug/change.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 43646 is the broadest representative of the cluster\u2019s recurring v5 initialization regressions, so it is the best single anchor if one issue must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the best representative of the one clearly related issue pair in this set, with 45357 reading like a narrower continuation of the same save_pretrained bug.", + "best_issue_reason": "Issue 43646 most clearly captures a central, reusable regression theme (custom model initialization breaking in v5) and is more general than the narrower model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Different failures: missing all_tied_weights_keys vs a model-load key error for a specific checkpoint." + "reason": "Both are model-loading failures, but they involve different models and different symptoms (missing all_tied_weights_keys vs failing to load a specific BERT checkpoint)." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "Unrelated code paths: AutoConfig trust_remote_code field loss vs Qwen3.5 PagedAttentionCache group-type crash." + "reason": "These are unrelated failures: missing fields from trust_remote_code config loading versus a PagedAttentionCache/group-type crash in Qwen3.5." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Trainer token-count tracking is unrelated to Qwen2.5-VL iterable dataset multi-image handling." + "reason": "Trainer token counting not updating and Qwen2.5-VL IterableDataset multi-image failure are different code paths and bug classes." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer special-token constructor conflict is unrelated to Qwen3Moe FP8 config loading." + "reason": "Tokenizer class construction error and Qwen3Moe FP8 config failure are unrelated issues despite both being loading problems." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both are Whisper-related, but one is batch_decode skip_special_tokens handling and the other is empty-transcription generation after align_special_tokens; not the same bug." + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription generation after align_special_tokens are distinct processor bugs." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is non-persistent buffers and the other is custom model initialization; different underlying problems." + "reason": "Related v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; not the same bug." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Mixtral torch dynamo grouped_mm tracing failure is unrelated to a template-node compilation TypeError." + "reason": "Torch dynamo tracing on Mixtral grouped_mm and template-node compilation errors are separate compilation failures." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Two distinct model-loading failures for different checkpoints and causes." + "reason": "Both are loading failures, but they affect different checkpoints and likely different root causes." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both mention GGUF loading, but one is a generic gguf_file API break and the other is a specific Qwen2.5-GGUF regression; not enough to treat as the same bug." + "reason": "gguf_file API breakage and Qwen2.5-GGUF loading failure overlap in area but are not the same concrete failure." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen3.5 linear_attention PagedAttentionCache crash are different defects." + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache linear_attention crash are different model/runtime bugs." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Dequantized model device_map offload failure is unrelated to Qwen2.5-GGUF loading." + "reason": "gpu+cpu offload loading of dequantized models and Qwen2.5-GGUF v5 loading failure are different loading paths with different symptoms." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Different Qwen issues: output_hidden_states bug vs still-image position-id scaling bug." + "reason": "Qwen3.5 output_hidden_states regression and Qwen2.5-VL rope-index temporal scaling bug are unrelated model logic issues." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to notebook custom-model initialization." + "reason": "Tokenizer method conflict and notebook custom model initialization breakage are separate regressions." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Molmo tied-weights attribute error is unrelated to tokenizer loading for est-roberta." + "reason": "Both concern loading/serialization in v5, but one is an attribute error on a model class and the other is a tokenizer load failure for a specific checkpoint." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Two separate model/tokenizer loading failures for different repos." + "reason": "Both mention v5 loading failures, but they are different artifacts and likely different breakpoints (model vs tokenizer, different checkpoints)." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Compilation/template-node error is unrelated to GraniteMoEHybrid invoking an invalid method." + "reason": "Template compilation failure and GraniteMoEHybrid invalid method call are not the same bug or code path." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both involve lm_head/tied weights, but one is a missing tie in Mistral3 and the other is device-dependent serialization inconsistency; related theme, not the same bug." + "reason": "Weight-tying bug in Mistral3ForConditionalGeneration and inconsistent lm_head serialization are related only by weights, not the same failure." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio chat-template sampling-rate default and MIMI padding-mask mismatch are different failures." + "reason": "Audio apply_chat_template sampling-rate default and MIMI batched-vs-single padding-mask mismatch are unrelated processor/encoder bugs." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization breakage is distinct from non-persistent buffer junk initialization." + "reason": "Custom model initialization breakage and non-persistent buffers filled with junk are both v5 regressions but distinct issues." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Siglip2 nonsensical outputs and OmDet-Turbo processor size mismatch are unrelated model/processor bugs." + "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor/model size mismatch are different vision-model problems." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Jupyter custom-model initialization regression is not the same as non-persistent buffer corruption." + "reason": "Notebook model initialization breakage and junk-filled non-persistent buffers are separate v5 regressions, not one bug." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Generic loading noise and Mistral regex patch crash are different problems." + "reason": "Generic loading noise and Mistral regex patch crash are unrelated; one is UX/noise, the other is an AttributeError in tokenizer patching." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-processing ValueError are separate multimodal issues." + "reason": "Llava Next segmentation fault and Qwen3OmniMoe video-input ValueError are different multimodal loading/processing failures." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI batched-vs-single padding issue is unrelated to DAC STE mismatch." + "reason": "MIMI padding-mask batching bug and DAC STE mismatch are both model correctness issues, but on different models and mechanisms." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3Moe FP8 config and template-node compilation are different code paths." + "reason": "Qwen3Moe FP8 config failure and template-node compile failure are unrelated errors in different subsystems." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT load failure is unrelated to the Mistral tokenizer regex crash." + "reason": "Loading a specific BERT model and crashing in fix_mistral_regex are different tokenizer-loading failures with different roots." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export with torch_compilable_check is unrelated to OmDet-Turbo processor image sizing." + "reason": "torch.export export failure with torch_compilable_check and OmDet-Turbo image-size mismatch are unrelated bugs." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "_torch_extract_fbank_features and compiled flex_attention on torch>=2.9 are different low-level failures." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Different failures for different models: loading physical-intelligence/fast vs Mistral regex patch crash." + "reason": "Model loading failure for a checkpoint and Mistral tokenizer regex patch crash are separate issues." }, { "left": "issue:45216", "right": "issue:45357", - "accept": true, - "reason": "Both describe the same Qwen3.5 save_pretrained serialization regression, with 45357 looking like a narrower follow-up about the visual encoder keys." + "accept": false, + "reason": "Both are Qwen3.5 save_pretrained regressions, but one concerns checkpoint correctness in general and the other incorrect visual-encoder keys specifically." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized flag breakage and mmap OOM avoidance are unrelated." + "reason": "_is_hf_initialized flag breakage and disabling mmap on Strix Halo to avoid OOM are unrelated bugs." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Tied lm_head weight bug in Mistral3 is not the same as LoRA merge collapse with extended vocabulary." + "reason": "Weight-tying in Mistral3ForConditionalGeneration and LoRA merge collapse with extended vocab are different tied-weights problems with different triggers." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Qwen3Moe FP8 loading and init_empty_weights unexpected argument are different regressions." + "reason": "Qwen3Moe FP8 config failure and init_empty_weights unexpected _is_hf_initialized argument are separate v5 compatibility issues." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken special-argument conflict is unrelated to loading physical-intelligence/fast." + "reason": "AddedToken special keyword conflict and model loading failure for physical-intelligence/fast are different tokenizer/model load errors." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "T5Tokenizer add_special_tokens conflict is unrelated to MobileLLM-125M key error." + "reason": "T5 tokenizer method conflict and MobileLLM-125M key error are unrelated loading failures." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded and conservative overall. The only clearly related pair in the packet is issue:45216 / issue:45357, which both describe the Qwen3.5 save_pretrained regression. The other soft-edge rejections stay appropriately cautious and do not overclaim 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 }, @@ -3143,8 +3104,8 @@ "issue:43296|issue:43531", "issue:44479|issue:45325", "issue:42915|issue:44589", - "issue:43531|issue:44877", "issue:43611|issue:44534", + "issue:43531|issue:44877", "issue:40444|issue:45325", "issue:36032|issue:41720", "issue:40990|issue:44811", @@ -3173,14 +3134,14 @@ "issue:41720|issue:45081", "issue:43644|issue:44079" ], - "prepared_review_unit_hash": "f8eb881f7354678f1e6530701a9135ce121719e506875f8ad2492d81687999cf", + "prepared_review_unit_hash": "6ff7c7438dde47f96e0241cb4a3b1759b5f7ccc643ce45d2e56b21e49ba70db5", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -3195,204 +3156,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. The soft pairs are generally only loose thematic matches, not strong duplicates; they span unrelated regressions across tokenizers, model loading, vision/video, packaging, and backend setup.", - "confidence": 0.62, - "canonical_issue_reason": "issue:45020 is the broadest open report in the set and best fits as a representative umbrella for recent Transformers-version regressions; the other issues are mostly narrow, model-specific failures.", + "summary": "This cluster is mostly a loose pile of unrelated Transformers regressions that all happen to mention v5/v5.3 loading or model behavior. Only one soft pair looks like a true duplicate: the non-persistent-buffer corruption regression.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the clearest duplicate anchor: it states the exact non-persistent-buffer corruption bug, and issue:44534 is the same regression with a slightly reworded title.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the most suitable representative issue because it has the widest scope and is still open, whereas the rest are more specialized bug reports that do not collapse into one concrete defect.", + "best_issue_reason": "issue:43644 is the best representative issue here because it matches the exact duplicate pair and is more specific than the other broad v5-regression reports in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "Different models and failures: PaddleOCR-VL load error vs Qwen3-MoE sliding-window behavior." + "reason": "Different problems: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen VL regressions, but one is a broad video-input regression across models and the other is a specific still-image rope-index scaling bug." + "reason": "Related vision/VL area, but different concrete bugs: broad video-input regression vs still-image position_id scaling." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "Qwen3-MoE FP8 config failure and Float8 storage lookup error are unrelated symptoms." + "reason": "Different failure modes: FineGrainedFP8Config issue vs float8 storage lookup TypeError." }, { - "left": "issue:43531", - "right": "issue:44877", + "left": "issue:43611", + "right": "issue:44534", "accept": false, - "reason": "Qwen3-MoE sliding_window bug vs granite_speech config strictness; different code paths." + "reason": "Both are v5 regressions, but they describe different bugs and code paths: base_model_prefix loading vs non-persistent buffer corruption." }, { - "left": "issue:43611", - "right": "issue:44534", + "left": "issue:43531", + "right": "issue:44877", "accept": false, - "reason": "base_model_prefix loading failure and non-persistent buffer junk are distinct v5 regressions." + "reason": "Unrelated models and failures: sliding_window on Qwen3-MoE vs strict config blocking granite_speech." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Multi-image iterable-dataset finetuning failure is not the same bug as still-image temporal position scaling." + "reason": "Both involve VL models, but one is multi-image IterableDataset training and the other is temporal position_id scaling." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer method-name conflict and Qwen3 auto device-map cuda assert are unrelated." + "reason": "Tokenizer AttributeError vs device-mapping cudaErrorAssert; no shared underlying bug." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "High perplexity on gpt-oss-20b is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "Different functionality: perplexity evaluation issue vs Whisper batch_decode skip_special_tokens bug." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "TensorFlow backend packaging issue vs noisy model-loading output; same area, different bug." + "reason": "Different issue types: torch/backend compatibility vs logging noise when loading a transformer." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "AutoTokenizer error messaging for Voxtral and Apple Silicon warmup TypeError are different failures." + "reason": "AutoTokenizer error messaging vs Apple Silicon warmup TypeError are unrelated." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "_is_hf_initialized flag breakage is not the same as junk-filled non-persistent buffers." + "reason": "Both are v5 regressions, but they affect different internals: _is_hf_initialized flag vs non-persistent buffers." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states KeyError; unrelated internals." + "reason": "Different model families and bugs: ModelOutput key assignment vs Gemma-4 FSDP2 shared_kv_states reconstruction." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both mention TensorFlow/backend packaging, but one is missing TF in the torch extra and the other is a tensorboard-only/no-TF backend error; not the same concrete defect." + "reason": "Packaging/backend installation issue vs tf backend undefined error; not the same bug." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Ignoring tokenizer.json and wrong Granite token IDs are related to AutoTokenizer, but they are not the same bug." + "reason": "Both tokenizer-related, but one is repo tokenizer.json resolution and the other is Granite token ID regression." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime special-token post-processing and add_special_tokens=True not adding BOS/EOS are different tokenizer issues." + "reason": "Different tokenizer runtime behaviors: post-processor update vs add_special_tokens not adding BOS/EOS." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper runtime failure and Float8 storage TypeError are unrelated." + "reason": "Whisper runtime failure vs float8 storage object error; unrelated." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA temporal-attention/weight-sharing defect and Qwen3.5 save_pretrained key regression are separate problems." + "reason": "Different Video-LLaVA/Qwen3.5 issues: missing temporal attention/shared weights vs wrong visual encoder keys on save_pretrained." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "device_map=auto offload failure and Qwen3 MoE memory inefficiency are different training/loading issues." + "reason": "Different loading/offload and training-memory problems; same broad area only." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "General loading noise vs tokenizer loading failure for EMBEDDIA/est-roberta are not duplicates." + "reason": "Loading noise vs tokenizer loading failure for a specific model; not duplicates." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and remote_code loading regressions are unrelated." + "reason": "Torch >=2.9 compiled flex_attention failure is unrelated to remote_code model breakage." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Qwen2-VL multi-GPU training and Qwen2.5-VL multi-image iterable finetuning are different bugs." + "reason": "Both mention multi-image training, but they are different model-specific failures with different setups." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading failure and CLIPTokenizer max-length behavior are unrelated." + "reason": "Qwen2.5-GGUF loading failure vs CLIPTokenizer model_max_length bug; different issues." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Flash-attention training degeneration and single-label zero-loss classification are different failures." + "reason": "Both are training/classification-related, but one is flash_attention_3 degeneration and the other is single-label zero-loss behavior." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is not the same as multi-image iterable dataset finetuning." + "reason": "Different generation/training contexts and different bugs." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Unexpected _is_hf_initialized argument and missing Float8 storage object are separate v5 breakages." + "reason": "init_empty_weights/_is_hf_initialized TypeError vs float8 storage lookup error; different code paths." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer junk are unrelated internals." + "reason": "ModelOutput key assignment bug is unrelated to non-persistent buffer corruption." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "GGUF loading failure and missing pad_token_id on Llama4Config are different model-loading issues." + "reason": "Different models and failures: Qwen2.5-GGUF loading vs Llama4Config missing pad_token_id." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI padding-mask mismatch and Gemma3 sliding-window NaNs are distinct sequence-processing bugs." + "reason": "Both involve batching, but one is MIMI padding-mask mismatch and the other is Gemma3 NaN embeddings with sliding-window attention." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading breakage and ModelOutput key assignment do not share a concrete code path." + "reason": "FSDP_CPU_RAM_EFFICIENT_LOADING breakage vs ModelOutput key assignment; not the same bug." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window and _is_hf_initialized init_empty_weights issues are unrelated." + "reason": "Qwen3-MoE sliding_window issue vs init_empty_weights argument mismatch are unrelated." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and Mistral regex patch crash are unrelated." + "reason": "cudaErrorAssert on auto device mapping vs Mistral regex patch crash; unrelated." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer junk and ModelOutput key assignment are separate regressions." + "reason": "Non-persistent buffer corruption and ModelOutput key assignment are different bugs." } ] }, @@ -3500,20 +3461,20 @@ "issue:39692|issue:40444", "issue:43572|issue:45759" ], - "prepared_review_unit_hash": "7e74b579cac3e9664cefa4313a0559ce4b6dfaaa2a4c4b08f565ccc49d4394dc", + "prepared_review_unit_hash": "c67d1b240a71fbc1ae5845d05ae4eac54d8e0028d6cd1bb3982d85e41344111f", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37521, + "serialized_chars": 37524, "estimated_input_tokens": 9381, "estimated_eval_tokens": 19018 }, @@ -3522,205 +3483,220 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated model-specific regressions with some superficial overlap around v5 loading/offload behavior; I would not merge any soft pair as duplicates.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43611 is the most concrete representative of the broader Transformers v5 load/init regression theme; it has a clear failure mode and could serve as the best hub issue in this set.", + "summary": "This cluster is mostly a grab-bag of unrelated bug reports across models and framework regressions; I only found one likely duplicate pair in the Qwen2.5-VL position-id/rope-index area.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45325 is the clearest root report for the Qwen2.5-VL position-id/rope-index regression; issue:45381 looks like the same underlying bug surfaced through video input.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43611 is the strongest overall issue candidate because it is specific, actionable, and closest to the cluster\u2019s repeated v5 model-loading breakages.", + "best_issue_reason": "issue:45325 is the best anchor in this set because it names the concrete function/regression most directly and matches the accepted duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Different bugs: multi-image finetuning with IterableDataset vs Qwen2.5-VL video position-id computation." + "reason": "Both are Qwen2.5-VL-related, but one is a finetuning/IterableDataset multi-image failure and the other is a video position-id bug; different concrete problems." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "Unrelated failure modes: torch.export/torch_compilable_check vs bfloat16 dtype mismatch in inference." + "reason": "torch.export with torch_compilable_check is unrelated to bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "One is a v5 flag/init regression, the other is a dtype mismatch; no shared code-path bug." + "reason": "_is_hf_initialized init behavior is a different regression from dtype mismatches during inference." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Packaging/backend dependency issue vs model loading failure in vLLM; not the same bug." + "reason": "Missing TensorFlow in the torch extra and PaddleOCR-VL loading are unrelated install/load issues." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is custom model initialization; too different to merge." + "reason": "Both mention Transformers v5 model loading/init breakage, but they describe different failures and code paths." }, { "left": "issue:45325", "right": "issue:45381", - "accept": false, - "reason": "Both touch Qwen2.5-VL position ids, but one is still-image temporal scaling and the other is video vision_position_ids; not clearly the same bug." + "accept": true, + "reason": "Both report the same Qwen2.5-VL position-id/rope-index bug, one via still-image temporal ids and the other via video vision_position_ids." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different models and symptoms: SigLIP2 nonsensical outputs vs Perceiver non-default resolution failure." + "reason": "Different models and failures: SigLIP2 nonsensical outputs vs Perceiver resolution/interpolate_pos_encoding behavior." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both involve lm_head weights, but one is tied-weight setup causing garbage generation and the other is serialization/device-dependent inconsistency." + "reason": "One is missing/tied lm_head weights in AyaVision; the other is inconsistent serialization of lm_head.weight by device." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Qwen2.5-VL position-id bug vs Kosmos2.5 long-OCR index error; unrelated." + "reason": "Qwen2.5-VL position-id handling and Kosmos2.5 long-OCR index errors are different model-specific bugs." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper runtime failure vs meta-device loading regression; different models and code paths." + "reason": "Whisper runtime failure is not the same as meta-device loading regressions in other models." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Different model families and failures: FineGrainedFP8Config on Qwen3Moe vs GGUF loading on v5." + "reason": "FineGrainedFP8Config failure for Qwen3Moe and GGUF loading failure in Qwen2.5 are different issues." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "One is int4 quantization memory growth, the other is a generic video-model OOM; not the same concrete issue." + "reason": "Generic OOM under int4 quantization is not the same as SAM3Video running out of memory." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "Model load failure in vLLM vs a torch/transformers version incompatibility; no shared bug." + "reason": "PaddleOCR-VL/vLLM loading and torch/Transformers version incompatibility are unrelated." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Different loading failures for different models; only broad 'cannot load' similarity." + "reason": "Different model-loading failures with no shared concrete bug." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint key mismatch and Qwen2.5-VL position-id scaling are unrelated." + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL rope-index calculation." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Qwen VL video-position regression vs Gemma4 device mismatch; different model-specific bugs." + "reason": "Video regression in Qwen models and Gemma4 PLE device mismatch are different code-path problems." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected `_is_hf_initialized` arg during init_empty_weights vs strict config rejection for granite_speech; distinct issues." + "reason": "Unexpected _is_hf_initialized argument during init and strict config rejection are separate regressions." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on gpt-oss and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are different bugs." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch vs torch.export failure; unrelated." + "reason": "CPU-offload device mismatch for Bark and torch.export failures are unrelated." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offload behavior vs tokenization changes; no common underlying bug." + "reason": "Offloading/quantization problems and tokenization changes are too broad to be the same concrete bug." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "BarkProcessor voice preset bug vs progressive generation inputs_embeds/past_key_values issue; unrelated." + "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer padding/truncation crash vs torch.export/torch_compilable_check failure; different components." + "reason": "Tokenizer padding/truncation crashes are unrelated to torch.export/torch_compilable_check failures." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both mention offloading, but one is dequantized model loading on gpu+cpu offload and the other is quantization/offloading behavior; not the same concrete bug." + "reason": "Both mention offloading, but one is dequantized model loading with device_map=auto and the other is a broad quantization/offloading complaint." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation bug vs Qwen3 auto device map CUDA assert; unrelated." + "reason": "Progressive generation cache handling and Qwen3 auto device-map CUDA asserts are different failures." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "CPU offload device mismatch vs DAC.from_latents/STE mismatch; different code paths." + "reason": "Bark CPU-offload mismatch and DAC.from_latents STE mismatch do not share the same code-path problem." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Tied lm_head weights bug vs Bamba SDPA/torch.compile failure; unrelated." + "reason": "AyaVision lm_head tying and Bamba torch.compile/SDPA issues are unrelated." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Different models and failure signatures: cudaErrorAssert on A800 vs Gemma4 PLE device mismatch." + "reason": "Both involve device placement, but the model families and failure modes are different." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI padding-mask batching bug vs OmDet-Turbo processor size mismatch; unrelated." + "reason": "MIMI padding-mask divergence and OmDet-Turbo processor/model image-size mismatch are unrelated." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Weight-tying/serialization issue vs dtype mismatch in inference; not the same bug." + "reason": "lm_head weight tying/serialization and bf16 dtype mismatches are different issues." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Batching/padding-mask encoder bug vs Perceiver interpolation/resolution bug; different problems." + "reason": "Different models and problems: padding-mask behavior vs Perceiver resolution handling." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors vs Qwen2.5-VL multi-image finetuning failure; different model families and symptoms." + "reason": "SigLIP2 docs/example errors and Qwen2.5-VL multi-image finetuning failures are not the same bug." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig vs AutoModelForCausalLM.from_config not unwrapping text_config; unrelated loading/config regressions." + "reason": "Missing pad_token_idx in StableLmConfig and text_config unwrapping for composite Qwen configs are different regression classes." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pair rejections are conservative and the only accepted edge, issue:45325 \u2194 issue:45381, is plausibly the same Qwen2.5-VL positional-index regression by title and scope. I don't see an overstatement that would justify rejecting this review." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -3820,226 +3796,226 @@ "issue:45081|issue:45812", "issue:43295|issue:45701" ], - "prepared_review_unit_hash": "78fc81e1e83a11a278348d90f99a5621353e824aac0890f2051909002342dd04", + "prepared_review_unit_hash": "a8a75d2178fdbd3592ae5c485596274578ef9341d9bc696e1ef17638088fb7ba", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37479, - "estimated_input_tokens": 9370, - "estimated_eval_tokens": 18996 + "serialized_chars": 37484, + "estimated_input_tokens": 9371, + "estimated_eval_tokens": 18998 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a collection of distinct Transformers bug reports rather than true duplicates. The strongest hub candidate is the still-open quantization/offloading report, but none of the soft-edge pairs are close enough to merge.", - "confidence": 0.76, - "canonical_issue_reason": "issue:43873 is the best duplicate hub: it is still open, has meaningful discussion/inbound references, and describes a broad, user-facing offloading/quantization failure that is likely to attract related reports.", + "summary": "The cluster is heterogeneous: trainer API regressions, tokenizer issues, model-loading/offload bugs, and unrelated model-specific failures. The soft pairs are mostly thematic overlaps rather than true duplicates, so no pair should be merged as the same underlying issue.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items cover distinct bugs across training, tokenization, loading/offload, and model-specific runtime paths.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the most suitable canonical issue because it is active, broadly scoped, and already has cross-references, making it a better triage target than the narrower one-off reports.", + "best_issue_reason": "Issue 43873 is the broadest open report with multiple references, but it still is not a safe canonical for the rest of this heterogeneous set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "One is a processor label-reduction bug; the other is a missing tied-weights attribute on a causal LM. Different code paths and symptoms." + "reason": "Different bugs: BeitImageProcessorFast label reduction vs Molmo missing all_tied_weights_keys." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "Both mention loading/runtime regressions, but one is about FSDP CPU RAM-efficient loading and the other about non-persistent buffers in v5. Not the same bug." + "reason": "Different mechanisms: broken FSDP CPU-efficient loading vs non-persistent buffer initialization junk." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "Unrelated failures: one is a model loading issue for glm-4v-9b, the other is tokenizer loading for a specific roberta model." + "reason": "Different symptoms and code paths: GLM-4V load failure vs tokenizer loading regression for est-roberta." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both involve save/serialization regressions, but one is about lm_head weight consistency and the other about incorrect visual encoder keys. Different underlying defects." + "reason": "Both involve tied/serialized lm_head weights, but one is generic v5 serialization/device behavior and the other is a Qwen3.5 visual encoder save regression." }, { "left": "issue:43296", "right": "issue:44877", "accept": false, - "reason": "Different components and failure modes: vLLM loading of PaddleOCR-VL vs strict config blocking granite_speech config." + "reason": "Model loading failures, but one is PaddleOCR-VL in vLLM and the other is strict config parsing for granite_speech." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "AyaVision lm_head tying and BitNet packed-weight unpacking are separate model-specific bugs with different causes." + "reason": "Both mention weight handling, but AyaVision lm_head tying and BitNet packed-weight unpacking are separate bugs." }, { "left": "issue:43792", "right": "issue:44683", "accept": false, - "reason": "Whisper-large-v2 runtime failure and compiled flex_attention on torch>=2.9 are unrelated." + "reason": "Whisper model loading failure is unrelated to compiled flex_attention failing on torch >= 2.9." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Both are CUDA/runtime crashes, but one is auto device mapping for Qwen3 and the other is a Llava Next load segfault. Too different to merge." + "reason": "Both are CUDA crashes, but one is Qwen3 auto device mapping and the other is Llava Next loading segfaults." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Both affect Qwen-family models, but one is an A800 cudaErrorAssert during auto device mapping and the other is a flash-attention generation bug. Not the same issue." + "reason": "Different Qwen3-family failures: cudaErrorAssert during device mapping vs flash-attention generation bug." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Both are model-forward device/variable issues, but Gemma 3n audio mask collision is unrelated to Gemma4 PLE device mismatch." + "reason": "Both are Gemma-family, but one is a variable-name collision in forward() and the other is a device mismatch in PLE during forward." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "One is high perplexity for gpt-oss on WikiText-2; the other is a tokenization-change regression. Same broad area, but not the same bug." + "reason": "Both discuss tokenizer/tokenization, but one is perplexity quality on a specific model and the other is a broad tokenization-version change report." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Both concern save/load behavior, but one is a trust_remote_code cache-path collision and the other is permission propagation during save_pretrained." + "reason": "Both concern save/load workflows, but cache-path collisions and read-only permission propagation are unrelated." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Both touch multimodal weight handling, but one is lm_head tying in Mistral3 and the other is incorrect visual encoder keys in Qwen3.5. Different defects." + "reason": "Both touch weight tying in multimodal models, but Mistral3 lm_head tying and Qwen3.5 visual encoder save keys are distinct." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "A script execution problem vs a FineGrainedFP8Config model failure; no concrete shared code-path." + "reason": "Both are runtime failures, but 3d_parallel.py execution and Qwen3Moe FineGrainedFP8Config errors are separate problems." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel inference." + "reason": "Both are dtype/processing-related, but tokenizer crashes and dtype mismatches in CI are not the same bug." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Different bug classes: BLIP2 dtype propagation vs Trainer resume/max_steps calculation." + "reason": "BLIP2 dtype propagation and Trainer resume/max_steps calculation are unrelated." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both involve memory behavior, but one is int4 quantization OOM in Transformers 5.x and the other is an apex RMSNorm memory leak in T5." + "reason": "Different memory issues: int4 quantization OOM in Transformers v5 vs apex FusedRMSNorm memory leak in T5." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both mention offloading/memory pressure, but one is quantization offloading behavior and the other is mmap on Strix Halo. Not the same failure." + "reason": "Both relate to offloading/memory, but quantization offload behavior and Strix Halo mmap OOM are not the same failure." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Continuous-batching multimodal crash with a string tensor conversion error is unrelated to a Gemma4Processor missing _tokenizer attribute." + "reason": "Different serve-path crashes: multimodal continuous batching vs Gemma4Processor missing _tokenizer." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech and tokenizer loading regression for EMBEDDIA/est-roberta are separate problems." + "reason": "Strict config loading and tokenizer loading regressions are separate issues." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "CustomTrainer.compute_loss signature mismatch and a 3d_parallel.py runtime failure are unrelated." + "reason": "Trainer compute_loss API breakage is unrelated to the 3d_parallel.py execution failure." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "A Trainer API compatibility error vs a Qwen3Moe FP8 loading failure; different issues." + "reason": "Unexpected num_items_in_batch kwarg in Trainer is unrelated to Qwen3Moe FP8 loading/runtime failure." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batched-shape bug does not match a version-dependent tokenization regression." + "reason": "Video processor batch-shape bug and broad tokenization-change report are not the same code-path." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "Both are model/runtime failures, but one is torch.compile with SDPA on Bamba and the other is packed-weight unpacking in BitNet. Different root causes." + "reason": "Both involve model loading/compile paths, but torch.compile SDPA failure for Bamba and BitNet packed-weight unpacking are separate." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both concern image processor sizing, but they affect different models and different expected resolutions; not the same underlying bug." + "reason": "Processor output size mismatch for OmDet-Turbo is unrelated to Perceiver non-default-resolution failure." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crash vs processor output-size mismatch; no shared code-path." + "reason": "NER/batched-padding tokenizer crash and OmDet processor resolution mismatch are different processor bugs." }, { "left": "issue:44479", "right": "issue:44918", "accept": false, - "reason": "Qwen video-input regression and Qwen3.5 embedding unpacking with TRL SFT are distinct multimodal/training issues." + "reason": "Video-input regression in v5.3.0 and unpacking Qwen3.5 embeddings with TRL SFT trainer are different failures." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "FineGrainedFP8Config failure and device_map=auto dequantized-model offload failure are different loading bugs." + "reason": "Qwen3Moe FineGrainedFP8 failure and device_map=auto dequantized offload failure are separate loading paths." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Auto device mapping CUDA assert and a failing parallel script are unrelated." + "reason": "CUDA assert on Qwen3 auto mapping is unrelated to not being able to run 3d_parallel.py." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Different quantized/offload problems: Bamba torch.compile+SDPA vs Bark CPU offload device mismatch." + "reason": "Both mention offload/compile-ish behavior, but Bamba SDPA torch.compile failure and Bark CPU-offload device mismatch are not the same bug." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral regex patch crash and Granite tokenizer ID regression are different tokenizer bugs." + "reason": "Tokenizer patch crash on Mistral and Granite token ID regression are separate tokenizer bugs." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression in v4.57.5 and a general tokenization-change report are not the same concrete defect." + "reason": "Processor.tokenizer regression is a custom model API break, not the same as a general tokenization-change report." } ] }, @@ -4131,8 +4107,8 @@ "issue:44779|issue:45812", "issue:41720|issue:43746", "issue:43606|issue:43720", - "issue:43550|issue:43653", "issue:43873|issue:45704", + "issue:43550|issue:43653", "issue:44514|issue:44545", "issue:44625|issue:45479", "issue:43531|issue:45759", @@ -4142,214 +4118,208 @@ "issue:34567|issue:35707", "issue:45759|issue:45859" ], - "prepared_review_unit_hash": "b6d980ac51cabd8995c1e5cfbc77c5436a6d6f24efc097df9e08494a54a565c8", + "prepared_review_unit_hash": "e228b4a92d1421e73aec5cbfaf7bc6e457f861209adf96852e8bbb918b667c8e", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35680, - "estimated_input_tokens": 8920, - "estimated_eval_tokens": 18096 + "serialized_chars": 35682, + "estimated_input_tokens": 8921, + "estimated_eval_tokens": 18098 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is largely heterogeneous and spans unrelated bugs across many models and subsystems. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template crash reported twice; the rest should stay separate.", - "confidence": 0.97, - "canonical_issue_reason": "issue:44514 is the earliest and more canonical report of the exact apply_chat_template crash; issue:44545 is a later duplicate with the same failure mode and title.", + "summary": "This cluster is mostly heterogeneous: config, tokenizer, generation, serving, and model-loading bugs across different model families. Only one clear duplicate pair stands out: the two Qwen2_5_VLProcessor apply_chat_template crash reports.", + "confidence": 0.95, + "canonical_issue_reason": "issue:44514 is the earlier and clearer report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input/padding=False crash as issue:44545.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the best issue to keep as the representative duplicate target because it matches the exact bug and is the original report in the duplicate pair.", + "best_issue_reason": "issue:44514 is the best canonical issue candidate because it cleanly represents the only confirmed duplicate pair and has the more stable, complete report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Both are config-related, but one is Gemma3TextConfig API mismatch and the other is GPTNeoXConfig rotary_pct persistence; different bugs." + "reason": "Different bugs: Gemma3TextConfig API mismatch vs GPTNeoX rotary_pct not persisting on reload." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast label reduction and glm-4v loading failure are unrelated failure paths." + "reason": "Unrelated issues affecting different models/components: BeitImageProcessorFast label reduction vs glm-4v-9b load failure." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM regression and Qwen3 recurrent-state reset are different model/code-path issues." + "reason": "Different model-path failures: CamemBERT masked LM regression vs Qwen3.5 recurrent state reset in cache logic." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Both mention Qwen3/MoE, but one is a cudaErrorAssert in auto device mapping and the other is a tensor-parallel limitation in GPT-OSS; not the same bug." + "reason": "Both involve Qwen/MoE, but the concrete failures differ: device-mapping cudaErrorAssert vs tensor-parallel not working." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx loss are unrelated." + "reason": "Distinct config/runtime bugs in different models: Qwen3-MoE sliding_window vs StableLmConfig pad_token_idx." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Runtime special-token post-processing and Granite tokenizer ID regression are different tokenizer bugs." + "reason": "Both are tokenizer-related, but one is runtime post-processor syncing and the other is wrong Granite token IDs; not the same bug." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "Importing GenerationMixin and running 3d_parallel.py are unrelated issues." + "reason": "Completely different problems: GenerationMixin import error vs inability to run 3d_parallel.py." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Qwen3OmniMoe video input processing and Qwen3VL weight_only loading failure are distinct multimodal loading bugs." + "reason": "Different multimodal model failures: Qwen3OmniMoe video input ValueError vs Qwen3VL weight_only/from_pretrained error." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "LayoutLMv2Tokenizer padding/truncation crash and MLukeTokenizer AttributeError are different tokenizer defects." + "reason": "Both tokenizer issues, but they affect different tokenizers and failure modes; not the same underlying bug." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated." + "reason": "Unrelated: Qwen3Moe FP8 config failure vs Llava Next segmentation fault on load." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "A dataloader label-drop bug and SAM3 text/bounding-box odd behavior are not the same underlying issue." + "reason": "Different subsystems and symptoms: gather_for_metrics label truncation vs SAM3 PCS text/bbox behavior." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve and Qwen3.5 streaming inference failure are separate problems." + "reason": "Different serve/runtime failures: Gemma4Processor missing _tokenizer vs Qwen3.5-0.8B streaming chat/completions issue." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Both concern tokenization, but one is runtime special-token updates and the other is a broader version-to-version tokenization change; no concrete same bug shown." + "reason": "The latter is a broad version-dependent tokenization regression, not the same specific special-token post-processor bug." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Void segmentation map processing and BarkProcessor voice_preset are unrelated." + "reason": "Both mention processors, but the affected processors and bugs are unrelated." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error with a CLIP model and progressive generation with inputs_embeds/past_key_values are different failures." + "reason": "No concrete overlap: MacOS bus error with CLIP vs progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo in cache strategies and outdated docs for removed pipelines are unrelated." + "reason": "Typo in cache strategies is unrelated to docs mentioning removed pipeline tasks." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Deepseek tokenizer regression and Granite AutoTokenizer wrong IDs are different model-specific tokenization bugs." + "reason": "Both are tokenizer regressions, but they concern different model families and different failure mechanisms." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 device-mapping cuda assert and GraniteSpeech PEFT local checkpoint loading are unrelated." + "reason": "Different issues: Qwen3 auto device mapping crash vs GraniteSpeech PEFT local checkpoint loading failure." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "CPU-offload device mismatch in Bark-small and BitNet packed-weight unpacking are different accelerate/loading issues." + "reason": "Both involve loading/offload paths, but the concrete bugs differ: CPU offload device mismatch vs packed BitNet weights not unpacked." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Bamba torch.compile SDPA failure and BigBirdTokenizer mask-token registration are unrelated." - }, - { - "left": "issue:43873", - "right": "issue:45704", - "accept": false, - "reason": "Quantization/offloading behavior and T5 apex memory leak are different bugs." + "reason": "Completely different areas: torch.compile/SDPA failure vs tokenizer special-token registration." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "These are the same apply_chat_template crash on batched input with padding=False, reported twice with identical failure mode." + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False; these are duplicates." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Qwen3.5 num_labels propagation and a generic zero-loss classification bug are not the same issue." + "reason": "Related to num_labels, but one is config propagation in Qwen3.5 and the other is a generic zero-loss training bug; not the same issue." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and composite Qwen config unwrapping are different code paths." + "reason": "Both involve Qwen3 variants, but one is sliding_window behavior and the other is config unwrapping in from_config." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Device-mismatch offload failure and BigBird mask-token decode issue are unrelated." + "reason": "No shared underlying bug: CPU offload device mismatch vs BigBirdTokenizer mask-token registration." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." + "reason": "Different model families and failure modes; nothing indicates the same bug." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Runtime special-token post-processing and DiaConfig save/load token-ID loss are different config/tokenizer bugs." + "reason": "One is runtime tokenizer post-processing; the other is config save/load losing custom token IDs." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counting and progressive generation cache handling are unrelated." + "reason": "Training-state accounting bug vs generation behavior bug; unrelated." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Composite config unwrapping in from_config and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different problems." + "reason": "Both are Qwen3.5-related, but one is composite config handling and the other is CPU-offload incompatibility in rotary embedding; distinct fixes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The representative duplicate choice is well grounded: issue:44514 is the earlier report and matches issue:44545 on the exact same apply_chat_template crash. The summary stays conservative about the rest of the cluster, and the soft-edge verdicts avoid over-claiming duplication where the packet only shows title-level similarity." + "feedback": "Grounded overall: the packet supports a mostly heterogeneous cluster with only one clearly confirmed duplicate pair (issue:44514 / issue:44545). The soft-edge rejections are conservative and stay within the evidence from titles/bodies. Minor caution: phrases like \"more stable, complete report\" are not directly evidenced in the packet, but they do not materially affect the clustering judgment." }, "evaluator_used": true, "retried": false, @@ -4426,8 +4396,8 @@ "issue:43531|issue:44589", "issue:44077|issue:44625", "issue:44987|issue:45042", - "issue:43421|issue:43653", "issue:44062|issue:44451", + "issue:43421|issue:43653", "issue:44918|issue:44936", "issue:45588|issue:45759", "issue:39692|issue:44610", @@ -4449,208 +4419,208 @@ "issue:43540|issue:44008", "issue:30064|issue:36010" ], - "prepared_review_unit_hash": "455ba3a5914cf15aef8bb5c83ee8537411bd500e58482d3fa84e76130b1a68c5", + "prepared_review_unit_hash": "d307e2304c3faaef1c71c358b8dcb7b6bc21a303703bc1dc293f01510f8e4c94", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36088, - "estimated_input_tokens": 9022, - "estimated_eval_tokens": 18300 + "serialized_chars": 36092, + "estimated_input_tokens": 9023, + "estimated_eval_tokens": 18302 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: these issues cover many unrelated bugs across tokenizers, multimodal models, quantization, offline mode, and training/runtime behavior. None of the soft pairs look like the same concrete underlying defect.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This cluster is highly heterogeneous: most items are unrelated bugs across different models, tokenizers, quantization/offload paths, and docs/examples. None of the soft pairs look like the same underlying fix, so I would not merge any of them as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No clear canonical issue: the items span many distinct code paths and model families, so there is no single issue that cleanly represents the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a clear canonical duplicate target; the items do not converge on one shared bug or code path.", + "best_issue_reason": "No single issue is a strong global representative; even the closest-looking pairs differ by model, failure mode, or subsystem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are model-loading failures, but they involve different models and likely different tokenizer/config regressions; not the same concrete bug." + "reason": "Both are tokenizer/model loading problems, but they affect different models and failure modes; not the same bug." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Same Qwen family, but one is a tie_word_embeddings warning during LoRA fine-tuning and the other is an incorrect visual-encoder key save_pretrained regression." + "reason": "One is a tie_word_embeddings warning during LoRA fine-tuning; the other is incorrect visual-encoder keys being saved. Different issues." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different subsystems and failure modes: image interpolation/classification vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel." + "reason": "Both involve vision/inference, but one is a positional-encoding image-size bug and the other is a bfloat16 dtype mismatch." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Both touch Qwen VL, but one is a weight_only from_pretrained loading error and the other is a video position-id bug; different code paths." + "reason": "Different Qwen variants and different failures: weight_only loading error vs wrong video vision_position_ids." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the Float8 storage type error in quantized loading." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the Float8 storage lookup error." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "One is about invalid optional post_init on patchtsmixer; the other is num_labels propagation in Qwen3.5 config." + "reason": "One is about optional post_init in patchtsmixer; the other is num_labels propagation in Qwen3.5 configs." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different problems: loading a specific model in 5.1+ vs image processors incorrectly depending on torchvision in v5.4.0." + "reason": "Different subsystems: loading a specific model on 5.1.0 vs PIL image processors incorrectly requiring torchvision." }, { - "left": "issue:43421", - "right": "issue:43653", + "left": "issue:44062", + "right": "issue:44451", "accept": false, - "reason": "Both involve special tokens, but one is runtime post-processor updates in TokenizersBackend and the other is BigBird mask-token registration/empty decode." + "reason": "AddedToken keyword collision is unrelated to the model-loading failure in EMBEDDIA/est-roberta." }, { - "left": "issue:44062", - "right": "issue:44451", + "left": "issue:43421", + "right": "issue:43653", "accept": false, - "reason": "AddedToken keyword duplication is unrelated to the ScandiBERT loading failure." + "reason": "Both involve special tokens, but one is runtime post-processor updates and the other is tokenizer registration/empty decode behavior." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "One is TRL SFT embedding unpacking; the other is trainer.evaluate failing after train\u2014different trainer states and symptoms." + "reason": "TRL embedding unpacking after SFT is unrelated to trainer.evaluate() failing after train()." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Different failures in different paths: flash_attention sink handling vs composite Qwen config unwrapping." + "reason": "Different code paths: flash-attention sink handling vs from_config not unwrapping text_config for composite Qwen models." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both mention processor/model mismatch, but one is a SigLIP2 docs example with quantization issues and the other is OmDet-Turbo image size mismatch." + "reason": "Both are vision-related, but they concern different models and different mismatches (docs/quantization vs processor input size)." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to the ZeRO-3 gradient explosion bug." + "reason": "Gradient checkpointing in eval mode and ZeRO-3 gradient explosion are separate training bugs." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both are offline tokenizer failures, but one is a generic cache-load problem and the other is specifically _patch_mistral_regex calling model_info() in offline mode." + "reason": "Both touch offline hub behavior, but one is cache loading and the other is an unconditional model_info() call in mistral regex patching." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Different tokenizer bugs: post-processor not updating on token changes vs LayoutLMv2Tokenizer crashing on NER/padding paths." + "reason": "Special-token post-processor updates are not the same as LayoutLMv2 tokenizer crashes on NER padding/truncation." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and failure modes: SigLIP2 incorrect outputs vs CamemBERT masked-LM predictions regression." + "reason": "Different models and symptoms: SigLIP2 bad outputs vs CamemBERT masked-LM prediction regression." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Llama4Config missing pad_token_id is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Missing pad_token_id in Llama4Config is unrelated to Qwen3-MoE sliding_window behavior." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "T5 apex RMSNorm memory leak and Qwen3_5Moe CPU-offload incompatibility are unrelated." + "reason": "T5 apex memory leak and Qwen3_5Moe CPU offload incompatibility are different problems." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both mention fix_mistral_regex/tokenizer regressions, but one is a crash on Mistral tokenizer and the other is a Kimi-K2.5 codec-handling regression with a misleading warning." + "reason": "Both concern tokenizer patching/regression, but one is a crash in _patch_mistral_regex and the other is Kimi-K2.5 codec handling." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device-map cudaErrorAssert is unrelated to the T5 RMSNorm memory leak." + "reason": "Qwen3 auto device mapping cuda asserts and T5 apex RMSNorm memory leak are unrelated." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Zero3/BertModel index error is unrelated to the fbank feature extraction issue." + "reason": "Deepspeed ZeRO-3/BertModel indexing error is unrelated to the fbank feature extraction issue." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 doc/model mismatch is unrelated to quantization offloading behavior." + "reason": "SigLIP2 docs/quantization issues are not the same as quantized offloading not working as expected." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT local-checkpoint loading and quantization offloading are different code paths and symptoms." + "reason": "PEFT local checkpoint loading for GraniteSpeech is a different bug from quantization/offloading behavior." }, { "left": "issue:42915", "right": "issue:45759", "accept": false, - "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to config unwrapping in AutoModelForCausalLM.from_config." + "reason": "Both are Qwen config-related, but one is FP8Config failure and the other is from_config handling of composite multimodal configs." }, { "left": "issue:40444", "right": "issue:43746", "accept": false, - "reason": "Qwen2.5-VL iterable dataset failure has nothing to do with GraniteSpeech PEFT checkpoint loading." + "reason": "Multiple-image Qwen2.5-VL finetuning failure and GraniteSpeech adapter loading are unrelated." }, { "left": "issue:43746", "right": "issue:45381", "accept": false, - "reason": "Different domains: PEFT adapter loading vs Qwen2.5-VL video position ids." + "reason": "Different models and paths: PEFT adapter loading vs Qwen2.5-VL video position IDs." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio fbank extraction and add-new-model-like tokenizer mapping logic are unrelated." + "reason": "Audio fbank feature extraction is unrelated to add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Qwen3OmniMoe video processing error and Gemma 3n variable-name collision are distinct bugs." + "reason": "Qwen3OmniMoe video-input processing and Gemma3n forward variable collision are different bugs." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Image processor void segmentation map failure is unrelated to the GenerationMixin ImportError." + "reason": "Void segmentation map processing and GenerationMixin import failure are completely unrelated." } ] }, @@ -4753,202 +4723,202 @@ "issue:44561|issue:45020", "issue:39692|issue:45834" ], - "prepared_review_unit_hash": "150ee0d54e730d6e8afca6da7a73033732d779c445ec61c2aa100301ae2e27ab", + "prepared_review_unit_hash": "88def731470adc5aff4713ca9bd0170e64b8d24b6f48bd0cb74ec0b8505e4673", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36340, - "estimated_input_tokens": 9085, - "estimated_eval_tokens": 18426 + "serialized_chars": 36342, + "estimated_input_tokens": 9086, + "estimated_eval_tokens": 18428 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of distinct Transformers regressions and docs issues, not true duplicates. The soft-similarity links mostly reflect shared subsystems (tokenizers, multimodal models, v5 regressions) rather than the same underlying bug.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue fits the cluster well: the items span unrelated failures across tokenizer, vision, dtype, remote_code, and docs paths, so the soft pairs should not be merged into one duplicate set.", - "canonical_pr_reason": "No PRs are present in this cluster.", - "best_issue_reason": "issue:44843 is the best single representative only in the sense that it is a concrete, well-scoped regression with a clear code path (_patch_mistral_regex); however, it still does not represent the broader cluster.", - "best_pr_reason": "No PRs are present in this cluster.", + "summary": "This cluster is broad and noisy: most items are distinct Transformers regressions across unrelated models, tokenizers, multimodal loaders, and docs. I do not see strong duplicate evidence for most soft edges; the closest ones share a subsystem name or helper function, not the same concrete bug.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45020 is the broadest umbrella here and best matches the recurring theme of recent-version regressions breaking model loading/usage, especially around remote-code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most representative issue for the cluster\u2019s general regression theme; it is broader and more central than the highly specific model- or tokenizer-only reports.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different models and failure modes: BitNet accelerate-loading unpacking vs DAC STE mismatch." + "reason": "Different models and failures: packed-weight loading vs DAC latent reconstruction/STE mismatch." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Both are v5 regressions, but one is CLIP attentions metadata and the other is RTDetrV2 output drift." + "reason": "Both are v5 behavior regressions, but they affect different models and code paths." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Unrelated bugs: SwitchTransformers sparse-layer construction vs Smollm3 RoPE-layer dropping." + "reason": "Unrelated architecture bugs: sparse-layer config creation vs RoPE layer dropping." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Validation/error-message issue vs classification-loss degeneration; different code paths and symptoms." + "reason": "Different validation/loss bugs in separate code paths; not the same underlying issue." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Tokenizer special-token registration bug is unrelated to BitNet packed-weight loading." + "reason": "Tokenizer special-token registration bug vs packed-weight loading bug; unrelated." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Different multimodal components and errors: image dtype mismatch vs Gemma3n variable-name collision AttributeError." + "reason": "Different multimodal model failures with different root causes." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Both involve processors/tokenizers, but one is a processor-tokenizer API regression and the other is a Kosmos2.5 OCR index error." + "reason": "Both are multimodal input issues, but one is processor/tokenizer regression and the other is OCR index handling." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken special-argument conflict is not the same as failing to load a remote tokenizer/model." + "reason": "Different tokenizer problems: AddedToken construction vs loading a specific model/tokenizer." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Different subsystems: torch.compile/SDPA failure in Bamba vs bf16 dtype mismatch in unrelated models." + "reason": "Different bugs in different models: torch.compile/SDPA vs dtype mismatch under bfloat16." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Completely different models and code paths; no shared underlying bug." + "reason": "No shared underlying code path; compile/SDPA failure is unrelated to DAC latents." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next loading segfault is unrelated to the MPS histogram/int dtype error in moe." + "reason": "Both are runtime errors, but in unrelated models and hardware backends." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer crash on NER/padding is unrelated to Perceiver image-classification resolution handling." + "reason": "Tokenizer/padding crash vs image classification resolution/interpolation failure; unrelated." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both touch _patch_mistral_regex, but one is an offline hub call bug and the other is a backend_tokenizer AttributeError; not the same bug." + "reason": "Same helper name, but different failures: offline hub call vs missing backend_tokenizer attribute." }, { "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 MoE bug is unrelated to remote-code config field loss." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype loading bug vs Video-LLaVA temporal-attention/weight-sharing issue." + "reason": "BLIP2 dtype propagation bug vs Video-LLaVA weight/temporal-attention issue." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Missing returned config fields under trust_remote_code is different from strict config blocking granite_speech loading." + "reason": "Both involve loading configs/models, but the concrete failures differ." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Special-token post-processor refresh bug is not the same as Whisper batch_decode skip_special_tokens handling." + "reason": "Post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; not the same issue." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration bug vs DAC latent/forward mismatch are unrelated." + "reason": "Tokenizer special-token registration and DAC latent decoding are unrelated." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both are output regressions, but they affect different models and different failure mechanisms." + "reason": "Both are v5 output regressions, but on different model families and causes." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Different multimodal video bugs in different Qwen model families and different tensor bookkeeping paths." + "reason": "Both mention video/Qwen, but they target different model variants and different input bugs." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Weight-only loading error is unrelated to flash-attention generation failure." + "reason": "Different Qwen issues: from_pretrained weight-only loading vs flash-attention generation behavior." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation bug is unrelated to Kosmos2.5 long OCR index errors." + "reason": "Whisper empty-transcription generation bug is unrelated to Kosmos2.5 long-OCR indexing." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer initialization regression is unrelated to FSDP CPU RAM efficient loading." + "reason": "Non-persistent buffer initialization bug vs FSDP CPU RAM-efficient loading regression." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "RoPE implementation discussion is not the same as stale docs mentioning removed pipeline tasks." + "reason": "RoPE implementation discussion vs outdated docs; no shared bug." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor-from-URL loading and non-template-node compilation errors are unrelated." + "reason": "ImageProcessor URL loading and template-node compilation are unrelated." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline task error-message cleanup is unrelated to SigLIP2 model/processor mismatch and quantization failure." + "reason": "Error-message/docs issue vs SigLIP2 wrong results from AutoModel/pipeline mismatch; not the same bug." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both mention remote_code breakage, but one is a specific removed helper and the other is a broad report about recent versions." + "reason": "Related broad area (remote-code regressions), but these are distinct breakages with different triggers." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 docs/example and Kosmos2.5 OCR failure are unrelated model-specific issues." + "reason": "Different model-specific multimodal/documentation problems with no shared failure mode." } ] }, @@ -5049,20 +5019,20 @@ "issue:43329|issue:45600", "issue:43742|issue:45020" ], - "prepared_review_unit_hash": "df0026f88181c04e0966d70d033f0e58121dcd22d4c3836ae2279846bb40ac38", + "prepared_review_unit_hash": "12bef7c90d365f5e98ee5cfd51b697f82a88dbe9d609ba49bdbcb2a5137aa5c5", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35509, + "serialized_chars": 35510, "estimated_input_tokens": 8878, "estimated_eval_tokens": 18012 }, @@ -5071,174 +5041,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair looks like a true duplicate: the items span unrelated bugs across processors, tokenizers, loading, training, and device/dtype behavior. There are no PRs in this cluster.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue is appropriate here; the set is too heterogeneous, and none of the soft pairs describe the same underlying defect closely enough to merge.", + "summary": "This cluster is mostly heterogeneous: the soft pairs share broad themes like model loading, tokenizers, or v5 regressions, but not the same concrete bug or change. I would not merge any of the listed soft edges.", + "confidence": 0.89, + "canonical_issue_reason": "No single issue is a true canonical duplicate target for this mixed set; if one must be chosen, issue 45020 is the broadest umbrella regression report, but it is not actually the same bug as most others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the closest thing to an umbrella anchor because it is broad and user-facing, but it still should not be treated as a duplicate target for the others.", + "best_issue_reason": "Issue 45020 is the most general, user-facing open regression and the closest thing to an umbrella report in this set, but it should still not absorb the unrelated issues as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Different subsystems: image processor segmentation maps vs trainer loss signature handling." + "reason": "Different subsystems and failures: image segmentation processing vs trainer compute_loss signature mismatch." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "CUDA auto-device-map crash and tokenizer loading failure are unrelated code paths." + "reason": "Unrelated problems: CUDA assert during device mapping vs tokenizer load failure for a specific model." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Related theme, but one is a runtime error message and the other is documentation drift after pipeline removals." + "reason": "Both concern v5 pipeline removal, but one is an error message bug and the other is stale docs; not the same underlying defect." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Different bugs: Qwen VL video regression vs Kosmos2.5 long-OCR index error." + "reason": "Different models and code paths: Qwen video regression vs Kosmos2.5 OCR long-input index error." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Both involve tokenizers, but the concrete failures and affected components differ." + "reason": "Different tokenizer bugs affecting different models and inputs." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "Blip2 dtype loading bug vs Apple Silicon allocator TypeError; not the same defect." + "reason": "Model dtype loading bug vs Apple Silicon allocator TypeError; unrelated." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask-shape IndexError is unrelated to evaluation failing after training." + "reason": "Mask shape IndexError in one workflow vs evaluate-after-train failure; not the same bug." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "Masked LM prediction drift is not the same as generic tokenization changes across versions." + "reason": "CamemBERT prediction regression is not the same as a broad tokenization-change report." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Both mention mismatches, but the model/code paths and failure modes are different." + "reason": "Different inference issues: CPU offload device mismatch vs bf16 dtype mismatch." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice preset handling and GenerationMixin import failure are unrelated." + "reason": "Bark voice preset behavior and GenerationMixin import error are unrelated." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor output size mismatch is a different bug from bfloat16 dtype mismatch." + "reason": "Processor image-size mismatch vs dtype mismatch; different code paths." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Model loading weight_only error vs RoPE position-id scaling bug; different mechanics." + "reason": "Both are VL-related, but one is weight_only loading and the other is rope/position-id scaling; distinct bugs." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Packaging/backend dependency omission is unrelated to the tokenizer regex patch crash." + "reason": "Packaging/backend dependency issue vs tokenizer regex patch crash; unrelated." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight loading during accelerate init is a different failure than dtype mismatches." + "reason": "Packed-weight loading failure is a different issue from generic dtype mismatches." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Both are tokenizer/processor errors, but they fail in different places and for different reasons." + "reason": "Similar symptom class (missing attribute), but different processors and different failure sites." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP RAM-efficient loading and init_empty_weights argument compatibility are distinct loading bugs." + "reason": "Both involve loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights arg compatibility." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Incorrect tied-weights mapping is not the same as a warning about tie_word_embeddings configuration." + "reason": "Different warnings and model families; not the same tied-weights problem." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon warmup TypeError and CUDA AMP loss crash are unrelated platform-specific failures." + "reason": "Apple Silicon warmup TypeError vs CUDA AMP loss crash; unrelated." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "One is a low-signal request, the other is an isolated reproduction of a separate upstream issue." + "reason": "One is a meaningless report and the other is a reproduction of a specific upstream issue; no duplicate relation." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio sampling-rate defaulting and DAC latent/forward mismatch are different code paths." + "reason": "Audio sampling-rate default bug and DAC latent consistency bug are unrelated." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Label reduction in BeitImageProcessorFast is not the same as RTDetrV2 output divergence." + "reason": "Beit reduce_labels bug and RTDetrV2 output regression are different model-specific defects." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "A label-shape bug is unrelated to the MPS attention correctness issue." + "reason": "Completely different topics: label reduction vs MPS attention correctness." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation failure and tokenizer max_length sentinel are unrelated." + "reason": "Template compilation error vs tokenizer max_length bug; unrelated." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Both concern error handling, but they are different anti-patterns in different functions." + "reason": "Both involve exception handling, but they are different components and different swallowed-error paths." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "Bark voice preset behavior and CustomTrainer.compute_loss signature mismatch are unrelated." + "reason": "BarkProcessor preset issue and trainer compute_loss signature issue are unrelated." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Multimodal token counting bug vs stale auto_mappings config reference; different failures." + "reason": "Multimodal token counting bug vs stale auto_mappings references; different files and problems." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "MobileLLM key error is not the same as the broader remote_code regression report." + "reason": "A key error for one model is not the same concrete bug as a broad remote_code compatibility regression." } ] }, @@ -5328,9 +5298,9 @@ "issue:43550|issue:44912", "issue:44568|issue:45488", "issue:43531|issue:45042", - "issue:43329|issue:44479", "issue:43262|issue:44811", "issue:44568|issue:45356", + "issue:43329|issue:44479", "issue:36010|issue:36331", "issue:43749|issue:44756", "issue:43643|issue:45759", @@ -5342,220 +5312,220 @@ "issue:43257|issue:43756", "issue:45823|issue:45859" ], - "prepared_review_unit_hash": "30c63c487a2d726d76c3fdfa70f9db8dba6a62318f1c915e1e20e57561622afa", + "prepared_review_unit_hash": "cd1b29853f7b678ac1af1421b6785f381e8db2826db35f41db53272d0af337a3", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36388, - "estimated_input_tokens": 9097, - "estimated_eval_tokens": 18450 + "serialized_chars": 36389, + "estimated_input_tokens": 9098, + "estimated_eval_tokens": 18452 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: most pairs share broad subsystem keywords but not the same concrete bug. I\u2019d reject nearly all soft edges; only the video-regression pair looks plausibly duplicate-adjacent.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue is evident: the items span unrelated training, loading, tokenizer, multimodal, and device/offload bugs.", + "summary": "These items are mostly unrelated closed issues spanning tokenizer regressions, loading/offload bugs, multimodal processor problems, and trainer behavior changes. The only plausible duplicate-style match is the Qwen VLM video-input regression pair.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45701 is the broadest open, user-facing umbrella here: it directly describes version-dependent tokenization changes and has inbound references, making it the best representative for the tokenizer-regression slice of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the broadest issue here and a reasonable anchor for the tokenization-regression subset, but it is not a true duplicate target for the cluster as a whole.", + "best_issue_reason": "issue:45701 is the best single issue to keep as a representative anchor because it is open, broadly worded, and closest to the recurring tokenizer-regression theme seen among several items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Different failures: BigBirdTokenizer special-token registration vs gpt-oss MXFP4 quantization fallback." + "reason": "Special-token decode failure vs MXFP4 quantization/loading fallback; different subsystems and symptoms." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Both are training-related, but one is a zero3/Bert index error and the other is evaluate-after-train state handling." + "reason": "Zero3/Bert indexing failure vs post-train evaluate failure; not the same code path." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both concern tokenizer loading, but they point to different models and different breakages in the loading path." + "reason": "Generic gguf/from_pretrained breakage vs a specific tokenizer regression for EMBEDDIA/est-roberta." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Weight conversion with accelerate/deepspeed is unrelated to a video token-counting helper bug." + "reason": "Qwen3 MoE weight conversion/loading issue vs multimodal video token counting bug; different defects." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup and PIL/torchvision dependency handling are unrelated." + "reason": "Float8 storage lookup error vs image processor torchvision dependency regression." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT local checkpoint loading and Whisper batch_decode skip_special_tokens are different code paths." + "reason": "PEFT checkpoint loading bug vs Whisper batch_decode skip_special_tokens bug." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Both are model-specific generation/loading regressions, but not the same model or concrete defect." + "reason": "GLM-4.7-Flash loading test failure vs Qwen3.5 flash-attention generation bug." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MoE weight conversion and GPT-oss OOM are different issues." + "reason": "Loading/conversion bug vs out-of-memory report; unrelated failure modes." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Unexpected zero probabilities is not the same bug as processor output size mismatch." + "reason": "SigLIP2 probability issue vs OmDet processor size mismatch; different model behavior." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "GenerationMixin import failure and Jupyter custom-model initialization breakage are different regression types." + "reason": "GenerationMixin import error vs Jupyter custom-model initialization regression." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation/example errors and quantization fallback are unrelated." + "reason": "Docs example errors for SigLIP2 vs GPT-oss quantization loading; no shared bug." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent decoding mismatch and DiaConfig save/load token IDs are unrelated model bugs." + "reason": "DAC forward/from_latents mismatch vs config save/load custom token ID loss." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "lm_head tying in AyaVision and BLIP2 dtype loading are separate model-loading problems." + "reason": "Weight-tying/text generation bug vs dtype propagation bug during Blip2 load." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight unpacking during accelerate load is unrelated to a processor shape mismatch." + "reason": "Accelerate packed-weight unpacking bug vs processor image-shape mismatch." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile SDPA failure and MXFP4 quantization fallback are different issues." + "reason": "torch.compile/SDPA failure vs quantization fallback; different concrete problems." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but one is add_special_tokens/BOS-EOS handling and the other is pre-tokenizer override behavior." + "reason": "Both tokenizer-related, but one is add_special_tokens/BOS-EOS handling and the other is pre-tokenizer override; not the same bug." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to PIL backend image processor dependencies." - }, - { - "left": "issue:43329", - "right": "issue:44479", - "accept": true, - "reason": "Both center on the video multimodal path; 43329 identifies the concrete undefined-helper bug that plausibly underlies the broader video input regression in 44479." + "reason": "Sliding-window MoE bug vs PIL backend torchvision dependency issue." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate defaults and Whisper decode flags are different processor behaviors." + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode special-token skipping." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but the specific breakage and model families differ." + "reason": "Different tokenizer regressions with different root causes and model families." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both point to the same video-input multimodal path; 43329 names the broken video token-counting branch, and 44479 reports a broad video regression on the same VLM families." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Importing GenerationMixin and compute_loss signature changes are separate API breakages." + "reason": "Import-time GenerationMixin breakage vs compute_loss signature change; unrelated API regressions." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM efficient loading and mmap OOM on Strix Halo are different memory/loading issues." + "reason": "FSDP CPU-RAM-efficient loading bug vs mmap/OOM issue; different loading paths and causes." }, { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "Missing fields with trust_remote_code and composite-config unwrapping are different config-loading bugs." + "reason": "Missing fields from trust_remote_code AutoConfig vs AutoModelForCausalLM.from_config text_config unwrap issue." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Hidden-states output and tokenization changes are unrelated." + "reason": "Qwen3.5 hidden-states bug vs broad tokenization-change report; not the same underlying defect." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE weight conversion and Qwen3-VL shape mismatch are different loading failures." + "reason": "Qwen3 MoE loading under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Token counting in the video branch and batched video processor output shape are not the same defect." + "reason": "Video token counting undefined-symbol bug vs batched video processor shape bug." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token-count tracking and GPT-oss perplexity are unrelated." + "reason": "Trainer token-count tracking bug vs GPT-oss perplexity being too high; different layers of the stack." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Both involve CPU offload, but the affected models and failure modes differ." + "reason": "Both involve CPU offload/device placement, but they affect different models and code paths." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer behavior are unrelated." + "reason": "MoE loading/conversion bug vs Smollm3 RoPE-layer discrepancy; unrelated model behavior." }, { "left": "issue:45823", "right": "issue:45859", "accept": false, - "reason": "Gemma4 device-map mismatch and Qwen3_5MoE rotary embedding CPU-offload incompatibility are different model-specific device bugs." + "reason": "Gemma4 device mismatch during forward vs Qwen3_5MoE rotary embedding CPU-offload incompatibility." } ] }, "evaluator_result": { "accept": true, - "feedback": "The evaluation is grounded in the packet and stays conservative overall. The cluster-level conclusion that this is mostly heterogeneous is supported, and the one accepted soft edge is plausibly the same video-regression family without overstating certainty. The rejected pairs also read as appropriately cautious rather than forced duplicates." + "feedback": "The summary is broadly grounded in the packet, and the soft-edge decisions are conservative overall. The lone accepted pair (issue:43329 / issue:44479) is reasonably mergeable as a single video-input regression fix. No clear overstatement of evidence stands out." }, "evaluator_used": true, "retried": false, @@ -5660,238 +5630,238 @@ "issue:43493|issue:43994", "issue:44291|issue:44821" ], - "prepared_review_unit_hash": "f7932867ad6e0f78a5c4dba8c11c767963f0b3b619633f339ccfd149d043712c", + "prepared_review_unit_hash": "5dc0fadce238ac3a62d405c3f7c7c1eb6a5b261e13a5d3d08c8b2c8debd0aabe", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37715, - "estimated_input_tokens": 9429, - "estimated_eval_tokens": 19114 + "serialized_chars": 37719, + "estimated_input_tokens": 9430, + "estimated_eval_tokens": 19116 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of Transformers bug reports. The only strong duplicate-looking subsets are the Qwen VL video-input regression pair and the SigLIP2 output-discrepancy pair; most other items are related only at the subsystem level and should stay separate.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 44479 is the broadest and most representative concrete bug report here: a versioned video-input regression affecting multiple Qwen VL variants, with a clear shared code path and duplicate potential.", + "summary": "The set is mostly heterogeneous, with a few small duplicate-like mini-clusters around Qwen2.5-VL video/position-id regressions, tokenizer loading, and SigLIP2 behavior.", + "confidence": 0.71, + "canonical_issue_reason": "issue:44479 is the closest thing to a canonical anchor because it matches the Qwen2.5-VL video/position-id regression cluster; the rest of the items are too diverse to form one true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "44479 is the best anchor issue because it captures the widest plausible duplicate set around the Qwen VL video-input regression and is more general than the narrower follow-up report.", + "best_issue_reason": "issue:44479 is the strongest local hub among the issues, but it is still only suitable as a partial anchor rather than a global representative of the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different models and different problems: RTDetrV2 output drift vs Qwen3.5 save_pretrained key corruption." + "reason": "Different bugs: RTDetrV2 output drift vs Qwen3.5 save_pretrained key serialization." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Both involve loading/tokenizer behavior, but one is a model load failure and the other is a tokenizer max_length bug for CLIPTokenizer." + "reason": "Different failure modes: tokenizer load regression vs CLIPTokenizer max_length misuse." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Unrelated bugs: init_empty_weights unexpected argument vs PIL backend requiring torchvision." + "reason": "One is init_empty_weights argument handling; the other is PIL backend unexpectedly requiring torchvision." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "Different subsystems and symptoms: gguf/from_pretrained loading vs Apple Silicon caching_allocator_warmup TypeError." + "reason": "Unrelated issues: gguf_file loading breakage vs Apple Silicon allocator warmup TypeError." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Both are model-loading failures, but they affect different models and likely different code paths." + "reason": "Different model-specific load bugs: Qwen3Moe FP8 config vs MobileLLM key error." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Different issue classes: Qwen3Moe FP8 failure vs AutoImageProcessor URL loading." + "reason": "Different bugs and code paths; no shared concrete failure is evident." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "No shared underlying bug: MacOS bus error on CLIP vs TrainerState token-count tracking." + "reason": "MacOS bus error on CLIP model is unrelated to TrainerState token counting." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Different failure modes and components; only very broad loading-related similarity." + "reason": "Qwen3Moe FP8 config issue is unrelated to CLIPTokenizer model_max_length." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Different model families and errors: device-map cuda assert vs video-input ValueError." + "reason": "Both involve Qwen models, but the concrete failures differ: auto device mapping assert vs video-input ValueError." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Strong duplicate signal: both report the same Qwen VL video-input / vision position-id regression in v5.3.0." + "reason": "Both describe the same Qwen2.5-VL video/position-id regression in 5.3.0; one is a broader report and the other a narrower repro." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Device-map/offload loading bug vs CLIPTokenizer max_length issue; not the same change or code path." + "reason": "Device-map offload failure is unrelated to CLIPTokenizer max_length." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Different model/config bugs: DiaConfig token IDs persistence vs MXFP4 quantization fallback." + "reason": "Different topics: DiaConfig save/load token IDs vs MXFP4 quantization fallback." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Unrelated backend/install packaging issue vs missing Qwen2_5_VL import." + "reason": "Packaging/backend availability issue is unrelated to missing Qwen2_5_VL import." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both mention dtype mismatches, but they concern different models and different inference paths." + "reason": "Both are dtype-related, but they affect different model classes and code paths." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are video-related, but the models and concrete failures differ (Qwen3OmniMoe ValueError vs Qwen3-vl-embedding StopIteration)." + "reason": "Both are video-input bugs, but they are different model families and different exceptions." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Related theme, but different model families and distinct video-processing failures." + "reason": "Different Qwen model families and different concrete failures; not the same bug." }, { "left": "issue:44062", "right": "issue:44991", - "accept": false, - "reason": "Tokenizer errors for different root causes; not enough evidence of one underlying bug." + "accept": true, + "reason": "Likely the same tokenizer bug: the AddedToken/special-argument TypeError appears to be the concrete cause of the EMBEDDIA/est-roberta tokenizer load failure." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Packaging/backend inclusion issue vs Python 3.14 load failure; no shared code-path evidence." + "reason": "Backend packaging issue vs Python 3.14 load failure are unrelated." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error is unrelated to a PEFT version pin issue." + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT minimum-version bump." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Different problems: backend installation contents vs model loading failure for a specific checkpoint." + "reason": "Missing tensorflow backend in installation is unrelated to a model-loading failure." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Both are vision-model bugs, but they are different models and different symptoms, so not the same underlying change." + "reason": "Different models and symptoms: SigLIP2 zero probabilities vs Perceiver non-default-resolution failure." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU placement bug vs flash-attention sinkless-model crash; unrelated." + "reason": "XLNet CPU arange bug is unrelated to flash-attention s_aux=None handling." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "ImportError due to missing PyTorch vs a model-loading regression for Llama 3.2 Vision." + "reason": "Generic PyTorch import error vs Llama 3.2 Vision model-loading regression are different issues." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "Different loading failures with different triggers; no shared concrete code path." + "reason": "MobileLLM key error and init_empty_weights unexpected-arg failure are separate loading problems." }, { "left": "issue:43906", "right": "issue:45381", "accept": false, - "reason": "One is a reproduction of an older issue, the other is a Qwen2.5-VL video-position bug; not enough to merge." + "reason": "A reproduction of another issue is not the same as the Qwen2.5-VL video-position-id bug." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Different training/loading paths: Qwen3Moe FP8 config vs FSDP CPU RAM efficient loading." + "reason": "Qwen3Moe FP8 config failure is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both involve RoPE/position logic, but they affect different models and distinct bugs." + "reason": "Smollm3 RoPE-layer behavior and Qwen2.5-VL temporal position-id scaling are different model-specific bugs." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "ImageProcessor label reduction bug vs sequence-classification zero-loss bug; unrelated." + "reason": "BEiT fast processor label reduction bug is unrelated to zero-loss classification behavior." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "ImportError for missing PyTorch vs LayoutLMv3 error-message clarity; not duplicates." + "reason": "PyTorch import availability and LayoutLMv3 box-clarity messaging are unrelated." }, { "left": "issue:43540", "right": "issue:45823", "accept": false, - "reason": "Different models and failure types: Qwen3OmniMoe video processing vs Gemma4 device mismatch." + "reason": "Qwen3OmniMoe video input failure is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "AppleSilicon TypeError in warmup vs Qwen3VL weight_only load error; unrelated." + "reason": "Apple Silicon allocator warmup TypeError and Qwen3VL weight_only load error are different bugs." }, { "left": "issue:43493", "right": "issue:43994", "accept": true, - "reason": "Strong model-level overlap: both report SigLIP2 producing wrong/nonsensical outputs, likely the same implementation bug." + "reason": "Same underlying SigLIP2 implementation problem: one compares HF to JAX, the other reports the same model giving nonsensical outputs in HF usage." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Unexpected init_empty_weights argument vs URL loading for AutoImageProcessor; different code paths." + "reason": "init_empty_weights argument handling and AutoImageProcessor URL loading are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The two accepted duplicate pairs are supported by the issue titles, and the rest of the pairwise rejections avoid overclaiming shared causality. 44479 is a reasonable anchor for the Qwen VL video regression cluster, and the overall heterogeneous-cluster summary matches the evidence." + "feedback": "The assessment stays grounded in the packet and the soft-edge calls are mostly conservative. The accepted duplicate-like pairs are reasonably supported by the titles, and the rejected pairs do not overclaim similarity. The summary also appropriately treats 44479 as a local anchor rather than a global canonical issue." }, "evaluator_used": true, "retried": false, @@ -5990,20 +5960,20 @@ "issue:44936|issue:45823", "issue:44464|issue:44743" ], - "prepared_review_unit_hash": "ea338bf2d828722ab7257ebaff837c67aba2b038d0590515adeca7e3e4c36988", + "prepared_review_unit_hash": "0ae36d680b47f599b6150580f72028d554fda88aa5602b06bebc624ebbdacdc0", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36117, + "serialized_chars": 36118, "estimated_input_tokens": 9030, "estimated_eval_tokens": 18316 }, @@ -6012,180 +5982,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are clearly different bugs despite superficial overlap in Transformers subsystems. The only pair that looks like the same underlying regression is the v5 tied-weights/save_pretrained issue family.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45005 is the better anchor because it describes the broader tied-weights regression in v5, which encompasses the narrower lm_head.weight serialization symptom in issue:44466.", + "summary": "This cluster is mostly a set of unrelated transformer bugs. The only soft edge that looks like a true duplicate is the pair about v5 tied-weights/save_pretrained serialization behavior.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44466 is the better canonical issue: it states the underlying tied-weights serialization regression more generally and describes the device-dependent behavior directly; issue:45005 reads like a model-specific manifestation of the same v5 tied-weights problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the most representative user-facing report in the cluster: it captures the underlying tied-weights problem without being limited to one device-specific manifestation.", + "best_issue_reason": "issue:44466 is the best representative of the cluster's only plausible duplicate thread because it captures the core regression rather than a downstream translation-model symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "Both are trainer-related, but one is a token-count tracking bug and the other is a custom compute_loss signature break; different code paths." + "reason": "Different trainer problems: token counting state vs a custom compute_loss signature mismatch." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Tokenizer construction error vs. Mistral regex patch crash; both involve tokenizers but the failure modes and causes differ." + "reason": "Both tokenizer-related, but one is an AddedToken kwarg error and the other is a Mistral regex/backend_tokenizer crash." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Both mention config fields, but one is a trust_remote_code AutoConfig issue and the other is a strict schema field drop in a specific config class; not the same bug." + "reason": "Both involve config fields, but the first is missing fields from trust_remote_code AutoConfig output and the second is a specific Qwen3.5MoeVisionConfig field dropped by strict parsing." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both are model loading failures, but the affected APIs/models and the reported failures differ; no clear shared underlying defect." + "reason": "Both are loading failures, but they involve different APIs and likely different root causes." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Different multimodal models and different failures (test generation vs load failure); too broad to call duplicates." + "reason": "Both are multimodal model loading/test failures, but they concern different models and symptoms." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Image processor torchvision dependency regression vs tokenizer regex crash; unrelated code paths." + "reason": "Unrelated subsystems: image processor torchvision dependency vs tokenizer regex crash." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "KeyError while loading one model versus missing Float8 storage type during serialization/loading; distinct loader issues." + "reason": "Different load-time failures with different low-level causes; one is a model key error, the other is a missing float8 storage object." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior vs AutoImageProcessor URL loading; no meaningful bug overlap." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to AutoImageProcessor URL loading." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "One is an unreleased version constraint on PEFT, the other is a stale config reference breaking CI; different problems." + "reason": "PEFT version pinning and auto_mappings referencing removed configs are separate CI/release issues." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Aux-loss normalization in MoE models vs FSDP CPU RAM efficient loading; different functionality and root cause." + "reason": "Aux-loss normalization and FSDP CPU RAM efficient loading are different bugs in different code paths." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "Different regressions: RTDetrV2 output changes vs Qwen3.5 save_pretrained checkpoint corruption." + "reason": "One is RTDetrV2 output drift in v5; the other is Qwen3.5 checkpoint saving. Same general version area, different bug." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "save_pretrained shard size question vs AutoProcessor kwargs not forwarded to cached_file; unrelated." + "reason": "Save shard size question is unrelated to AutoProcessor kwargs forwarding." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL issues, but one is apply_chat_template batching/padding and the other is video vision_position_ids; different symptoms and likely different fixes." + "reason": "Both are Qwen2.5-VL issues, but one is chat-template batching crash and the other is incorrect vision_position_ids for video input." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight loading bug vs Perceiver interpolation failure; not the same code path." + "reason": "BitNet packed-weight loading and Perceiver image-classification resolution handling are unrelated." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both touch trust_remote_code, but one is missing fields in AutoConfig output and the other is a v5 utility removal breaking remote code; different regressions." + "reason": "Different regressions: missing config fields under trust_remote_code vs removal of is_torch_fx_available breaking remote-code models." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed ZeRO-3 model loading error vs trainer.evaluate() after train(); no duplicate evidence." + "reason": "DeepSpeed ZeRO-3 loading crash and trainer.evaluate-after-train failure do not share a concrete code path." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "Tokenizer/model loading breakage vs BLIP2 dtype propagation; unrelated issues." + "reason": "GGUF loading regression and BLIP2 dtype propagation are distinct model-loading issues." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Both are config-related, but one is num_labels propagation and the other is strict loading blocking granite_speech; different defects." + "reason": "Both involve config handling, but one is num_labels propagation and the other is strict config blocking granite_speech loading." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed training RuntimeError vs LayoutLMv3 error message clarity; not duplicates." + "reason": "Training DeepSpeed weight shape error is unrelated to a LayoutLMv3 box-validation message issue." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Generic loading breakage vs offline model_info call in _patch_mistral_regex; different failure mechanisms." + "reason": "GGUF loading breakage and mistral_regex offline/model_info behavior are different tokenizer/loading problems." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer/processor regressions, but one is processor.tokenizer/image passing and the other is Kimi-K2.5 codec handling and warning text; not the same bug." + "reason": "Both are regressions around processor/tokenizer behavior, but one is processor.tokenizer/image handling and the other is Kimi-K2.5 codec/fix_mistral_regex warnings." }, { "left": "issue:44466", "right": "issue:45005", "accept": true, - "reason": "Both describe the same tied-weights/save_pretrained regression in v5: one gives the specific lm_head.weight/device symptom, the other frames the broader translation-model tied-weights issue." + "reason": "Both point to the same v5 tied-weights serialization regression: inconsistent lm_head.weight handling for tied-weight models, with translation models as a specific affected case." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC forward/from_latents mismatch vs Whisper batch_decode skip_special_tokens; unrelated." + "reason": "DAC forward/from_latents mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Same API family, but batched padding=False crash and assistant tool-call/no-content crash are distinct edge cases." + "reason": "Both affect apply_chat_template in Qwen2.5-VL, but the failure modes differ: batched padding=False vs assistant tool-call messages without content." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both involve special tokens, but one is BigBird mask token registration and the other is BOS/EOS insertion for mdeberta; different model/tokenizer behavior." + "reason": "Different tokenizer special-token regressions for different models and different APIs." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "trainer.evaluate() lifecycle issue vs Gemma4 PLE device mismatch under device_map='auto'; unrelated." + "reason": "Trainer evaluation failure and Gemma4 PLE device mismatch are not the same bug." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Both concern generation correctness, but one is compiled forward inconsistency and the other is recurrent-state reset in a specific modular model; not the same bug." + "reason": "Both concern generation/runtime correctness, but one is chunked generation with compiled forward and the other is recurrent state reset in a specific Qwen modular file." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative on the soft-edge pairs. The only accepted duplicate, issue:44466 vs issue:45005, is well supported by the titles and descriptions as the same tied-weights/save_pretrained regression with one narrower device-specific symptom. The summary does not overstate weaker matches." + "feedback": "The summary is grounded in the packet, and the single accepted soft edge (44466/45005) is a conservative, well-supported duplicate call based on the issue titles. The other pair verdicts are appropriately cautious given the limited evidence." }, "evaluator_used": true, "retried": false, @@ -6272,14 +6242,14 @@ "issue:44977|issue:45356", "issue:43531|issue:43742", "issue:43653|issue:43927", - "issue:43329|issue:45325", "issue:44464|issue:44977", + "issue:43329|issue:45325", "issue:43577|issue:45588", "issue:43550|issue:43927", "issue:43577|issue:44554", "issue:44479|issue:44625", - "issue:43550|issue:44361", "issue:43122|issue:43906", + "issue:43550|issue:44361", "issue:32090|issue:34567", "issue:41720|issue:44464", "issue:43504|issue:43550", @@ -6291,244 +6261,244 @@ "issue:44849|issue:45357", "issue:43278|issue:43688" ], - "prepared_review_unit_hash": "24329f43457dd66abe5fa80e012514dd5aad9f20f0b392a2db73b2ef9f4da4d7", + "prepared_review_unit_hash": "630e90614d011610badb22f0bc629ce280caeb875a8d45dd4c5a844bdbdfe2ce", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 35, - "serialized_chars": 37683, - "estimated_input_tokens": 9421, - "estimated_eval_tokens": 19098 + "serialized_chars": 37686, + "estimated_input_tokens": 9422, + "estimated_eval_tokens": 19100 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes tokenizer regressions, model-loading failures, training/runtime bugs, multimodal/video issues, and unrelated backend-specific problems. None of the soft edges look like the same underlying bug or change, so they should all be kept separate.", - "confidence": 0.97, - "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated failure modes and code paths, so there is no shared underlying bug to anchor the cluster.", + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly connect unrelated bugs/regressions across tokenization, multimodal processing, training, loading, and dtype/device-specific behavior. I would keep all of the proposed soft edges separated.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items are only superficially similar and do not share one underlying bug or change.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative for duplication: the cluster is not one bug family, but many distinct regressions and feature requests.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Different problems: MoE auxiliary-loss normalization vs DeepSeekV3 implementation divergence." + "reason": "MoE-related, but one is auxiliary-loss normalization and the other is a broader implementation divergence; not the same bug." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Image processor segmentation-map handling is unrelated to Qwen2-VL multi-GPU training." + "reason": "Image-processor segmentation bug vs multi-GPU training topic; unrelated." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero probabilities and processor.tokenizer regression are separate code paths." + "reason": "Different failure modes: zero probabilities in SigLIP2 vs processor/tokenizer API regression." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "MLukeTokenizer task failure is unrelated to AutoTokenizer loading FastSpeech2ConformerTokenizer." + "reason": "Different tokenizer classes and errors; no shared code-path bug." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird mask-token decode bug is not the same as MLukeTokenizer AttributeError." + "reason": "BigBird special-token decode issue vs MLukeTokenizer AttributeError; unrelated." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "MoE loss normalization and T5 apex RMSNorm memory leak are unrelated." + "reason": "Both model-internal issues, but different components and failure mechanisms; not mergeable as one fix." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Missing PyTorch import and Trainer _gpu_broadcast_one NoneType error are different failures." + "reason": "ImportError for missing PyTorch vs trainer broadcast NoneType error; unrelated." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "BLIP2 dtype propagation is unrelated to RTDetrV2 output mismatch in v5." + "reason": "Different model families and different dtype/output issues." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "PyTorch ImportError and MacOS CLIP bus error do not share the same root cause." + "reason": "Import-time dependency error vs MacOS bus error; unrelated." }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to an isolated reproduction of another issue." + "reason": "No evidence they reproduce the same concrete crash path; only loose similarity around model loading/runtime failures." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "Missing backend import and embedding reinitialization after resize are different bugs." + "reason": "PyTorch missing dependency vs embedding reinitialization bug; unrelated." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model init breakage in Transformers 5.0 is not the same as FSDP CPU RAM-efficient loading." + "reason": "Custom model initialization regression vs FSDP CPU RAM efficient loading breakage; different code paths." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Auxiliary-loss normalization and tokenization changes are unrelated issues." + "reason": "Generic tokenization/version-change report vs MoE auxiliary-loss normalization; not the same defect." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "HF_HUB_OFFLINE model_info patching is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Offline hub metadata call bug vs tokenizer load failure for a specific model; different causes." }, { "left": "issue:44977", "right": "issue:45356", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug and Kimi-K2.5 tokenizer codec/regex regression are distinct." + "reason": "Qwen3.5 flash-attention generation bug vs Kimi tokenizer/codec regression; unrelated." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and MobileLLM key error are different model-loading/runtime issues." + "reason": "Qwen3-MoE sliding_window bug vs MobileLLM loading key error; different issues." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBird special-token decode bug is unrelated to DiaConfig custom token ID persistence." + "reason": "Tokenizer special-token registration bug vs config token-ID persistence bug; not the same underlying problem." }, { - "left": "issue:43329", - "right": "issue:45325", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "Undefined video-token helpers in multimodal counting are unrelated to Qwen2.5-VL rope-index scaling." + "reason": "Compiled-forward chunked generation inconsistency vs flash-attention generation bug; distinct failures." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:43329", + "right": "issue:45325", "accept": false, - "reason": "Compiled chunked-generation inconsistency is not the same as flash-attention generation failure." + "reason": "Different VLM video/rope bugs; both multimodal, but not the same code-path problem." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "BLIP2 dtype propagation and flash_attention s_aux=None crash affect different subsystems." + "reason": "Different model/dtype bug vs flash-attention crash on s_aux=None; unrelated." }, { "left": "issue:43550", - "right": "issue:43927", + "right": "issue:44361", "accept": false, - "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig token ID loss." + "reason": "torch.compile SDPA failure in Bamba vs tokenizer AttributeError; unrelated." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype handling does not match the MPS attention correctness issue." + "reason": "Blip2 dtype loading issue vs MPS attention correctness issue; different subsystems." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression in Qwen models is unrelated to Qwen3.5 num_labels propagation." + "reason": "Video-input regression in Qwen VL models vs num_labels propagation bug; unrelated." }, { - "left": "issue:43550", - "right": "issue:44361", + "left": "issue:43122", + "right": "issue:43906", "accept": false, - "reason": "Bamba compile/SDPA bug and MLukeTokenizer task AttributeError are unrelated." + "reason": "Tokenizer version-difference report vs isolated reproduction of another issue; not enough evidence of the same bug." }, { - "left": "issue:43122", - "right": "issue:43906", + "left": "issue:43550", + "right": "issue:44361", "accept": false, - "reason": "Tokenizer version-change behavior and an isolated reproduction of another issue are not the same bug." + "reason": "Model compile/SDPA crash vs tokenizer task AttributeError; unrelated." }, { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "NoneType GPU broadcast error and TrainerState token counter not updating are different trainer issues." + "reason": "Trainer broadcast NoneType error vs TrainerState counter not updating; different trainer behaviors." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to compiled chunked-generation output differences." + "reason": "Device-mapping cuda assert vs compiled chunked-generation inconsistency; different code paths." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "BEiT legacy-field loading failure and Bamba torch.compile/SDPA bug are unrelated." + "reason": "Beit pretrained-load legacy field bug vs Bamba compile/SDPA failure; unrelated." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "BEiT pretrained-load legacy-field bug does not match BitNet packed-weight loading failure." + "reason": "Legacy field preset loading issue vs packed-weight accelerate loading issue; different models and fixes." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer broadcast NoneType and Qwen2-VL multi-GPU training are separate problems." + "reason": "Trainer broadcast error vs Qwen2-VL multi-GPU training request; unrelated." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "BEiT legacy-field load failure is unrelated to OmDet-Turbo processor/model input-size mismatch." + "reason": "Different processor/model mismatch problems; not the same bug." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file loading breakage and inability to load a specific tokenizer are different loading paths." + "reason": "GGUF loading breakage vs inability to load a specific tokenizer; related only at a high level, not the same defect." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "MobileLLM key error and AutoImageProcessor-from-URL failure do not share a root cause." + "reason": "MobileLLM key error vs AutoImageProcessor-from-URL failure; unrelated." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug and save_pretrained visual-key regression are different regressions." + "reason": "Qwen3.5 output_hidden_states bug vs save_pretrained visual-key regression; distinct issues." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Embedding dtype drift in eval is unrelated to MoE auxiliary-loss normalization." + "reason": "Embedding dtype drift in eval vs MoE auxiliary-loss normalization; unrelated." } ] }, @@ -6637,20 +6607,20 @@ "issue:44423|issue:44734", "issue:43575|issue:45405" ], - "prepared_review_unit_hash": "86ecab0b66435cdf7d506536e5d6d71b450ad5878b109ea8e962f216352e12e6", + "prepared_review_unit_hash": "f6a5d4ae2a8b3b8bd5e9059c973e6013f160924cb383e52e6648af87a771b309", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38208, + "serialized_chars": 38205, "estimated_input_tokens": 9552, "estimated_eval_tokens": 19360 }, @@ -6659,216 +6629,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous and does not contain any clear duplicate pair; most items are unrelated model- or subsystem-specific regressions. I would use the broad tokenizer regression in issue 45812 as the nearest representative only because it is actionable and high-impact, not because the cluster is a true duplicate set.", + "summary": "The set is heterogeneous, and none of the soft pairs look like the same underlying bug or change. Most are either different model families, different failure modes, or one is documentation while the other is a runtime bug.", "confidence": 0.91, - "canonical_issue_reason": "No strong canonical duplicate exists; if one must be chosen, issue 45812 is the best representative because it describes a broad, user-visible tokenizer regression with clear impact.", + "canonical_issue_reason": "issue:43824 is the most central issue in the set by discussion/inbound references and is a clear user-facing regression.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the most suitable cluster representative because it is a concrete behavioral regression and more general than the other narrowly scoped reports.", + "best_issue_reason": "issue:43824 has the strongest signal of impact and cross-linking, so it is the best representative issue among these unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different failure modes: missing BOS/EOS vs wrong token IDs." + "reason": "Both are tokenizer-related, but one is missing BOS/EOS behavior for mdeberta while the other is wrong token IDs for Granite models; different bugs and model families." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training is unrelated to the GenerationMixin import error." + "reason": "Qwen2-VL multi-GPU training and a GenerationMixin import failure are unrelated code paths." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation mismatch and pipeline translation-task messaging are different bugs in different paths." + "reason": "SigLIP2 implementation mismatch vs an incorrect pipeline translation-task message; not the same issue." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Both mention rotary behavior, but one is a Smollm3 RoPE-layer bug and the other is GPTNeoX config reload persistence." + "reason": "Both mention rotary/rope settings, but one is a Smollm3 layer-selection bug and the other is GPTNeoX config reload regression." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio chat-template sample-rate handling is unrelated to the processor.tokenizer regression in multimodal custom code." + "reason": "Audio sampling-rate defaulting and processor.tokenizer/image handling are different regressions." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both involve missing pad-token fields, but they are separate model configs and likely separate regressions." + "reason": "Both are missing pad-token fields, but for different configs and different attribute names; not the same bug." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch and DiaConfig token-ID persistence are different code paths." + "reason": "CPU offload device mismatch in bark-small is unrelated to DiaConfig token-ID loss on save/load." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is unrelated to config save/load token ID loss." + "reason": "Packed-weight unpacking during accelerate loading is a different failure than DiaConfig losing custom token IDs." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug and MacOS CLIP bus error are not the same underlying issue." + "reason": "Tokenizer spacing bug vs MacOS bus error loading a CLIP model; completely different problems." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL is unrelated to a custom trainer signature change." + "reason": "Qwen2-VL multi-GPU training and a custom trainer loss-signature TypeError are unrelated." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Documentation example errors are not the same as the SigLIP2 HF-vs-JAX implementation discrepancy." + "reason": "One is documentation example errors, the other is an implementation discrepancy for SigLIP2; not the same concrete bug." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated failures." + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 parity mismatch and ZeRO-3 gradient growth are different problems." + "reason": "SigLIP2 parity discrepancy vs a ZeRO-3 gradient inflation bug; different code paths and symptoms." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Training/eval dtype drift is unrelated to the ZeRO-3 gradient anomaly." + "reason": "Embedding dtype drift in eval is not the same as abnormal gradient growth with DeepSpeed ZeRO-3." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "A model-loading regression and a dependency version bump are not the same bug." + "reason": "Phi-4 multimodal loading failure vs an unmet PEFT version requirement; different root causes." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Qwen3OmniMoe video-input processing and Qwen3.5 flash-attention generation are different issues." + "reason": "Qwen3OmniMoe video-processing ValueError and Qwen3.5 flash-attention generation issues are different bugs." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Qwen3-VL pad-token config loss and StableLmConfig pad_token_idx loss are separate model-specific regressions." + "reason": "Qwen3-VL pad_token_id missing and StableLmConfig pad_token_idx missing are separate model/config regressions." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Weight conversion during accelerate/deepspeed loading and Qwen3VL MoE loading failure are not the same code-path problem." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed differs from Qwen3VL MoE loading breakage in dev; not mergeable as one fix." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are unrelated." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated." }, { "left": "issue:43756", "right": "issue:45083", "accept": false, - "reason": "Smollm3 RoPE layer handling and a Qwen3 Omni feature-length helper bug are different model internals." + "reason": "Smollm3 RoPE layer dropping and a qwen3_omni_moe helper length bug are not the same defect." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Model loading failure for cjvt/sleng-bert and offline model_info patch behavior are distinct issues." + "reason": "Model loading failure for sleng-bert and an offline-mode model_info call bug in AutoTokenizer are different issues." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL and unwrapping text_config for AutoModelForCausalLM are different multimodal config bugs." + "reason": "Both touch Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is incorrect unwrapping of text_config in from_config." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Defaulting mm_token_type_ids to zeros and Gemma4 multi-GPU OOM are unrelated." + "reason": "Text-only finetuning defaults vs multi-GPU inference OOM are different Gemma4 problems." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization regression and lm_head.weight serialization inconsistency are separate problems." + "reason": "Custom model initialization breakage and device-dependent lm_head.weight serialization are distinct v5 regressions." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Importing Qwen2_5_VLForConditionalGeneration and Float8 storage lookup failures are unrelated." + "reason": "Qwen2_5_VL import failure and Float8 storage lookup failure are unrelated." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage handling and flash_attn import_utils KeyError are different errors." + "reason": "A Float8 storage TypeError and a flash_attn import-utils KeyError on Python 3.13 are different failures." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "A SigLIP2 docs example issue is not the same as the processor.tokenizer regression." + "reason": "Documentation example mistakes are not the same as the processor.tokenizer regression in v4.57.5." }, { "left": "issue:45229", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is OOM and the other is a device mismatch in PLE forward pass." + "reason": "Gemma4 OOM during multi-GPU inference and a device mismatch with device_map='auto' are separate runtime problems." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS CLIP bus error and Qwen2-VL multi-GPU training are unrelated." + "reason": "MacOS bus error on CLIP loading and Qwen2-VL multi-GPU training are unrelated." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing pad_token_idx and unexpected _is_hf_initialized in init_empty_weights are different compatibility issues." + "reason": "Missing pad_token_idx in StableLmConfig and an unexpected _is_hf_initialized arg under init_empty_weights are different compatibility issues." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field model load failure and CPU offload device mismatch are unrelated." + "reason": "Legacy-field model loading in Beit is unrelated to bark-small CPU offload device mismatch." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "DiaConfig token-ID persistence and MLukeTokenizer task AttributeError are different bugs." + "reason": "DiaConfig save/load token-ID loss and MLukeTokenizer task AttributeError are separate bugs." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both are serve regressions, but one is a multimodal '.to' crash and the other is KV-cache continuation indexing." + "reason": "Both are transformers serve crashes, but on different endpoints and different failure mechanisms: str.to on multimodal batching vs KV-cache tensor indexing." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "A tensor-parallel OOM and an unreleased dependency version bump are not the same issue." + "reason": "Qwen2-57B-A14B-Instruct OOM with tensor parallelism and a PEFT version bump are unrelated." } ] }, @@ -6941,8 +6911,8 @@ "issue:45823" ], "soft_pairs": [ - "issue:43504|issue:43653", "issue:44464|issue:45584", + "issue:43504|issue:43653", "issue:43994|issue:44493", "issue:43278|issue:43701", "issue:43606|issue:44112", @@ -6968,190 +6938,190 @@ "issue:43295|issue:44610", "issue:44466|issue:45127" ], - "prepared_review_unit_hash": "fbcffc2941192bd80d7ea3f8d8b620c50d5b129583411064c4ca3e6b9e85e14f", + "prepared_review_unit_hash": "e74357a0621218bf8748faddbced3295a90b489831a38446172933b0bde42115", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 35018, - "estimated_input_tokens": 8755, - "estimated_eval_tokens": 17766 + "serialized_chars": 35013, + "estimated_input_tokens": 8754, + "estimated_eval_tokens": 17764 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a duplicate cluster: the issues span many unrelated bugs across tokenization, model loading, generation, multimodal processing, device/dtype handling, and CI. The soft-similarity links are mostly broad subsystem overlap rather than the same concrete defect.", - "confidence": 0.96, - "canonical_issue_reason": "No canonical issue for dedupe: the items do not share one underlying bug, so selecting a representative would conflate unrelated reports.", + "summary": "The soft links are mostly false positives: the issues cover different models, subsystems, and failure modes, so none should be merged as duplicates. If a representative issue is needed, the most active and referenced one is the multimodal chat-template bug.", + "confidence": 0.56, + "canonical_issue_reason": "No clear canonical duplicate: the candidate pairs do not point to one shared underlying bug, so the cluster is better treated as heterogeneous.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue for dedupe: none of these issues is a true umbrella for the others, and any choice would be arbitrary.", + "best_issue_reason": "issue:44521 is the best representative by activity and inbound references, and it sits on an active multimodal API regression with broader user impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43504", - "right": "issue:43653", + "left": "issue:44464", + "right": "issue:45584", "accept": false, - "reason": "Different bugs: BEiT pretrained loading with a legacy field vs BigBird tokenizer special-token registration/decode output." + "reason": "Both involve generation, but they are different bugs in different code paths/models: compiled forward chunked generation vs Whisper empty-transcription behavior." }, { - "left": "issue:44464", - "right": "issue:45584", + "left": "issue:43504", + "right": "issue:43653", "accept": false, - "reason": "Both involve generation, but the defects are different: compiled chunked generation inconsistency vs Whisper empty-transcription failure after special-token alignment." + "reason": "Unrelated failures: BEiT semantic-segmentation loading with a legacy field vs BigBirdTokenizer special-token registration." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Unrelated symptoms: SigLIP2 wrong outputs in AutoModel/pipeline vs many models getting an unexpected position-id key." + "reason": "Different symptoms and likely different fixes: SigLIP2 output discrepancy vs a broad position-id key warning across many models." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Different subsystems: trainer token-count dtype/state update vs resume-from-checkpoint key mismatch." + "reason": "One is a dtype drift in evaluation; the other is a resume_from_checkpoint key mismatch." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both are device-related, but one is a CPU offload mismatch for Bark while the other is a stale GraniteSpeech CI test failure." + "reason": "Both mention device-related CI behavior, but they affect different models and different mechanisms (CPU offload mismatch vs stale override test)." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Completely unrelated: janus image-generation test failure vs a PEFT version pin being unreleased." + "reason": "Janus image-generation test failure and an unreleased PEFT version bump are unrelated." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "No shared defect: Flash Attention 2 test failures vs MLflow not reporting failed jobs." + "reason": "Flash Attention 2 test failures and MLflow reporting after training failure are unrelated subsystems." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Same chat-template area, but not the same bug: all-zero assistant masks for multimodal inputs vs a crash on tool-call assistant messages with no content." + "reason": "Both are chat-template related, but one is all-zero assistant masks for multimodal inputs while the other is a crash on tool-call assistant messages; not the same bug." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token-count updates." + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token-count tracking." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Different issues: a template-node compilation TypeError vs a flash_attn import KeyError on Python 3.13." + "reason": "Non-template-node compile errors and a flash_attn import KeyError on Python 3.13 are separate issues." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "Distinct problems: DeepSpeed ZeRO-3 gradient blow-up vs model loading breakage in latest transformers." + "reason": "ZeRO-3 gradient blow-up and a model-loading regression under latest transformers are different failure modes." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "Different defects: torch.jit.script GeneratorExp export failure vs SigLIP2 producing zero probabilities." + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability outputs." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "Both touch attention, but one is an MPS correctness bug and the other is a PyTorch version check/import bug." + "reason": "MPS attention correctness with differing head dims is not the same as a flex_attention import/version check bug." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation-state update after sync_gpus is unrelated to a Qwen3.5 output_hidden_states bug." + "reason": "Generation kwargs sync_gpus behavior and Qwen3.5 hidden-states output are different code paths." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Different code paths: Apple Silicon allocator warmup TypeError vs Qwen3.5 flash-attention generation issues." + "reason": "Apple Silicon allocator TypeError and Qwen3.5 flash-attention generation problems do not share a concrete bug." }, { "left": "issue:44568", "right": "issue:45701", "accept": false, - "reason": "Both mention tokenization, but 44568 is a specific BOS/EOS addition regression while 45701 is a broad version-dependent tokenization change report." + "reason": "A specific tokenizer regression for BOS/EOS addition is broader than a general report that transformers version changes tokenization." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Both are multimodal, but one is a video-processing ValueError and the other is missing/default mm_token_type_ids for text-only fine-tuning." + "reason": "Video processing failure in Qwen3OmniMoe and Gemma 4 text-only mm_token_type_ids behavior are separate multimodal issues." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Different failure modes: flash_attention utility import error vs meta-device model loading breakage." + "reason": "Flash-support import failure and meta-device model-loading regressions are unrelated." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Unrelated: dtype mismatches in bfloat16 inference vs Wav2Vec2 save_pretrained/tokenization failure." + "reason": "Bfloat16 dtype mismatches in CI and Wav2Vec2 save/tokenization failures are not the same underlying bug." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Same Gemma 4 family, but different bugs: missing mm_token_type_ids default vs device mismatch with device_map='auto'." + "reason": "Both are Gemma 4 related, but one is about defaulting mm_token_type_ids for text-only finetuning while the other is a device-map mismatch during forward." }, { "left": "issue:44315", "right": "issue:44849", "accept": false, - "reason": "No overlap in underlying bug: Liger Kernel not applied via model_init vs Qwen3.5 hidden-states behavior." + "reason": "Liger Kernel application via model_init and Qwen3.5 hidden-state output are unrelated." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "Attention correctness on MPS is unrelated to Kimi-K2.5 tokenizer codec handling/regression warning behavior." + "reason": "An MPS attention correctness issue and a Kimi-K2.5 tokenizer codec regression are unrelated." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy is unrelated to checkpoint resume key mismatches." + "reason": "SigLIP2 HF-vs-JAX discrepancy and checkpoint resume key mismatch do not describe the same bug." }, { "left": "issue:43873", "right": "issue:44464", "accept": false, - "reason": "Different issues: quantization/offloading behavior vs compiled-forward chunked generation inconsistency." + "reason": "Quantized offloading behavior and compiled-forward chunked generation inconsistencies are different problems." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Both involve processor/model interfaces, but one is a tokenizer/images regression and the other is an input-size mismatch for OmDet-Turbo." + "reason": "Processor/tokenizer API regression and OmDet-Turbo input-size mismatch are separate processor/model integration bugs." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Different bugs: device-dependent lm_head.weight serialization vs LoRA merge collapse with extended vocab and tied embeddings." + "reason": "Both involve tied weights/serialization themes, but one is device-dependent serialization of lm_head.weight while the other is LoRA merge collapse with extended vocab." } ] }, @@ -7253,17 +7223,17 @@ "issue:44360|issue:44561", "issue:43262|issue:43450", "issue:43819|issue:45083", - "issue:43819|issue:44112", - "issue:44610|issue:44857" + "issue:44610|issue:44857", + "issue:43819|issue:44112" ], - "prepared_review_unit_hash": "beb5a656a9268e294867a39076fcf3eccf996f7b71097431ecaa18a6c1bf98cc", + "prepared_review_unit_hash": "bb9f7b824f4c51fad782bbea4d688ce3891cf0b84271fd8da6259724d161dc06", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -7278,199 +7248,202 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a heterogeneous similarity cluster with no true duplicate pairs; the apparent matches are mostly broad topical similarities (tokenizer/config, multimodal processors, docs, training, or import issues) rather than the same underlying bug or change.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "This is a mostly heterogeneous cluster with one clearly plausible duplicate pair (the two SigLIP2 issues) and many adjacent-but-distinct processor/model regressions. For the representative issue, I\u2019d anchor on the most activity-heavy broad multimodal/import case rather than overclaiming that issue:44521 is strongest overall.", + "confidence": 0.8, + "canonical_issue_reason": "issue:43824 is a broad, high-activity multimodal import regression with the strongest discussion/inbound signal among the cluster\u2019s more generic candidates, so it works better as the representative issue than issue:44521.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43824 is the best global issue representative because it is active, broadly phrased, and sits near the center of the cluster\u2019s multimodal/import failures without relying on an unsupported claim of being the most active overall.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Different bugs: Perceiver image-resolution failure vs RT-DETR memory not released on deletion." + "reason": "Different model families and failure modes: Perceiver resolution handling vs RT-DETR memory release." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are import problems, but for different symbols and code paths; not the same missing export." + "reason": "Both are import errors, but they involve different symbols, models, and code paths." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLm pad token config bug vs a model loading failure for a different repository." + "reason": "Different loading regressions on different models; similarity is too broad." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Unrelated regressions: Janus test failure vs stale warning in paged generate()." + "reason": "Janus image-generation test failure and a stale generate warning are unrelated." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Both involve loading/config issues, but the symptoms and affected model families differ." + "reason": "Both concern loading, but the underlying problems are distinct." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 text-only fine-tuning mask IDs vs Qwen2.5-VL rope-index scaling bug are different code paths." + "reason": "Different Gemma/Qwen multimodal bugs affecting different preprocessing logic." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Same symptom class, but different model configs; no evidence of one shared fix." + "reason": "Same general missing-field pattern, but different configs and no clear shared defect." }, { "left": "issue:43844", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed gradient instability is unrelated to trainer.evaluate() failing after training." + "reason": "ZeRO-3 training instability and post-train evaluate failure are different issues." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Different processor regressions: processor.tokenizer/image handling vs batched video shape output." + "reason": "Processor.tokenizer regression and batched video-shape handling hit different paths." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Qwen2.5-VL import error and DeepSeek detokenization are unrelated." + "reason": "Different model-specific import/detokenization failures." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor.tokenizer regression vs quantization/offloading behavior; different failures." + "reason": "Custom processor API breakage and quantization offloading are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Both are label-handling bugs, but in different utilities and with different failure modes." + "reason": "Both mention labels, but one is metric gathering and the other is image-processor label reduction." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3 MoE FP8 config failure is unrelated to missing Qwen2_5_VL import export." + "reason": "Different model families and different failure classes." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 zero-probabilities bug vs RT-DETR memory leak are unrelated." + "reason": "SigLIP2 output mismatch and RT-DETR memory release are unrelated." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing bug vs embedding reinitialization after resize are different underlying issues." + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization on resize." }, { "left": "issue:38175", "right": "issue:43493", - "accept": false, - "reason": "Same model family, but zero probabilities and HF-vs-JAX discrepancy are not clearly the same bug." + "accept": true, + "reason": "Both are SigLIP2 implementation discrepancies, and the zero-probability symptom plausibly stems from the same HF-vs-reference mismatch." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Different tokenizer/model failures: AutoTokenizer loading vs Wav2Vec2 save/tokenization breakage." + "reason": "Tokenizer-related, but different model families and different bugs." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Embedding dtype mismatch during evaluate is unrelated to Qwen3 MoE memory usage." + "reason": "Embedding dtype drift in eval is not the same as MoE memory inefficiency." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both touch apply_chat_template in Qwen2.5-VL, but one is a padding crash and the other is empty assistant masks." + "reason": "Same processor area, but batched-input crashing and all-zero assistant masks are different failures." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they reference different deprecated APIs and pages." + "reason": "Both are docs issues, but they reference different APIs and regressions." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token bug vs GraniteSpeech CI test staleness are unrelated." + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device-override behavior are unrelated." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor batch-shape bug vs Qwen2 load OOM are unrelated." + "reason": "Batched video preprocessing and tensor-parallel OOM are different problems." }, { "left": "issue:43450", "right": "issue:43493", "accept": false, - "reason": "Different vision issues: batched video shape vs SigLIP2 implementation discrepancy." + "reason": "Batched video shape bug and SigLIP2 implementation mismatch are different paths." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Audio feature-extractor crash and Whisper empty-transcription regression are different bugs." + "reason": "Unsupported feature-extractor arg and empty-transcription generation are distinct audio bugs." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA/extended-vocab collapse and category-count limit are unrelated." + "reason": "LoRA/extended-vocab collapse and category-count limit errors are unrelated." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer design issue vs removed helper breaking trust_remote_code are unrelated." + "reason": "DSA indexer activation choice and a removed helper affecting trust_remote_code are different subsystems." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate default vs video batched-shape bug are unrelated." + "reason": "Audio apply_chat_template sampling-rate default and video batching are unrelated processor bugs." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent mismatch and qwen3_omni_moe length helper behavior are different code paths." + "reason": "DAC latent/forward mismatch is unrelated to qwen3_omni_moe feature-length handling." }, { - "left": "issue:43819", - "right": "issue:44112", + "left": "issue:44610", + "right": "issue:44857", "accept": false, - "reason": "DAC reconstruction bug vs GraniteSpeech CI test failure are unrelated." + "reason": "Preprocessing size mismatch and AMP crash are different issues." }, { - "left": "issue:44610", - "right": "issue:44857", + "left": "issue:43819", + "right": "issue:44112", "accept": false, - "reason": "OmDet processor size mismatch and LwDetr AMP crash are different vision bugs." + "reason": "DAC latent mismatch is unrelated to a GraniteSpeech CI test failure." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level summary is mostly conservative, but one key claim is not grounded in the packet: issue:44521 is not clearly the strongest by discussion/inbound activity (e.g. issue:43824 has higher discussion and inbound counts). Because the representative-issue choice is justified with an unsupported superlative, I can\u2019t accept this as fully precise. The soft-edge verdicts themselves look mostly conservative, including the single SigLIP2 acceptance, but the top-line rationale needs to be toned down or corrected." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -7539,27 +7512,27 @@ "issue:37428|issue:43425", "issue:43278|issue:43825", "issue:43335|issue:43550", - "issue:43606|issue:45083", "issue:44857|issue:45593", + "issue:43606|issue:45083", "issue:44448|issue:44743", "issue:37428|issue:44351", "issue:45200|issue:45588", - "issue:45290|issue:45381", "issue:29942|issue:32090", + "issue:45290|issue:45381", "issue:43611|issue:43931", "issue:39692|issue:43526", "issue:44206|issue:45823", "issue:44462|issue:44561", "issue:44561|issue:44779", "issue:43335|issue:43653", - "issue:43819|issue:44493", "issue:45310|issue:45478", + "issue:43819|issue:44493", "issue:45588|issue:45859", "issue:29127|issue:32090", "issue:43550|issue:44112", "issue:44292|issue:45520", - "issue:43335|issue:43504", "issue:44855|issue:45083", + "issue:43335|issue:43504", "issue:39692|issue:44493", "issue:43335|issue:45083", "issue:44448|issue:44779", @@ -7571,244 +7544,232 @@ "issue:34689|issue:35141", "issue:44448|issue:44977" ], - "prepared_review_unit_hash": "7e1fd1c2e78f343c388bca4d2a9f8690ab60e60bf9174c9b91048401266314df", + "prepared_review_unit_hash": "c8dc2187a0f187fa7f0b87cc35eb4edf4079ec7fea45342beefdf18d770cd956", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 34, - "serialized_chars": 37356, - "estimated_input_tokens": 9339, - "estimated_eval_tokens": 18934 + "serialized_chars": 37360, + "estimated_input_tokens": 9340, + "estimated_eval_tokens": 18936 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: nearly all soft edges are false positives across different models/code paths. The only clear duplicate is the Qwen3.5 MoE from_pretrained regression (issue:45310 and issue:45478), which should be merged as one.", - "confidence": 0.94, - "canonical_issue_reason": "issue:45310 is the better canonical for the only true duplicate pair: it reports the same Qwen3.5 MoE from_pretrained regression as issue:45478 and is the earlier report.", + "summary": "This cluster is mostly heterogeneous; the only clear duplicate pair is the identical Qwen3.5 MoE from_pretrained regression. The rest are related only at a very broad Transformers/modeling level, not the same underlying bug.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45310 is the cleanest canonical issue because it exactly matches issue:45478 in title and failure mode, and it was filed first.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue represents the whole set well because the cluster is mixed; if a representative is needed, issue:45310 is the strongest candidate since it is the only accepted duplicate pair's canonical node.", + "best_issue_reason": "issue:45310 is the best overall representative of the only true duplicate pair in this set; other issues are unrelated one-off bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Both are load/runtime failures, but one is a flash-attention import issue and the other is torch 2.10 incompatibility; different root causes." + "reason": "Different failures: a missing flash-attention utility import vs a general torch-version incompatibility." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype drift in eval vs a pipeline error-message regression; unrelated bugs." + "reason": "One is an embedding dtype regression; the other is a pipeline error-message issue." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Sparse-layer construction bug vs Bamba torch.compile/SDPA failure; different models and code paths." + "reason": "Different code paths: SwitchTransformers sparse-layer config vs Bamba torch.compile/SDPA failure." }, { - "left": "issue:43606", - "right": "issue:45083", + "left": "issue:44857", + "right": "issue:45593", "accept": false, - "reason": "CPU offload device mismatch in bark-small vs a qwen3_omni_moe helper-length bug; not the same issue." + "reason": "AMP/CUDA loss crash and missing auxiliary losses are separate bugs." }, { - "left": "issue:44857", - "right": "issue:45593", + "left": "issue:43606", + "right": "issue:45083", "accept": false, - "reason": "AMP crash in LwDetrImageLoss vs missing auxiliary losses in D-FINE; different functionality." + "reason": "CPU offload device mismatch is unrelated to qwen3_omni_moe feature-length behavior." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus output regression vs qwen modular recurrent-state reset; different code paths." + "reason": "Different models and root causes: Pegasus output drift vs Qwen recurrent-state reset." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Different import failures: missing flash-attention helper vs missing HybridCache export." + "reason": "Both are import errors, but for different symbols and different broken APIs." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Gemma4 token-type-id defaulting vs flash_attention.py crashing on s_aux=None; not the same bug." + "reason": "Gemma4 token-type-id defaulting and flash_attention aux-state crash are unrelated." }, { - "left": "issue:45290", - "right": "issue:45381", + "left": "issue:29942", + "right": "issue:32090", "accept": false, - "reason": "Chat-template tool-call crash vs qwen2.5-vl video position-id bug; distinct features." + "reason": "Flash Attention 2 test failures and a trainer broadcast NoneType error do not share the same bug." }, { - "left": "issue:29942", - "right": "issue:32090", + "left": "issue:45290", + "right": "issue:45381", "accept": false, - "reason": "Flash Attention 2 test failure vs Trainer/_gpu_broadcast_one NoneType error; unrelated." + "reason": "Chat-template tool-call handling and qwen2.5-vl video position IDs are different issues." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "base_model_prefix loading regression vs Qwen3-VL weight-shape mismatch; different failure modes." + "reason": "Both are loading errors, but they point to different models and different shape/attribute problems." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 docs example problems vs BeitImageProcessorFast reduce_labels bug; unrelated." + "reason": "Documentation example mistakes are not the same as BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "Unsupported center arg in LasrFeatureExtractor vs Gemma4 PLE device mismatch; different components." + "reason": "Different regressions: unsupported center argument vs Gemma4 PLE device mismatch." }, { "left": "issue:44462", "right": "issue:44561", "accept": false, - "reason": "Tokenizer.json precedence bug vs removal of is_torch_fx_available breaking trust_remote_code; unrelated." + "reason": "Tokenizer.json precedence and removal of is_torch_fx_available are unrelated breakages." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "One is a removed helper breaking remote-code models; the other is a DeepSeek tokenizer behavior regression." + "reason": "A removed utility breaking trust_remote_code is unrelated to Deepseek tokenizer output changes." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer bug vs BigBirdTokenizer special-token registration issue; different subsystems." - }, - { - "left": "issue:43819", - "right": "issue:44493", - "accept": false, - "reason": "DAC.from_latents mismatch vs widespread position-id key warnings; not the same underlying bug." + "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are different subsystems." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Same title and same Qwen3.5 MoE from_pretrained regression on transformers>=5.4.0; looks like a duplicate report." + "reason": "Exact same Qwen3.5 MoE from_pretrained error, same title and same failure mode." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC latent/forward mismatch and unexpected position-id keys are unrelated." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "flash_attention.py s_aux=None crash vs Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility; distinct code paths." + "reason": "Both mention offload/flash-attention-adjacent code, but the concrete failures differ." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "LayoutLMv3 error-message clarity issue vs Trainer broadcast NoneType error; unrelated." + "reason": "LayoutLMV3 error-message clarity and GPU broadcast NoneType are not the same bug." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba SDPA torch.compile failure vs stale GraniteSpeech CI test; different bugs." + "reason": "Bamba SDPA compile failure and GraniteSpeech stale device override are different code paths." }, { "left": "issue:44292", "right": "issue:45520", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure vs Python 3.13 flash_attn import KeyError; different triggers." + "reason": "Qwen-3 NVFP4 runtime failure and Python 3.13 flash_attn import KeyError are distinct." }, { - "left": "issue:43335", - "right": "issue:43504", + "left": "issue:44855", + "right": "issue:45083", "accept": false, - "reason": "Sparse-layer creation bug vs BEiT legacy-field loading bug; different models and mechanisms." + "reason": "Python parser/import failure versus qwen3_omni_moe feature-length helper behavior." }, { - "left": "issue:44855", - "right": "issue:45083", + "left": "issue:43335", + "right": "issue:43504", "accept": false, - "reason": "Python 3.13 DebertaV2Model import parse error vs qwen3_omni_moe feature-length helper behavior; unrelated." + "reason": "Different BEiT/SwitchTransformers loading bugs with no shared root cause." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 docs/example issues vs position-id warnings across many models; not the same bug." + "reason": "SigLIP2 docs issues and position-id key warnings are unrelated." }, { "left": "issue:43335", - "right": "issue:45083", + "right": "issue:45072", "accept": false, - "reason": "SwitchTransformers sparse-layer misconfiguration vs qwen3_omni_moe helper output issue; different scope." + "reason": "Sparse-layer creation and dtype mismatch in bfloat16 inference are different bugs." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus v4/v5 output difference vs DeepSeek tokenizer regression; unrelated." + "reason": "Pegasus output differences and Deepseek tokenizer regression are unrelated." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Zero-probability SigLIP2 issue vs unexpected position-id keys; different symptoms and code paths." - }, - { - "left": "issue:43335", - "right": "issue:45072", - "accept": false, - "reason": "Sparse-layer config bug vs dtype mismatch in bfloat16 inference; different root causes." + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as position-id key warnings." }, { - "left": "issue:43335", - "right": "issue:45593", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "SwitchTransformers config bug vs D-FINE auxiliary-loss omission; unrelated." + "reason": "Different models and failures: Pegasus v4/v5 output drift vs Qwen3.5 flash-attention generation." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "BitNet packed-weight loading bug vs GraniteSpeech stale test failure; unrelated." + "reason": "Packed-weight unpacking in BitNet and GraniteSpeech device override are unrelated." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs Gemma4 multi-GPU OOM; different problem classes." + "reason": "Qwen3.5 flash-attention generation issue and Gemma4 multi-GPU OOM are different problems." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Llama 3.2 Vision model-loading regression vs output-embedding reinit after resize; different issues." - }, - { - "left": "issue:44448", - "right": "issue:44977", - "accept": false, - "reason": "Pegasus output regression vs Qwen3.5 flash-attention generation problem; different models and failures." + "reason": "Llama 3.2 Vision model loading and post_init reinitializing embeddings are separate bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet: it correctly identifies the cluster as mostly heterogeneous, avoids overclaiming on the false soft edges, and only elevates the one deterministic duplicate pair (issue:45310 / issue:45478). The canonical choice is conservative and supported by the titles and timestamps." + "feedback": "The assessment is grounded in the packet: the only accepted pair, issue:45310 and issue:45478, is an exact title/failure match and the earlier filing is correctly identified as canonical. The remaining soft-edge verdicts stay conservative and are consistent with the issue titles and the packet\u2019s lack of stronger link evidence. No overstatement detected." }, "evaluator_used": true, "retried": false, @@ -7889,15 +7850,15 @@ "issue:43927|issue:45698", "issue:45459|issue:45712", "issue:44977|issue:45563", - "issue:43701|issue:45083", "issue:29942|issue:35141", "issue:37428|issue:44987", - "issue:43550|issue:45083", + "issue:43701|issue:45083", "issue:43873|issue:44493", "issue:45372|issue:45588", + "issue:43550|issue:45083", "issue:32090|issue:33357", - "issue:43278|issue:43329", "issue:43412|issue:45588", + "issue:43278|issue:43329", "issue:43873|issue:45229", "issue:45245|issue:45563", "issue:43976|issue:45440", @@ -7911,20 +7872,20 @@ "issue:43278|issue:43381", "issue:44448|issue:44466" ], - "prepared_review_unit_hash": "fbeca0750cb3c621b0ad176a184ca24cda0a247b7929498a9629068fed627518", + "prepared_review_unit_hash": "662ae0457c8e3485879e77b74eea02b7ab301bdd14bb22df5265639a2afd891e", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37206, + "serialized_chars": 37205, "estimated_input_tokens": 9302, "estimated_eval_tokens": 18860 }, @@ -7933,198 +7894,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is broadly heterogeneous, and none of the soft pairs look like the same underlying bug or change. Most pairs only share a subsystem or symptom category, not the concrete failure mode or code path.", - "confidence": 0.86, - "canonical_issue_reason": "No single issue cleanly represents this mixed set; the closest to a central, user-facing regression is issue 45372 because it is an explicit processor-loading break in the v5 line, but it is still not a true duplicate anchor for the whole cluster.", + "summary": "No duplicate cluster emerges here: every soft pair spans different concrete bugs or unrelated subsystems, so none should be merged.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue fits this set; the soft pairs are heterogeneous and appear to be false positives rather than duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45372 is the strongest standalone representative among these because it describes a concrete, high-impact loading failure with clear reproduction and impact, unlike many of the more niche or test-specific reports.", + "best_issue_reason": "No single issue is a good representative duplicate target because there is no accepted overlap across the candidates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43723", "right": "issue:44554", "accept": false, - "reason": "Tokenizer loading in v5 and an MPS attention correctness bug are unrelated code paths and failure modes." + "reason": "Tokenizer loading in v5 and an MPS attention correctness issue are unrelated code paths." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "IterableDataset multimodal batching and BeitImageProcessorFast label reduction are different components and bugs." + "reason": "Multimodal iterable-dataset failures and BEiT label reduction are different bugs." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch in Bark is unrelated to an AMP float16 crash in LwDetrImageLoss." + "reason": "CPU offload device mismatch and float16 AMP loss crashes are different failure modes." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Both mention dtype-related symptoms, but one is a loss crash under AMP and the other is a torch.compile SDPA issue in a different model path." + "reason": "Both involve dtype issues, but they affect different models and concrete code paths." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token handling and Gemma3 mixed-length sliding-window NaNs are different failures in different layers." + "reason": "Tokenizer special-token handling and Gemma3 NaN attention outputs are unrelated." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Extra whitespace after added tokens is a tokenizer formatting bug, not the Trainer/_gpu_broadcast_one NoneType error." + "reason": "Tokenizer spacing after added tokens is unrelated to Trainer broadcast NoneType errors." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and a bare-except swallowing errors in Fuyu tensor conversion are distinct bugs." + "reason": "Wav2Vec2 save/tokenization failures and a bare-except tensor conversion bug are not the same issue." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load behavior, but one loses custom token IDs in config while the other loads the wrong custom module; different root causes." + "reason": "Config token-ID persistence and wrong custom-module loading after save/load are different bugs." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Hidden tokenizer errors due to protobuf handling is unrelated to dummy classes leaking into dir(transformers) and repo checks." + "reason": "Tokenizer error masking and leftover dummy classes are unrelated maintenance issues." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "A flash-attention generation correctness issue and a stale warning for num_return_sequences are not the same defect." - }, - { - "left": "issue:43701", - "right": "issue:45083", - "accept": false, - "reason": "Checkpoint resume key mismatch and a multimodal feature-length helper bug are unrelated." + "reason": "Flash-attention generation failure and a stale warning for num_return_sequences are different symptoms." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Failed Flash Attention 2 tests and embedding reinitialization after resize are different bugs in different model setup paths." + "reason": "Flash Attention 2 test failures and embedding reinitialization on resize are unrelated." }, { "left": "issue:37428", "right": "issue:44987", "accept": false, - "reason": "An import error in flash attention utilities and a model loading failure for a specific repo are not the same concrete issue." + "reason": "Missing flash-attention import symbol and a specific model-load failure are not the same bug." }, { - "left": "issue:43550", + "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Bamba compile/SDPA failure and qwen3_omni_moe helper behavior are unrelated." + "reason": "Checkpoint resume key mismatch is unrelated to qwen3_omni_moe feature-length calculation." }, { "left": "issue:43873", "right": "issue:44493", "accept": false, - "reason": "Quantization offloading issues do not match unexpected position-id keys during model loading/serialization." + "reason": "Quantization offload issues and unexpected position-id keys are different problems." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import failure and a flash_attention.py crash on s_aux=None are different code paths and failure mechanisms." + "reason": "Gemma4 processor import failure and a flash-attention s_aux=None crash are unrelated." }, { - "left": "issue:32090", - "right": "issue:33357", + "left": "issue:43550", + "right": "issue:45083", "accept": false, - "reason": "Trainer broadcast NoneType and MacOS bus error with community CLIP are not the same bug." + "reason": "torch.compile/SDPA failure for Bamba is unrelated to qwen3_omni_moe output-length logic." }, { - "left": "issue:43278", - "right": "issue:43329", + "left": "issue:32090", + "right": "issue:33357", "accept": false, - "reason": "Embedding dtype drift in eval and undefined video-branch symbols in multimodal token counting are unrelated." + "reason": "Trainer GPU broadcast errors and a MacOS bus error on CLIP are different code paths." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export guard/support issues are not the same as sink-less flash-attention attribute handling." + "reason": "Executorch export guard/erfinv support and flash-attention None handling are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in eval and a multimodal token-count video-branch bug are unrelated." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization/offloading and multi-GPU OOM are both resource-related but clearly different problems." + "reason": "Quantization offloading and multi-GPU CUDA OOM are different resource-management issues." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-cardinality limits and a stale generate warning are unrelated." + "reason": "Category-count overflow and a stale generation warning are unrelated." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility and DeepSeekV3 implementation divergence are unrelated." + "reason": "Python version compatibility and DeepSeekV3MoE implementation divergence are unrelated." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Auto device mapping cudaErrorAssert and Gemma4 multi-GPU OOM are different runtime failures." + "reason": "Qwen3 device-mapping assertion and Gemma4 OOM are different models and failures." }, { "left": "issue:28218", "right": "issue:29127", "accept": false, - "reason": "Tokenizer extra-space behavior and LayoutLMv3 error-message clarity are unrelated." + "reason": "Tokenizer added-space behavior and LayoutLMv3 box validation messaging are unrelated." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor import failure and transformers serve missing _tokenizer are different loader/serving bugs." + "reason": "Gemma4 processor import failure and Gemma4Processor missing _tokenizer are distinct bugs." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "ReasoningEffort import breakage and protobuf-related tokenizer error masking are separate issues." + "reason": "Gemma4 processor import failure and protobuf error masking are unrelated." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong tokenizer decoder type and AutoTokenizer ignoring tokenizer.json are related only at a broad tokenizer level, not the same concrete bug." + "reason": "Wrong tokenizer decoder type and ignoring tokenizer.json are different tokenizer-loading issues." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Dummy Conv2d cleanup in Sam3PixelDecoder and SigLIP2 nonsensical outputs are unrelated." + "reason": "Dummy Conv2d cleanup and Siglip2 nonsensical outputs are unrelated." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP loss crash on CUDA and flash_attention s_aux=None are different failures in different modules." + "reason": "AMP/CUDA loss crash and flash-attention s_aux=None crash are different bugs." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Training/eval dtype mismatch and gradient checkpointing being disallowed in eval mode are not the same bug." + "reason": "Eval dtype drift and gradient-checkpointing-in-eval behavior are unrelated." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Both concern v5 behavior differences, but one is output parity for Pegasus and the other is lm_head serialization varying by device; not the same code-path problem." + "reason": "Pegasus output differences and lm_head serialization inconsistency are different concrete issues." } ] }, @@ -8229,20 +8190,20 @@ "issue:29127|issue:33357", "issue:43526|issue:44186" ], - "prepared_review_unit_hash": "3fa3d455539afcc059be98c2be452643517b3c699c504d850ea9a6cec2765819", + "prepared_review_unit_hash": "e5d6edfd8bf45ca3a88fd2fb92b8d4db3d53bd5d4a7a8367ab6d1068470fb84e", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36671, + "serialized_chars": 36670, "estimated_input_tokens": 9168, "estimated_eval_tokens": 18592 }, @@ -8251,192 +8212,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pairs look like the same underlying bug/change; they mostly share only a broad subsystem or vague theme, so all candidates are rejected.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "These items do not appear to form a real duplicate cluster; they span many unrelated bugs, regressions, and feature requests. I would reject all soft edges.", + "confidence": 0.92, + "canonical_issue_reason": "issue:45588 is the most concrete, localized bug report in the set (`integrations/flash_attention.py` crash on `s_aux=None`), but even so this is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45588 is the strongest standalone triage target because it has a specific code path, clear failure mode, and a plausible one-PR fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Unrelated failures: pipeline top_k output shape vs Trainer broadcast NoneType error." + "reason": "Pipeline `top_k` nesting bug and trainer `_gpu_broadcast_one` NoneType error are unrelated code paths." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Different topics: Sam3PixelDecoder dummy Conv2d vs RT-DETR memory not released." + "reason": "Dummy `Conv2d` in Sam3PixelDecoder is unrelated to RT-DETR memory being retained after deletion." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Both mention Gemma-like models, but one is flash-attention s_aux handling and the other is device_map/forward device mismatch." + "reason": "Flash-attention `s_aux=None` crash and Gemma4 device-map mismatch are different failures." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "Tokenizer special-token decode bug vs Qwen recurrent-state/cache behavior; no concrete overlap." + "reason": "BigBird mask-token registration/empty decode is unrelated to Qwen3.5 recurrent-state reset behavior." }, { "left": "issue:44451", "right": "issue:45520", "accept": false, - "reason": "Model loading regression vs flash_attn import KeyError on Python 3.13; different root causes." + "reason": "Model loading failure for ScandiBERT and Python 3.13 `flash_attn` import-key handling are different issues." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "Different tokenizers and failure modes: LayoutLMv2 NER/padding crash vs Wav2Vec2 save_pretrained/tokenization failure." + "reason": "LayoutLMv2 NER/padding crash and Wav2Vec2 save/tokenization failure do not share the same bug." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Feature extractor center-arg crash is unrelated to Qwen video-input regression." + "reason": "Unsupported `center` arg in `LasrFeatureExtractor` is unrelated to the Qwen video-input regression." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 inference quality issue vs RT-DETR memory lifecycle bug." + "reason": "SigLIP2 wrong outputs with AutoModel/pipeline and RT-DETR memory leak are unrelated." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save/tokenization failure." + "reason": "MLuke tokenizer task AttributeError is not the same as Wav2Vec2 save_pretrained/tokenization failure." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id config inconsistency vs Qwen3.5 flash-attention generation bug; different code paths." + "reason": "Gemma eos token config mismatch and Qwen3.5 flash-attention generation issue are different underlying bugs." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Both are quantization-related, but one is FP8 config failure and the other is saving quantized models; not the same bug." + "reason": "Qwen3Moe FP8 config failure and quantized-model saving failure are separate quantization problems." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Tokenizer post-processor update bug is unrelated to Gemma3 NaN embeddings from sliding-window attention." + "reason": "Runtime special-token post-processor update bug is unrelated to Gemma3 NaN embeddings from sliding-window attention." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export/torch_compilable_check failure is unrelated to Wav2Vec2 tokenization/save_pretrained issues." + "reason": "`torch.export`/`torch_compilable_check` failure is unrelated to Wav2Vec2 save/tokenization behavior." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures vs void segmentation map processing; different subsystems and symptoms." + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems and failures." }, { "left": "issue:43257", "right": "issue:43278", "accept": false, - "reason": "Qwen3 MOE weight conversion issue is unrelated to embedding dtype changing during evaluation." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to embedding dtype changing at eval time." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE-layer omission vs DeepseekV3MoE implementation divergence; not the same concrete defect." + "reason": "Smollm3 RoPE-layer omission and DeepSeekV3MoE implementation divergence are not the same bug." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "CLIP MacOS bus error vs SigLIP2 docs/model-processor mismatch and quantization failure; only loosely related vision models." + "reason": "MacOS CLIP bus error and SigLIP2 docs example errors are different issues despite both involving vision models." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error-message clarity vs TrainerState token counter not updating; unrelated." + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count tracking are unrelated." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType error vs DeepSpeed Adafactor OOM; different failure classes." + "reason": "`_gpu_broadcast_one` NoneType crash and DeepSpeed Adafactor OOM are distinct failure modes." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "OmDet-Turbo processor size mismatch vs RT-DETR memory leak; no shared bug." + "reason": "OmDet-Turbo processor size mismatch and RT-DETR memory leak are unrelated." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotations crash vs mistral_common ReasoningEffort import error; unrelated." + "reason": "Auto-docstring annotation crash and mistral_common `ReasoningEffort` import error do not share a code path." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Specific missing Qwen2.5-VL import vs generic dependency issue; too broad and not the same defect." + "reason": "Missing `Qwen2_5_VLForConditionalGeneration` export and a generic dependency issue are not the same problem." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "DeepseekV3MoE implementation divergence vs D-FINE auxiliary-loss omission; different models and behaviors." + "reason": "DeepSeekV3MoE divergence and D-FINE auxiliary-loss behavior are unrelated model bugs." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction vs checkpoint key mismatch; unrelated problems." + "reason": "Gradient checkpointing in eval mode and checkpoint key mismatch are different checkpointing problems." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird special-token decode bug vs FastSpeech2ConformerTokenizer loading failure; different tokenizers and symptoms." + "reason": "BigBird special-token registration bug is unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." }, { "left": "issue:43761", "right": "issue:45859", "accept": false, - "reason": "CLIPVision hidden_states regression vs Qwen3_5Moe CPU-offload incompatibility; unrelated components." + "reason": "CLIPVisionModel hidden-state regression and Qwen3_5Moe CPU-offload incompatibility are separate issues." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs Kosmos2.5 long-OCR index error; different code paths." + "reason": "BeitImageProcessorFast `reduce_labels` bug is unrelated to Kosmos2.5 OCR index error." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention 2 tests vs GenerationMixin import error; not the same bug." + "reason": "Flash Attention 2 test failures and `GenerationMixin` import error are unrelated." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error messaging issue vs CLIP MacOS bus error; unrelated." + "reason": "LayoutLMv3 error clarity and CLIP MacOS bus error are not the same failure." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "ImageProcessorFast label handling vs LayoutLMv2 tokenizer crashes; different artifacts and failure modes." + "reason": "BeitImageProcessorFast label handling and LayoutLMv2 tokenizer crashes are different components and bugs." } ] }, @@ -8529,211 +8490,211 @@ "issue:44442|issue:45803", "issue:43761|issue:44188", "issue:43316|issue:43525", - "issue:42907|issue:45698", - "issue:44479|issue:45290", "issue:33290|issue:34567", + "issue:44479|issue:45290", "issue:44442|issue:44898", + "issue:42907|issue:45698", "issue:34689|issue:43611", "issue:39692|issue:45491", - "issue:33290|issue:35141", - "issue:33290|issue:34689" + "issue:33290|issue:34689", + "issue:33290|issue:35141" ], - "prepared_review_unit_hash": "363f2f0acec12602667396e675a1f6d80f392e5384d7d6888aa5515fb1935c11", + "prepared_review_unit_hash": "5dd5b343c5c64766c489f2cbe907fa395aafe98290de50da1f7a7ec15ef7479f", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35645, - "estimated_input_tokens": 8912, - "estimated_eval_tokens": 18080 + "serialized_chars": 35643, + "estimated_input_tokens": 8911, + "estimated_eval_tokens": 18078 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly unrelated Transformers issues across different models/features; none of the soft pairs look like the same underlying bug or change.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 45698 is the best single representative because it is concrete, actively discussed, and has inbound references, but the overall set is heterogeneous rather than one duplicate cluster.", + "summary": "This cluster is heterogeneous: the issues span unrelated model families, trainer/runtime regressions, tokenizer/processor bugs, and memory/loading failures. None of the proposed soft pairs look like the same underlying bug or change, so I would not merge any of them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is a grab bag of unrelated bugs, and the soft pairs are only superficial title-level similarities.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the strongest standalone issue: it has discussion/inbound refs and a clear reproducible loading regression.", + "best_issue_reason": "issue:45698 is the best representative only because it is open, has inbound references, and concerns a broad save/load custom-module regression with ongoing impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Different model families and failure modes: missing `pad_token_id` vs `sliding_window` behavior." + "reason": "Different failures: missing pad_token_id on Qwen3-VL text config vs a sliding_window/recurrent-state bug in Qwen3-MoE." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "Unrelated bugs: layoutlmv3 error messaging vs token-embedding resize/reinit." + "reason": "Unrelated bugs: unclear LayoutLMv3 box error messaging vs token-embedding resizing reinitializing the output embedding." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Different code paths: compiled chunked generation inconsistency vs Gemma3 sliding-window NaNs." + "reason": "Both involve generation, but one is compiled forward chunking inconsistency and the other is a Gemma3 sliding-window NaN issue with mixed-length batches." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Qwen3.5 flash-attention generation issue is unrelated to the category-count runtime error." + "reason": "Completely different problems: FlashAttention generation regression vs a categories-cardinality RuntimeError." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Different problems: FP8 config failure vs Qwen3-VL weight-shape loading mismatch." + "reason": "Different root causes: FineGrainedFP8Config failure on Qwen3Moe vs weight-shape mismatch when loading Qwen3-VL." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Tokenizer-loading failure and dtype-mismatch inference bug are not the same issue." + "reason": "Tokenizer-loading failure and bfloat16 dtype mismatch are unrelated despite both being model-adjacent." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Both are audio-related, but one is chat-template sampling-rate defaulting and the other is an unsupported `center` arg crash." + "reason": "Both touch audio processing, but one is a chat-template sampling-rate default bug and the other is an unsupported `center` argument crash." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "Different subsystems: label reduction in image preprocessing vs AMP crash in LwDetrImageLoss." + "reason": "Different subsystems: BEiT label reduction returns the wrong shape, while LwDetrImageLoss crashes under float16 AMP on CUDA." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs sync bug and metric-gather label truncation are separate training-time issues." + "reason": "Generation kwargs syncing and metric-gathering label truncation are separate code paths with different failure modes." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM and resume-from-checkpoint key mismatch are unrelated." + "reason": "OOM during tensor-parallel model load is not the same bug as a resume_from_checkpoint key mismatch." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both touch flash attention, but they describe different concrete crashes in different code paths." + "reason": "Both mention FlashAttention, but they hit different models and different failures: Ministral-3 regression vs `s_aux=None` AttributeError." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Recurrent-state caching bug is unrelated to Whisper batch_decode special-token handling." + "reason": "Unrelated: Qwen3.5 recurrent-state reset bug vs Whisper batch_decode ignoring skip_special_tokens." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both mention OOM, but one is tensor-parallel loading and the other is int4 quantization memory growth." + "reason": "Both are OOMs, but one is TP model loading and the other is increased reserved memory under int4 quantization; not the same issue." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "General custom-model init breakage is not the same as a specific model class's optional `post_init` issue." + "reason": "General custom-model initialization breakage in v5 is broader and distinct from the specific patchtsmixer post_init constraint." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe `sys.modules` access is unrelated to DeepSeekV3MoE divergence from remote implementation." + "reason": "sys.modules access safety bug and DeepSeekV3MoE implementation divergence are unrelated." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "A missing `hidden_states` output in CLIPVisionModel is not clearly the same as the ModelOutput key-assignment bug." + "reason": "The former is a CLIPVisionModel forward regression; the latter is a general ModelOutput key-assignment bug. Similar symptoms, different bugs." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer loading in v5 and degenerate zero-loss sequence classification are unrelated." + "reason": "Tokenizer auto-loading and degenerate zero-loss classification are completely different problems." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "AutoTokenizer load failure and bare-except swallowing errors are different bugs." + "reason": "Tokenizer loading failure is unrelated to Fuyu's bare-except tensor conversion bug." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "Hidden-states regression and attention-kernel divergence under `torch.compile` are different problems." + "reason": "Both are regressions, but one is hidden_states output handling and the other is attention-kernel divergence under torch.compile." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config API complaints, but they concern different models and different missing fields." + "reason": "Both involve config APIs, but one is a broad Gemma3TextConfig discrepancy and the other is a specific missing pad_token_id on Llama4Config." }, { - "left": "issue:42907", - "right": "issue:45698", + "left": "issue:33290", + "right": "issue:34567", "accept": false, - "reason": "Save/dequantization failure is unrelated to wrong custom-module loading after save_pretrained." + "reason": "Optimizer/deepspeed OOM and TrainerState token-count bookkeeping are unrelated." }, { "left": "issue:44479", "right": "issue:45290", "accept": false, - "reason": "Video-input regression and chat-template tool-call crash are separate user-facing failures." + "reason": "Video-input regression across Qwen VL variants is unrelated to chat-template crashes on tool-call assistant messages." }, { - "left": "issue:33290", - "right": "issue:34567", + "left": "issue:44442", + "right": "issue:44898", "accept": false, - "reason": "Deepspeed Adafactor OOM and TrainerState token counting are unrelated." + "reason": "Tokenizer loading and Perceiver non-default-resolution classification failure are not the same code path." }, { - "left": "issue:44442", - "right": "issue:44898", + "left": "issue:42907", + "right": "issue:45698", "accept": false, - "reason": "Tokenizer loading and Perceiver interpolation at non-default resolution are unrelated." + "reason": "Saving dequantized models and loading the wrong custom module after save_pretrained are different save/load bugs." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Both are model-loading regressions, but they arise from different root causes and model families." + "reason": "Both affect model loading, but they target different mechanisms: a specific Llama 3.2 vision regression vs base_model_prefix handling in v5." }, { "left": "issue:39692", "right": "issue:45491", "accept": false, - "reason": "Documentation example errors are unrelated to Gemma3 sliding-window NaNs." + "reason": "Documentation-example errors and Gemma3 NaN embeddings are unrelated." }, { "left": "issue:33290", - "right": "issue:35141", + "right": "issue:34689", "accept": false, - "reason": "Adafactor OOM in deepspeed is unrelated to embedding reinitialization on resize." + "reason": "OOM in deepspeed/adafactor and model-loading failure for Llama 3.2 vision are different issues." }, { "left": "issue:33290", - "right": "issue:34689", + "right": "issue:35141", "accept": false, - "reason": "OOM during training with Adafactor and broken model loading are different issues." + "reason": "Completely different: training-memory OOM vs embedding reinitialization during post_init after resizing tokens." } ] }, @@ -8811,16 +8772,16 @@ "issue:43493|issue:43827", "issue:43550|issue:44743", "issue:44077|issue:44479", - "issue:43316|issue:45759", "issue:41628|issue:45372", + "issue:43316|issue:45759", "issue:43720|issue:44442", "issue:43643|issue:44913", "issue:43749|issue:43931", "issue:43873|issue:44871", "issue:33290|issue:36331", "issue:34689|issue:36010", - "issue:44190|issue:44496", "issue:43381|issue:43856", + "issue:44190|issue:44496", "issue:44855|issue:45137", "issue:43381|issue:43742", "issue:43316|issue:44877", @@ -8830,206 +8791,200 @@ "issue:43673|issue:44743", "issue:43065|issue:44617", "issue:43404|issue:43441", - "issue:43388|issue:43688", "issue:44162|issue:44792", + "issue:43388|issue:43688", "issue:43295|issue:43825" ], - "prepared_review_unit_hash": "c7a33a310b454181bdb31f513354e13628988b28392df59b4de481bd64b4366f", + "prepared_review_unit_hash": "96e4d0039f1a1eed1367b69a33c25ed61b1c174dbb3820739befde62c59dd65a", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35277, - "estimated_input_tokens": 8820, - "estimated_eval_tokens": 17896 + "serialized_chars": 35282, + "estimated_input_tokens": 8821, + "estimated_eval_tokens": 17898 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab bag of unrelated Transformers issues across loading, generation, multimodal, training, and config behavior. The soft-similarity pairs share broad keywords in a few places, but none describe the same concrete bug or change, so no duplicate merge candidate is strong enough to accept.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "This cluster is highly heterogeneous: it mixes unrelated import, config, training, loading, and multimodal processing issues. None of the soft-similarity pairs look like the same underlying bug, so all soft edges are rejected.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43825 is the clearest, most self-contained user-facing regression in the set, with a specific problem statement and broad pipeline-v5 relevance.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43825 is the safest anchor if one issue must represent the cluster: it is actionable, well-scoped, and easier to generalize than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Both mention training/runtime behavior, but one is about gradient checkpointing in eval mode and the other about quantization offloading; different code paths and failures." + "reason": "Different failure modes: gradient checkpointing vs quantization/offloading." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Both are Qwen model-loading regressions, but one is a weight-shape mismatch at load time and the other is an incorrect save_pretrained checkpoint; not the same bug." + "reason": "Both involve Qwen models, but one is a shape mismatch on load and the other is a save_pretrained regression." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch vs documentation still referencing removed pipelines; one is model correctness, the other is docs cleanup." + "reason": "SigLIP2 implementation discrepancy is unrelated to docs still mentioning pipeline()." }, { "left": "issue:43550", "right": "issue:44743", "accept": false, - "reason": "One is a Bamba torch.compile/SDPA failure, the other is Qwen3.5 recurrent state reset with cache; unrelated model internals." + "reason": "Different model families and bugs: torch.compile/SDPA vs recurrent state reset with cache." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "PatchTSMixer post_init validation and Qwen video regression are different model-family issues with no shared failing code path." + "reason": "Unrelated model-specific validation issue versus a video-input regression." }, { - "left": "issue:43316", - "right": "issue:45759", + "left": "issue:41628", + "right": "issue:45372", "accept": false, - "reason": "Both are config/API discrepancies, but for different composite-model behaviors and different objects; not one underlying defect." + "reason": "Both are import errors, but caused by different packages and different processor paths." }, { - "left": "issue:41628", - "right": "issue:45372", + "left": "issue:43316", + "right": "issue:45759", "accept": false, - "reason": "Both are import errors, but one is missing AutoImageProcessor export and the other is a mistral_common ReasoningEffort dependency break." + "reason": "Related to config APIs, but these are different composite-model bugs with different triggers." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "Packed-weight loading for BitNet vs tokenizer loading for FastSpeech2Conformer are unrelated loader paths." + "reason": "Packed-weight loading bug is unrelated to tokenizer loading failure." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "One is missing fields when using trust_remote_code in AutoConfig, the other is GPTNeoXConfig losing rotary_pct on reload; similar theme, different bug." + "reason": "Both are config reload problems, but one drops missing fields while the other resets a specific parameter." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "Both involve loading, but one is FSDP CPU RAM efficient loading and the other is a specific Qwen3-VL shape mismatch; not the same fix." + "reason": "Different subsystems: FSDP CPU RAM loading versus Qwen3-VL shape mismatch." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization offloading problems and Gemma eos_token_id inconsistency are unrelated concerns." + "reason": "Quantization/offloading behavior is unrelated to Gemma eos_token_id inconsistency." }, { "left": "issue:33290", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed Adafactor OOM and CustomTrainer compute_loss signature mismatch are different training failures." + "reason": "Optimizer OOM in DeepSpeed is unrelated to a trainer method signature change." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Model loading regression vs GenerationMixin import failure; different subsystems and symptoms." + "reason": "Both are model/loading errors, but one is a specific model regression and the other is a missing import path." }, { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:43381", + "right": "issue:43856", "accept": false, - "reason": "Local dataset loading in a training script and unrecognized model config are unrelated issues." + "reason": "Eval-mode gradient checkpointing and Qwen3 MoE memory usage are different problems." }, { - "left": "issue:43381", - "right": "issue:43856", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "Both relate loosely to efficiency, but one is checkpointing behavior and the other is MoE memory usage during training; not the same defect." + "reason": "Dataset loading in a training script is unrelated to missing model_type in config.json." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 JIT parsing error and a DeepSpeed Zero3 deque error are completely separate failures." + "reason": "Python parser error in torch.jit scripting is unrelated to DeepSpeed deque underflow." }, { "left": "issue:43381", "right": "issue:43742", "accept": false, - "reason": "Gradient checkpointing in eval mode and a MobileLLM loading key error do not share the same code path." + "reason": "Gradient checkpointing eval-mode issue is unrelated to a MobileLLM key error." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Both concern config handling, but one is an API discrepancy between configs and the other is strict config rejection for granite_speech; different behavior." + "reason": "Both involve config handling, but one is an API discrepancy and the other is a strict loader rejection." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "TensorFlow backend packaging vs Gemma3TextConfig API mismatch are unrelated." + "reason": "Packaging/backend dependency issue is unrelated to Gemma3TextConfig API mismatch." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both touch multimodal/video processing, but one is a Qwen3OmniMoe video input error and the other is a batched apply_chat_template crash; different entry points." + "reason": "Both are multimodal input bugs, but they affect different processors, inputs, and code paths." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k nesting bug and missing PyTorch import are unrelated issues." + "reason": "Pipeline top_k nesting bug is unrelated to a missing PyTorch dependency error." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache, but one is a missing GenerationMixin cache in v5 chunked_prefill and the other is recurrent state reset in modular_qwen3_5; different mechanisms." + "reason": "Both mention cache, but one is GenerationMixin cache availability and the other is recurrent state reset logic." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Sam3PixelDecoder dummy layer discussion and Sam3Video OOM are not the same bug." + "reason": "Dummy Conv2d implementation in Sam3PixelDecoder is unrelated to a Sam3Video OOM report." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Both are Mistral-family issues, but one is tied-weight correctness and the other is FlashAttention compatibility; they would not be merged as one PR." - }, - { - "left": "issue:43388", - "right": "issue:43688", - "accept": false, - "reason": "Metric gathering drops tuple labels vs auxiliary loss normalization are separate training/eval bugs." + "reason": "Different Mistral-family issues: weight tying versus FlashAttention failure." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 model breakage and a Janus image-generation test failure are unrelated." + "reason": "Completely unrelated: ESM2 breakage versus a janus image-generation test failure." }, { - "left": "issue:43295", - "right": "issue:43825", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "Processor/tokenizer regression and pipeline translation-message regression are different v5 issues with different fixes." + "reason": "Metric label dropping and auxiliary-loss normalization are different training/eval bugs." }, { - "left": "issue:43550", - "right": "issue:44514", + "left": "issue:43295", + "right": "issue:43825", "accept": false, - "reason": "A torch.compile+SDPA model failure is not the same as a batched processor chat-template crash, despite both involving Qwen-family models." + "reason": "Both are regression reports, but one is processor/tokenizer behavior and the other is pipeline v5 messaging." } ] }, @@ -9135,220 +9090,220 @@ "issue:44464|issue:45563", "issue:43906|issue:45290" ], - "prepared_review_unit_hash": "e4b6a3c9bf97fe1d454232ee40ba6d1798ead916ac602810a4e80054e923047d", + "prepared_review_unit_hash": "e62f28d9f211f864c2b85fabe8c2daedf37fadee412a754206fd5fe0789812df", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36838, - "estimated_input_tokens": 9210, - "estimated_eval_tokens": 18676 + "serialized_chars": 36836, + "estimated_input_tokens": 9209, + "estimated_eval_tokens": 18674 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of unrelated issues; there is no meaningful duplicate cluster or shared underlying bug/change. No PRs are present.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue fits: the items span unrelated loading, generation, config, vision, trainer, and docs problems.", + "summary": "This cluster is a grab bag of mostly unrelated Transformers issues across loading, generation, multimodal processors, quantization/offload, and model-specific regressions. None of the soft pairs look like true duplicates or the same fixable code-path bug, so all soft edges should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:44521 is the best anchor: it is open, has the most external attention in this set, and describes a concrete reproducible multimodal bug rather than a vague question or isolated model-specific failure.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global representative for the cluster because the artifacts do not describe one common defect.", + "best_issue_reason": "issue:44521 is the strongest standalone issue to keep: specific reproduction, active discussion, and likely to benefit from continued triage/fix work.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "Unrelated: model_max_length question for DeBERTa-V3 vs missing transformers.onnx module." + "reason": "Different problems: a model_max_length question for DeBERTa-V3 vs a missing `transformers.onnx` module import error." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Different bugs: video-classification processor lookup vs SAM3 vision encoder missing attribute." + "reason": "Both involve video/SAM-like models, but one is a pipeline processor selection bug and the other is a missing attribute on a specific SAM3 output object." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Both are v5-related, but one is a PatchTSMixer post_init change and the other is non-persistent buffer initialization." + "reason": "Different code paths: one is about `post_init` being optional for patchtsmixer, the other about non-persistent buffers being filled with junk in v5." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Distinct symptoms and code paths: Gemma eos_token_id mismatch vs Whisper empty-transcription generation failure." + "reason": "Different bugs: Gemma-3 eos token configuration mismatch vs Whisper failing on empty transcription after `align_special_tokens`." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "Different model behaviors: LoRA merge/tied embeddings collapse vs Gemma 4 mm_token_type_ids defaulting." + "reason": "Both relate to embeddings/tokens, but one is LoRA merge collapse with tied embeddings and the other is Gemma 4 multimodal token type IDs defaulting to zeros." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "Unrelated: torch.compile SDPA OOM on Bamba vs resume_from_checkpoint key mismatch." + "reason": "Unrelated failures: a torch.compile/SDPA issue for Bamba vs a resume-from-checkpoint key mismatch." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Both are regressions, but they affect different models and failures: hidden_states=None vs video input regression." + "reason": "Different models and regressions: CLIPVision hidden_states handling vs Qwen video-input regression in v5.3.0." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "Device mismatch during CPU offload is unrelated to checkpoint key mismatch." + "reason": "Different issues: CPU offload device mismatch for bark-small vs checkpoint key mismatch during resume." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer API regression is not the same bug as the Gemma eos_token_id configuration mismatch." + "reason": "Processor/tokenizer regression vs Gemma-3 eos_token configuration mismatch; not the same underlying bug." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Different loading failures: tensor-parallel OOM for Qwen2-57B vs key error loading MobileLLM-125M." + "reason": "Both are load-time failures, but one is TP/OOM for Qwen2-57B and the other is a key error loading MobileLLM-125M." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3d_parallel.py execution failure is unrelated to Qwen2 model loading OOM." + "reason": "A runtime script failure (`3d_parallel.py`) and a model-loading OOM are different root causes." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "Both involve model loading, but gguf_file breakage and missing model_type are separate root causes." + "reason": "Different loading problems: `gguf_file` breaking tokenizer/model from_pretrained vs an unrecognized model type in config.json." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage is different from GPTNeoX rotary_pct not persisting on reload." + "reason": "Custom model initialization breakage in v5 is unrelated to GPTNeoX rotary_pct not persisting on reload." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Serve KV-cache indexing crash is unrelated to Gemma4Processor missing _tokenizer." + "reason": "Different subsystems: KV-cache continuation indexing in serve vs Gemma4Processor missing `_tokenizer`." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Legacy field handling in Beit loading is not the same issue as wrong custom module resolution after save/load." + "reason": "Legacy-field preset loading for BEiT is unrelated to wrong custom module resolution after save_pretrained." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty transcription handling." + "reason": "CUDA OOM in Sam3Video is not the same bug as Whisper failing on empty transcription." }, { "left": "issue:43575", "right": "issue:43873", "accept": false, - "reason": "Memory pressure on Qwen2 TP load and quantized offloading behavior are different problems." + "reason": "Both mention large-model memory behavior, but one is TP loading OOM and the other is quantization/offloading not working as expected." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is unrelated to missing load-balancing loss." + "reason": "Different model behaviors: CLIPVision hidden_states regression vs MoE load-balancing loss omission." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Weight-shape mismatch while loading Qwen3-VL is not the same as a wrong tied-weights warning in Qwen3-Next." + "reason": "Qwen3-VL weight-shape mismatch on load is unrelated to the Qwen3-Next tied-weights warning bug." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error and token-embedding reinit in post_init are separate bugs." + "reason": "Trainer `_gpu_broadcast_one` NoneType error and embedding reinitialization after resize are different failure modes." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "A stale num_return_sequences warning is unrelated to Whisper failing on empty transcription." + "reason": "A stale warning in paged generate is not the same as Whisper's empty-transcription crash." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward has no connection to Gemma eos_token configuration." + "reason": "Chunked generation inconsistency with compiled forward is unrelated to eos_token_id config mismatch." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation bug are different model-specific issues." + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation failure are separate model-specific regressions." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression across Qwen VL variants is unrelated to serve KV-cache indexing." + "reason": "Both are generation-related, but one is a video-input regression for specific Qwen VL models and the other is a serve KV-cache indexing crash." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as OmDet-Turbo processor producing the wrong image size." + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo processor size mismatch are different vision-model issues." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to SigLIP2 producing nonsensical outputs." + "reason": "Docs still referencing removed `pipeline()` APIs is unrelated to SigLIP2 producing nonsensical outputs." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk values and GPTNeoX rotary_pct reload loss are different state/config serialization bugs." + "reason": "Non-persistent buffer initialization junk and GPTNeoX rotary_pct not reloading are unrelated configuration/runtime bugs." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks for multimodal chat templates is different from wrong vision_position_ids in Qwen2.5-VL video input." + "reason": "Both are multimodal/video-adjacent, but one is assistant mask construction in chat templating and the other is wrong video vision position IDs." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "A missing image_utils import is unrelated to Qwen3.5 chat crashes." + "reason": "A missing import from `image_utils` is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Chunked generation inconsistency and paged-generate stale warning are different generation issues." + "reason": "Both touch generation, but one is a correctness regression in compiled chunked generation and the other is only a stale warning." }, { "left": "issue:43906", "right": "issue:45290", "accept": false, - "reason": "A reproduction of issue #38071 is not the same as tool-call assistant message crashes in apply_chat_template." + "reason": "An isolated reproduction of a prior issue and a chat-template crash on tool-call messages are different bugs." } ] }, @@ -9450,25 +9405,25 @@ "issue:36246|issue:38617", "issue:42175|issue:44279", "issue:44016|issue:44220", - "issue:44186|issue:44734", "issue:38175|issue:43065", + "issue:44186|issue:44734", "issue:44466|issue:44913", "issue:44683|issue:45446" ], - "prepared_review_unit_hash": "40a46983a8324e0b9ad5195be4b74e0b65e0d0b09363743337d4162c2222d3cb", + "prepared_review_unit_hash": "0a90c049b74b336552fa37f4498ae8821ed6b21ab2c3be03b103c757bbb18726", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37512, + "serialized_chars": 37510, "estimated_input_tokens": 9378, "estimated_eval_tokens": 19012 }, @@ -9477,204 +9432,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly unrelated Transformers bugs spanning imports, configs, serialization, pipelines, and device/runtime issues. The soft pairs look superficially similar but do not describe the same underlying defect, so none should be merged as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue fits this set: the artifacts cover many distinct codepaths and failure modes with only broad topical overlap.", - "canonical_pr_reason": "No PRs are present in the cluster.", - "best_issue_reason": "issue:41628 is the cleanest representative because it is a focused, reproducible import-regression report with a clear codepath and broad user impact.", + "summary": "This set is heterogeneous: most pairs are only superficially similar (often just the same subsystem or API area) and do not describe the same underlying bug. I would not merge any of the soft edges; the strongest recurring theme is import/API regressions, but even those are distinct failures.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 41628 is the broadest import/API-regression report among the set and is the closest match to the other import-related items, but the cluster overall is too mixed to treat it as a true duplicate center.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41628 is the best single representative if one must be chosen because it reflects a common user-facing import failure pattern, but it should not be used as a duplicate canonical for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Both involve loading problems, but one is a local dataset/script issue and the other is a specific Phi-4 multimodal model load failure." + "reason": "Both are model/script loading problems, but one is a local dataset issue in an example script and the other is a multimodal model load failure; different code paths." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 auto-device-map CUDA assert and a category-count limit error are unrelated failures in different codepaths." + "reason": "Qwen3 auto device mapping cuda assert is unrelated to the 2^24 category-limit runtime error." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "A janus test failure and stale auto_mappings config references are different bugs with different causes." + "reason": "Janus image-generation test failure and Sam3LiteText auto_mappings CI breakage are different failures in different areas." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import parsing of DebertaV2 is unrelated to the flash_attention crash on s_aux=None." + "reason": "Python 3.13 import parsing error in DebertaV2Model is unrelated to the flash_attention s_aux=None crash." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "XLNet CPU tensor allocation and a bare except swallowing errors in FuyuBatchFeature are not the same defect." + "reason": "XLNet CPU tensor creation bug and FuyuBatchFeature bare-except issue are distinct code-path problems." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge/tied-weights collapse and incorrect visual-encoder save keys are separate serialization issues." + "reason": "Both involve Qwen serialization, but one is LoRA/tied-weights collapse after merge while the other is incorrect visual encoder keys in save_pretrained." }, { "left": "issue:43388", "right": "issue:43582", "accept": false, - "reason": "Metric-gather label truncation and an Apple Silicon warmup TypeError are unrelated." + "reason": "gather_for_metrics label truncation and AppleSilicon warmup TypeError are unrelated." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Python 3.14 import failure and missing TensorFlow backend from the torch extra are different environment/setup problems." + "reason": "Both are packaging/backend-related, but one is Python 3.14 import failure and the other is missing TensorFlow backend wiring for tensorboard." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and DebertaV2 import parsing on Python 3.13 do not share a concrete bug." + "reason": "SwitchTransformers sparse-layer config bug and DebertaV2 Python 3.13 import parsing error are unrelated." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer corruption and GPTNeoX rotary_pct reload behavior are different state/serialization bugs." + "reason": "Non-persistent buffer serialization junk and GPTNeoX rotary_pct reload default are different state/serialization bugs." }, { "left": "issue:44077", "right": "issue:44079", "accept": false, - "reason": "Both touch internals, but one is about post_init validation and the other about ModelOutput key assignment; not the same issue." + "reason": "Optional post_init in patchtsmixer and ModelOutput key assignment from None are separate model-output/config issues." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "A general Python-version compatibility problem is not the same as the DebertaV2 import error." + "reason": "A Python-version compatibility complaint and a Python 3.13 indentation/import parsing error are not the same bug." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "A torch.compile SDPA assert in Bamba is unrelated to a local dataset loading failure in an example script." + "reason": "Bamba torch.compile SDPA failure is unrelated to the local dataset loading issue." }, { "left": "issue:44496", "right": "issue:45803", "accept": false, - "reason": "Unrecognized model config and a swallowed-exception bug in FuyuBatchFeature are unrelated." + "reason": "Unrecognized model/config.json failure and Fuyu bare-except bug are unrelated." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Gemma4 embedding resize propagation and Whisper empty-transcription handling are different multimodal bugs." + "reason": "Gemma4 embedding resize propagation and Whisper empty transcription handling are different model behaviors." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "A stale GraniteSpeech CI test and removed Sam3LiteText auto-mapping references are separate CI/config issues." + "reason": "Both involve CI failures, but one is a stale GraniteSpeech test and the other is auto_mappings referencing removed configs." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are different codepaths and symptoms." + "reason": "Qwen3OmniMoe video input validation and serve KV-cache continuation indexing are different code paths." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "The vague 'ADD THE DATA' issue is not the same as the _torch_extract_fbank_features problem." + "reason": "The vague 'ADD THE DATA' issue and the fbank extraction bug are not clearly the same underlying problem." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "SwitchTransformers config behavior and example-script dataset loading are unrelated." + "reason": "SwitchTransformers config construction bug and local dataset loading error are unrelated." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Both are regressions, but one is processor/tokenizer API breakage and the other is Gemma4 embedding resizing; not mergeable as one fix." + "reason": "Processor/tokenizer regression and Gemma4 resize_token_embeddings propagation bug are different API failures." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Empty multi-label classification results and local dataset loading are unrelated example-script bugs." + "reason": "Multi-label classification empty results and local dataset loading are unrelated example-script issues." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "A category-count overflow and Whisper empty transcription after token alignment are unrelated." + "reason": "Category cardinality limit error and Whisper empty transcription are unrelated." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification image-processor lookup and a notebook syntax error are clearly different problems." + "reason": "Video-classification pipeline image-processor lookup and notebook syntax error are not the same bug." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different export paths." + "reason": "Both are import failures, but they concern different symbols and different surfaces: a missing Qwen2.5 VL image processor submodule export versus top-level AutoImageProcessor import." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Phi4Multimodal default-config initialization and Qwen3_5MoeVisionConfig field dropping are distinct config bugs." + "reason": "Phi4MultimodalConfig default initialization and Qwen3.5MoEVisionConfig missing deepstack_visual_indexes are distinct config-schema bugs." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Missing Qwen2_5_VLImageProcessor and missing layer_type_validation are different import regressions." + "reason": "Both are import errors, but one is a missing model class export and the other is a missing utility from configuration_utils; different regressions." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Missing TensorFlow in the torch extra is a packaging issue, while the other is only a vague dependency complaint." + "reason": "TensorFlow backend packaging issue and generic dependency issue are too broad and not the same defect." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "A notebook syntax error and an fbank feature extraction issue do not describe the same bug." + "reason": "Notebook syntax error and audio feature extraction bug are unrelated." }, { - "left": "issue:44186", - "right": "issue:44734", + "left": "issue:38175", + "right": "issue:43065", "accept": false, - "reason": "LayoutLMv2Tokenizer padding/truncation crashes and serve KV-cache continuation indexing are unrelated." + "reason": "Unexpected zero probabilities in SigLIP2 and dummy Conv2d in Sam3PixelDecoder are unrelated model-specific bugs." }, { - "left": "issue:38175", - "right": "issue:43065", + "left": "issue:44186", + "right": "issue:44734", "accept": false, - "reason": "Unexpected zero probabilities in Siglip2 and a dummy Conv2d in Sam3PixelDecoder are separate model-specific issues." + "reason": "LayoutLMv2 tokenizer crashes on NER padding/truncation and serve KV-cache continuation indexing are unrelated." }, { "left": "issue:44466", "right": "issue:44913", "accept": false, - "reason": "Device-dependent lm_head serialization and GPTNeoX rotary_pct reload behavior are different weight/config persistence bugs." + "reason": "lm_head.weight serialization inconsistency and GPTNeoX rotary_pct reload default are different serialization/config issues." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both mention flex_attention/versioning, but one is a runtime compile failure and the other is a version-check import bug; not the same concrete fix." + "reason": "Compiled flex_attention on torch>=2.9 and incorrect PyTorch version check for AuxRequest import are unrelated compatibility bugs." } ] }, @@ -9772,228 +9727,228 @@ "issue:43976|issue:44279", "issue:43901|issue:44496", "issue:43381|issue:45593", - "issue:44496|issue:45759", "issue:43299|issue:45357", + "issue:44496|issue:45759", "issue:44351|issue:45278", "issue:44521|issue:45325", "issue:41950|issue:45020", "issue:44242|issue:45200", "issue:43976|issue:45405" ], - "prepared_review_unit_hash": "e56c23ccad0808971a10b64fc104053e08def1b42f9ac05a467c73693153dfb2", + "prepared_review_unit_hash": "f1e8b44bf8fe077aff81c724efe46eae2e196dff1028673c5ea7086f8001d020", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36961, - "estimated_input_tokens": 9241, - "estimated_eval_tokens": 18738 + "serialized_chars": 36956, + "estimated_input_tokens": 9239, + "estimated_eval_tokens": 18734 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated Transformers bugs centered on model loading/import regressions, processor/config mismatches, and a few isolated utilities/docs issues. None of the soft pairs look like true duplicates; the closest is the broad remote_code regression bucket, but the concrete reports still differ too much to merge.", - "confidence": 0.76, - "canonical_issue_reason": "issue:45020 is the best umbrella representative because it captures the broadest recurring theme in the set: recent-version regressions that break remote_code / model-loading paths across multiple models.", + "summary": "This cluster is dominated by Transformers regressions around model loading, remote_code, multimodal processors, and version upgrades, but the soft pairs mostly describe different concrete failure modes. I would not merge any of the proposed soft edges.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45020 is the broadest representative because it captures the shared theme of recent-version regressions breaking remote_code/model-loading flows across multiple models, while most other issues are narrower symptoms or model-specific variants.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the strongest single issue for triage since it is broad, recent, and closest to the shared theme of version-related loading failures; the rest are narrower, model-specific variants.", + "best_issue_reason": "issue:45020 is the best cluster anchor for duplicate triage because it is the most general, user-facing description of the underlying regression class.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44734", "right": "issue:45632", "accept": false, - "reason": "Different subsystems and failure modes: KV-cache continuation tensor indexing vs trust_remote_code cache-path collision." + "reason": "Serve KV-cache indexing failure and trust_remote_code cache-path collision are unrelated code paths." }, { "left": "issue:44561", "right": "issue:45632", "accept": false, - "reason": "Both involve trust_remote_code, but one is an import/API break from removing is_torch_fx_available while the other is a local cache-path collision." + "reason": "Both mention trust_remote_code, but one is a missing import/helper regression and the other is a cache-directory collision." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both are save/tie-weight related, but one is a missing lm_head weight tie in Mistral3 and the other is incorrect checkpoint saving for Qwen3.5 visual encoder keys." + "reason": "Different models and different failure modes: tied weights vs incorrect save_pretrained checkpoint contents." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness and unsafe sys.modules access are not the same underlying bug." + "reason": "Import slowness and unsafe sys.modules access are related only at a very broad level, not the same bug." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Both are config-loading problems, but one is missing fields with trust_remote_code and the other is an unrecognized model missing model_type." + "reason": "One is missing fields from AutoConfig with trust_remote_code; the other is an unrecognized model/config validation error." }, { "left": "issue:45406", "right": "issue:45632", "accept": false, - "reason": "Gemma4Processor _tokenizer crash in serve is unrelated to trust_remote_code cache path collisions." + "reason": "Gemma4Processor missing _tokenizer in serve is not the same as a trust_remote_code cache collision." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Different model-specific load failures: Phi-4 multimodal loading vs stale auto_mappings entries for removed Sam3LiteText configs." + "reason": "A model-load regression and stale auto_mappings references are different defects." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI device-override test failure is unrelated to a model load failure for Phi-4 multimodal." + "reason": "CI device-override flakiness is not the same as a user-facing model-loading failure." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor/tokenizer image-handling regression is unrelated to t5gemma resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Processor/tokenizer regression and decoder embedding resize behavior are separate bugs." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both are model-loading failures, but they affect different models and likely different code paths." + "reason": "They concern different models with different loading failures." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Same model family name, but saving dequantized Ministrals/Devstrals and FlashAttention loading for Ministral-3 are distinct bugs." + "reason": "Saving dequantized models and FlashAttention loading errors are distinct issues." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken keyword collision are unrelated." + "reason": "Notebook syntax error and AddedToken argument conflict have no shared code path." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version compatibility and a specific model loading failure are different issues." + "reason": "Python-version compatibility and a specific model-loading regression are not duplicates." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Generation cache chunked_prefill regression is unrelated to ANSI codes emitted by loading_report." + "reason": "Cache missing in generation and ANSI-code emission in loading_report are unrelated." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and auto_mappings cleanup are different code paths." + "reason": "Qwen3-MoE sliding_window behavior and removed auto_mappings configs are different problems." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Unrecognized model/config error is unrelated to PIL image processors requiring torchvision." + "reason": "Model-type recognition failure and PIL processor torchvision dependency are separate loading issues." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but they concern different missing exports in different modules and model families." + "reason": "These are both import errors, but for different symbols and different root causes." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BEiT image processor label reduction bug is unrelated to a documentation mismatch in TextClassificationPipeline." + "reason": "A processor bug and stale docs about return_all_scores are not the same underlying change." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Different functionality: embedding resizing vs Whisper generation on empty transcription." + "reason": "Embedding resize and Whisper empty-transcription handling are unrelated." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both involve Qwen regressions, but one is loading a multimodal model and the other is save_pretrained checkpoint correctness." + "reason": "Loading Qwen3VL MoE models and saving an incorrect Qwen3.5 checkpoint are different regressions." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "Optimizer argument handling and a Qwen3.5 tie_word_embeddings warning are unrelated." + "reason": "Unused SGD args and a tie_word_embeddings warning are not the same defect." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "General dependency complaint is too vague and not the same as the Python-version breakage." + "reason": "Python support mismatch and a vague dependency issue are too different to merge." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Docs-only pipeline wording issue vs actual model-loading failure." + "reason": "Documentation mismatch and config/model recognition failure do not share a concrete code path." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and D-FINE aux-loss behavior are unrelated." + "reason": "Gradient checkpointing eval-mode behavior and D-FINE auxiliary loss logic are unrelated." }, { - "left": "issue:44496", - "right": "issue:45759", + "left": "issue:43299", + "right": "issue:45357", "accept": false, - "reason": "Unrecognized model config and AutoModelForCausalLM.from_config text_config unwrapping are different config-loading bugs." + "reason": "Both involve Qwen models, but one is loading and the other is save_pretrained key serialization." }, { - "left": "issue:43299", - "right": "issue:45357", + "left": "issue:44496", + "right": "issue:45759", "accept": false, - "reason": "Both are Qwen3.5-related, but loading failure and incorrect save_pretrained visual-encoder keys are separate regressions." + "reason": "Unrecognized model config and AutoModel text_config unwrapping are different configuration-path bugs." }, { "left": "issue:44351", "right": "issue:45278", "accept": false, - "reason": "A specific missing import and a broad report of many import errors are not the same concrete bug." + "reason": "A specific missing export and a broad collection of import errors are not the same concrete issue." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Multimodal assistant_mask bug and Qwen2.5-VL rope-index scaling bug are different generation/path problems." + "reason": "Assistant-mask generation and temporal position-id scaling are different multimodal bugs." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification pipeline processor lookup bug is unrelated to remote_code model-loading regressions." + "reason": "Video pipeline processor lookup and remote_code breakage are only loosely related subsystem issues." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss logic and Gemma 4 mm_token_type_ids defaults are different training bugs." + "reason": "MoE load-balancing loss and Gemma4 token-type defaults are separate training bugs." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python compatibility issue and PEFT minimum-version bump are unrelated." + "reason": "Python compatibility and an unreleased minimum dependency bump are unrelated release issues." } ] }, @@ -10085,8 +10040,8 @@ "issue:45600|issue:45712", "issue:43450|issue:44279", "issue:43441|issue:45834", - "issue:42994|issue:43317", "issue:44466|issue:45276", + "issue:42994|issue:43317", "issue:44261|issue:44263", "issue:43065|issue:44898", "issue:43441|issue:44683", @@ -10097,14 +10052,14 @@ "issue:44671|issue:44779", "issue:43723|issue:44162" ], - "prepared_review_unit_hash": "915d9b5f00b92aae10516092fb77e44d57207a5bcc435e467e30162480196e3c", + "prepared_review_unit_hash": "20967b1afdafc5984258ae8d4f7afb7a1e33cef20ea51d74bfa2c12f9a66cc3a", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -10119,180 +10074,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very heterogeneous: most items are unrelated bug reports, and none of the soft pairs look like the same underlying issue strongly enough to merge. The only loose concentration is around tokenizer/model-loading regressions in v5, but the specific failures differ too much.", - "confidence": 0.86, - "canonical_issue_reason": "No single canonical issue stands out; the set spans unrelated bugs, with only loose overlap around tokenizer/model-loading regressions.", + "summary": "These issues are mostly unrelated model-specific bugs, regressions, and docs/config problems. None of the soft pairs look like true duplicates, so I would reject all soft edges.", + "confidence": 0.83, + "canonical_issue_reason": "If forced to pick one representative issue, 44261 is the strongest standalone report: it is open, detailed, and describes a concrete regression around a missing config value rather than a vague symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43723 is the broadest representative of the tokenizer-loading regressions in this set, but it should not be treated as a true duplicate umbrella for the whole cluster.", + "best_issue_reason": "44261 is the best anchor for the set because it is active, well-scoped, and technically actionable; however, the cluster overall does not appear to be a duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Both are config/attribute loading problems, but one is about default multimodal sub-configs and the other is a missing pad_token_id attribute; different bugs." + "reason": "Different bugs: multimodal config defaults vs a missing pad_token_id attribute on a different config class." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "A video pipeline processor lookup bug and a Gemma4 device-mismatch forward bug are unrelated." + "reason": "Different code paths: video pipeline processor lookup vs Gemma4 forward device placement mismatch." }, { "left": "issue:38175", "right": "issue:44617", "accept": false, - "reason": "SigLIP2 zero probabilities and Sam3Video CUDA OOM are entirely different failure modes." + "reason": "Unrelated symptoms: zero probabilities in SigLIP2 vs CUDA OOM in Sam3Video." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and missing TensorFlow backend detection are different subsystems and bugs." + "reason": "Different subsystems: torch flex_attention compilation vs TensorFlow backend installation detection." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "A generic dependency issue and an unreleased PEFT version bump are not the same underlying problem." + "reason": "A vague dependency issue is not the same as a specific PEFT minimum-version bump." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype propagation and Qwen3.5 save_pretrained checkpoint corruption are distinct code paths." + "reason": "Different regressions: dtype propagation for BLIP2 vs incorrect checkpoint saving for Qwen3.5." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both mention docs, but one is a broken SigLIP2 example and the other is stale pipeline documentation." + "reason": "Both mention docs, but they are different documentation problems for different pipelines and behaviors." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Both are model-loading regressions in v5, but they affect different models and likely different loading code." + "reason": "Different models and failures: Qwen3VL MoE loading vs GLM-4.7-Flash test loading." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Processor/tokenizer API breakage with images is not the same as Deepseek tokenizer incorrect results." + "reason": "Different tokenizer regressions: processor.tokenizer/image handling vs Deepseek tokenization output correctness." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashes are separate tokenizer-loading failures." + "reason": "Loading tokenizer.json from repo is unrelated to the Mistral regex patch AttributeError." }, { "left": "issue:41950", "right": "issue:43650", "accept": false, - "reason": "A pipeline processor lookup issue and an unrelated placeholder issue are not duplicates." + "reason": "A video-classification pipeline bug is unrelated to the placeholder 'ADD THE DATA' issue." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Tokenizer class metadata saved incorrectly and ignoring repository tokenizer.json are related only at a high level, not the same bug." + "reason": "Both touch tokenizer serialization/loading, but one is save_pretrained metadata mismatch and the other is AutoTokenizer ignoring tokenizer.json." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Deepspeed/Adafactor OOM and deepspeed calling optimizer.step twice are different optimizer integration bugs." + "reason": "Different deepspeed optimizer problems: OOM with Adafactor vs optimizer.step being called twice." }, { "left": "issue:44462", "right": "issue:44779", "accept": false, - "reason": "Both concern tokenizers in v5, but one is about loading repo tokenizer.json and the other about Deepseek tokenization behavior." + "reason": "Different regressions in different components: repository tokenizer loading vs Deepseek tokenizer output changes." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "A notebook syntax error and an init_empty_weights TypeError are unrelated." + "reason": "A notebook syntax error is unrelated to the init_empty_weights TypeError regression." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both are CI/repo hygiene issues, but they involve different stale symbol sources and failure modes." + "reason": "Both are CI hygiene issues, but they target different files and failures: stale mappings vs leftover dummy classes." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Incorrect batched video processor shapes and a dependency problem do not match." + "reason": "Incorrect batched video processor shapes are unrelated to a generic dependency issue." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention failure for Ministral-3 and an OCR index error in Kosmos2.5 are different model-specific bugs." + "reason": "Ministral FlashAttention failure and Kosmos2.5 OCR index error are different model-specific bugs." }, { - "left": "issue:42994", - "right": "issue:43317", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "Quantized model saving and dequantized model loading with gpu+cpu offload are different lifecycle stages and failures." + "reason": "Different serialization and resizing bugs affecting different models and code paths." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:42994", + "right": "issue:43317", "accept": false, - "reason": "Serialization of lm_head.weight and resize_token_embeddings not propagating in Gemma4 are related to weights, but not the same concrete bug." + "reason": "Quantized model saving failure is not the same as dequantized model loading with gpu/cpu offload." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "Missing rms_norm_eps in MLA layernorm and torch.split return handling in GlmMoeDsaIndexer are unrelated." + "reason": "Completely different bugs: MLA layernorm epsilon precision issue vs torch.split return values in a MoE indexer." }, { "left": "issue:43065", "right": "issue:44898", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and Perceiver non-default resolution failure are different vision-model issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to Perceiver non-default-resolution image classification failure." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "FlashAttention support for a model and compiled flex_attention on newer torch versions are separate attention-stack bugs." + "reason": "Different attention-stack bugs: FlashAttention failure in a model vs compiled flex_attention failure on newer torch." }, { "left": "issue:43650", "right": "issue:45290", "accept": false, - "reason": "A placeholder issue and a chat-template crash on tool-call assistant messages are unrelated." + "reason": "The placeholder issue has no concrete overlap with chat template crashes on tool-call assistant messages." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss handling." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor save files and gemma-4 zero3 from_pretrained are different save/load problems." + "reason": "Missing processor.save_pretrained files and gemma-4 zero3 loading are different bugs." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing save_pretrained files and a performance slowdown after upgrading are not the same underlying issue." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked LM regression and Deepseek tokenizer regression are different model-specific failures." + "reason": "Different model regressions: CamemBERT masked LM predictions vs Deepseek tokenizer output correctness." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "Tokenizer loading in v5 and ESM2 being broken are not the same bug." + "reason": "Generic tokenizer loading regression is not the same as the broader ESM2 breakage report." } ] }, @@ -10367,8 +10328,8 @@ "soft_pairs": [ "issue:43784|issue:45542", "issue:43262|issue:44521", - "issue:43673|issue:43761", "issue:44016|issue:45823", + "issue:43673|issue:43761", "issue:44361|issue:45593", "issue:43901|issue:43994", "issue:42175|issue:43976", @@ -10376,9 +10337,9 @@ "issue:43122|issue:43650", "issue:43824|issue:43976", "issue:43525|issue:43881", - "issue:44188|issue:44336", "issue:44297|issue:45397", "issue:43525|issue:43643", + "issue:44188|issue:44336", "issue:43295|issue:44617", "issue:44589|issue:45362", "issue:42491|issue:43299", @@ -10395,20 +10356,20 @@ "issue:36246|issue:42757", "issue:36246|issue:44351" ], - "prepared_review_unit_hash": "bb55f7a1f1c3e3cd6a2751999d3156ae3c8bbda6b7bd860dd7965ef74dc65ed2", + "prepared_review_unit_hash": "56ad95dd2a41d945ffc8306506b6c5deefc943d380ff17b5aef42bcab27c49c8", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35507, + "serialized_chars": 35506, "estimated_input_tokens": 8877, "estimated_eval_tokens": 18010 }, @@ -10417,186 +10378,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs mostly share broad themes like imports, multimodal processing, model loading, or regressions, but they do not look like the same concrete bug/change. I would not deduplicate any of them based on the available titles and metadata.", - "confidence": 0.97, + "summary": "This cluster is a loose thematic bucket of Transformers regressions, import/config errors, and unrelated model-specific bugs rather than a true duplicate set. None of the soft pairs share the same concrete failure path, so they should stay separate.", + "confidence": 0.94, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43824 is the strongest representative if one must be chosen: it is a concrete import regression, has the highest visible discussion/inbound activity among the set, and is clearly scoped compared with the more ambiguous reports.", + "best_issue_reason": "If one issue must represent the cluster, 43299 is the closest fit because it is a concrete v5 model-loading regression and is thematically central to the set, but it is still not a duplicate anchor for the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Both are import/runtime errors, but one is a missing nn symbol in sentence-transformers integration and the other is a missing tf backend symbol; different components and failure modes." + "reason": "Different code paths: sentence-transformers import NameError vs missing TensorFlow backend from a torch-only install." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Both involve chat/template-style multimodal handling, but one is an audio sampling-rate default and the other is assistant-mask generation for multimodal inputs; different code paths." + "reason": "Audio chat-template sampling-rate default bug vs multimodal assistant_mask generation bug." }, { - "left": "issue:43673", - "right": "issue:43761", + "left": "issue:44016", + "right": "issue:45823", "accept": false, - "reason": "Chunked-prefill cache behavior in generation and CLIP vision hidden_states output are unrelated bugs." + "reason": "Notebook syntax error is unrelated to the Gemma4 device-map forward mismatch." }, { - "left": "issue:44016", - "right": "issue:45823", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "A notebook syntax error is not the same issue as a Gemma4 device_map forward-device mismatch." + "reason": "Generation cache/chunked prefill regression vs CLIPVisionModel hidden_states being dropped; different subsystems." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "Tokenizer task AttributeError and D-FINE auxiliary-loss handling are unrelated model-specific bugs." + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; no shared failure mechanism." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "One is a docs/behavior mismatch for return_all_scores; the other is incorrect model outputs for SigLIP2." + "reason": "Docs mismatch for return_all_scores vs bad model outputs for siglip2." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "TensorFlow optional-dependency/backend installation and Python-version compatibility are separate issues." + "reason": "Packaging/backend dependency problem vs Python version support regression." }, { "left": "issue:44987", "right": "issue:45542", "accept": false, - "reason": "Model loading failure for physical-intelligence/fast is unrelated to the TensorFlow/tensorboard backend problem." + "reason": "Model-loading failure for a specific checkpoint vs tf backend undefined due to missing TensorFlow." }, { "left": "issue:43122", "right": "issue:43650", "accept": false, - "reason": "Tokenizer output changes from 4.57.3 to 5.0 and an unrelated placeholder/data issue are not the same bug." + "reason": "Tokenizer behavior regression vs an unrelated placeholder/data issue." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "A Qwen2.5-VL import regression and a Python 3.9/3.10 compatibility problem are different failures." + "reason": "Specific Qwen2.5-VL import export failure vs Python 3.9+/3.10+ compatibility problem." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Both are model-loading failures, but for different model families with different missing attributes/errors." - }, - { - "left": "issue:44188", - "right": "issue:44336", - "accept": false, - "reason": "Torch.compile attention-kernel divergence and ANSI codes in loading_report are unrelated." + "reason": "Missing pad_token_id in Llama4Config vs glm-4v loading failure; different model/config bugs." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer metadata saved incorrectly and Gemma4 zero3 from_pretrained loading are unrelated issues." + "reason": "Tokenizer metadata mismatch on save_pretrained vs a zero3/from_pretrained loading issue." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Missing pad_token_id on Llama4Config and missing fields from trust_remote_code are not the same code-path problem." + "reason": "Missing config attribute vs trust_remote_code stripping fields; related broadly to configs, but not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile attention-kernel divergence vs ANSI codes leaking from loading_report." }, { "left": "issue:43295", "right": "issue:44617", "accept": false, - "reason": "Processor/tokenizer regression and CUDA out-of-memory are different classes of bug." + "reason": "Processor/tokenizer regression vs CUDA OOM; completely different symptoms and fixes." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage import failure and Qwen3.5 chat crashes do not point to the same underlying defect." + "reason": "Float8 storage import/type error vs Qwen3.5 chat crash; no shared code path." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3/MoE regressions, but one is LoRA compatibility across branches and the other is base Qwen3VL-MoE model loading; too different to merge." + "reason": "Both involve Qwen3-MoE/Qwen3VL loading, but one is cross-version LoRA compatibility and the other is a main-branch loading regression; not the same concrete bug." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs inconsistency for return_all_scores and intermittent import slowness are unrelated." + "reason": "Documentation issue vs import-time slowness; unrelated." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "AutoImageProcessor import failure and Python-version incompatibility are separate issues." + "reason": "AutoImageProcessor export issue vs Python version compatibility regression." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A pytest-xdist file race and a stale generation warning are unrelated." + "reason": "pytest-xdist race in testing utilities vs stale warning for num_return_sequences." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "ANSI output from loading_report and a wrong docstring for position_ids are not the same bug." + "reason": "ANSI escape codes in loading_report vs incorrect docstring for position_ids." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Both involve losses, but they affect different models and different loss mechanisms." + "reason": "Missing load-balancing loss vs missing auxiliary losses; both are loss-related but for different models and logic." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "A model not supporting Flash Attention 2.0 and a generic Flash Attention 2.0 issue are too broad to treat as one concrete bug." + "reason": "Unsupported Flash Attention 2 warning for one model vs a generic Flash Attention topic; not the same fix." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification processor lookup and audio fbank extraction are different subsystems." + "reason": "Video-classification pipeline using image processors vs a torch fbank extraction issue." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Reading position_ids after buffer registration and Gemma4 zero3 loading are unrelated." + "reason": "Buffer/read error for position_ids vs Gemma4 loading failure; different layers." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Intermittent import slowness and an unrecognized-model config error are different problems." + "reason": "Import slowness vs unrecognized model config; unrelated." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Both are Qwen-family reports, but one is a sliding_window bug in Qwen3-MoE and the other is a chat crash in Qwen3.5-35B; not the same failure." + "reason": "Qwen3-MoE sliding_window bug vs Qwen3.5 chat crash; different model behaviors." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Different import errors affecting different symbols and libraries." + "reason": "Qwen2.5-VL image processor import error vs missing is_offline_mode export from huggingface_hub." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different classes (Qwen2.5-VL image processor vs HybridCache) and different code paths." + "reason": "Qwen2.5-VL image processor import error vs HybridCache import error; different missing symbols." } ] }, @@ -10701,20 +10662,20 @@ "issue:44368|issue:44977", "issue:44368|issue:45563" ], - "prepared_review_unit_hash": "30b32888d201e1ec2ff75800c4918e4904fbe8792ae4dac9056a4dc91252acb9", + "prepared_review_unit_hash": "bd564781241a6646892dac37a2f53ac9608df3867fe9f42aa9da9a4bebbd334e", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36905, + "serialized_chars": 36908, "estimated_input_tokens": 9227, "estimated_eval_tokens": 18710 }, @@ -10723,192 +10684,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most soft pairs are false positives: they share broad themes like configs, embeddings, Deepspeed, or tokenizers, but describe different code paths and fixes. The only plausible duplicate is the generic `resize_token_embeddings` failure and its t5gemma-specific manifestation.", - "confidence": 0.92, - "canonical_issue_reason": "issue:45292 is the broader, more general report of `resize_token_embeddings` not updating output embeddings, while issue:45335 looks like a model-specific instance of the same underlying bug.", + "summary": "This cluster is mostly a set of unrelated Transformers bug reports. The only soft duplicate-looking pair is the `resize_token_embeddings` regression across model variants; the rest are different code paths or different model families.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 45292 is the broadest formulation of the `resize_token_embeddings` bug, with 45335 looking like a model-specific instance of the same failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best representative because it describes the generic bug in the shared resizing path, not just one affected model.", + "best_issue_reason": "45292 is the best representative because it states the generic API break, while 45335 narrows it to t5gemma\u2019s decoder embeddings.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "Different subsystems and failures: BEiT label reduction vs unsafe `sys.modules` access." + "reason": "Different subsystems: image processor label reduction vs unsafe `sys.modules` access in modeling utils." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "Unrelated problems: ANSI coloring in loading reports vs GLM-5 RoPE implementation." + "reason": "Unrelated bugs: terminal ANSI output vs GLM-5 RoPE implementation." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Different bugs: module lookup crash vs pytest-xdist file race." + "reason": "Generic import/cache behavior and xdist file race are different code paths and failure modes." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Both affect v5 compatibility, but one is cache handling during chunked prefill and the other is `is_torch_fx_available` removal for `trust_remote_code` models." + "reason": "One is generation cache handling; the other is removal of a helper breaking `trust_remote_code` models." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Performance/import timing issue versus Gemma-3 EOS token config inconsistency." + "reason": "Import-time slowdown is unrelated to Gemma eos token config mismatch." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Different code paths: router load-balancing loss logic vs an AMP/CUDA crash in image loss." + "reason": "MoE load-balancing loss and AMP/CUDA image-loss crash are different model/components." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Import error for a missing Qwen2.5-VL class is unrelated to the Deepspeed/BertModel indexing crash." + "reason": "Qwen2.5 VL import error is unrelated to DeepSpeed ZeRO3/BertModel index failure." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma4 `mm_token_type_ids` are separate model-specific issues." + "reason": "ESM2 breakage and Gemma 4 token-type defaulting are different model bugs." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer correctness regression is not the same bug as `resize_token_embeddings` failing to update embeddings." + "reason": "Tokenizer behavior regression is not the same as resizing token embeddings not updating outputs." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Position-ids buffer access error is unrelated to the reported slowdown in data filtering after upgrading Transformers." + "reason": "Position ID buffer bug and v5 filtering slowdown do not share the same underlying cause." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both involve config handling, but one is strict loading failure for granite_speech while the other is a missing field silently dropped in Qwen3_5MoeVisionConfig." + "reason": "Both are config-related, but one is strict config loading and the other is a missing field in Qwen vision config." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize bug is unrelated to Whisper failing on empty transcription after special-token alignment." + "reason": "`resize_token_embeddings` and Whisper empty-transcription failure after token alignment are different regressions." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "One is a concrete embedding resize bug; the other is a broad tokenization regression report." + "reason": "Embedding resize bug is unrelated to tokenizer behavior changing across versions." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Pipeline docs mismatch has no relation to unsafe `sys.modules` access." + "reason": "Docs mismatch for pipeline output is unrelated to unsafe module lookup." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 CUDA OOM are unrelated." + "reason": "Offline cache loading failure is not the same as Gemma4 multi-GPU OOM." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model loading failure and protobuf decode-error masking are different tokenizer/load paths." + "reason": "Physical-intelligence model load failure differs from tokenizer decode-error masking." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Missing config field under strict config and `from_config` not unwrapping `text_config` are different composite-model config bugs." + "reason": "Missing config field in Qwen vision config is unrelated to `from_config` text_config unwrapping." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model/config issue is unrelated to the `sys.modules` access problem." + "reason": "Unrecognized model config and unsafe `sys.modules` access are different issues." }, { "left": "issue:43232", "right": "issue:44336", "accept": false, - "reason": "Generation kwargs update after `sync_gpus` is unrelated to ANSI output in loading reports." + "reason": "Generation kwargs update bug is unrelated to ANSI rendering in loading_report." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Both mention Deepspeed, but MOE weight conversion and ZeRO3 deque underflow are different failure modes." + "reason": "Qwen3 MOE weight conversion and DeepSpeed deque underflow are different problems." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "SAM3 pixel decoder dummy conv and sam3_video/sam3_tracker type warning are not the same bug." + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as sam3_video/sam3_tracker config mismatch." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sampling-rate default bug is unrelated to import latency." + "reason": "Audio processor sampling-rate defaulting is unrelated to import-time slowdown." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "Stale indexer cache crash is unrelated to Gemma4 missing default `mm_token_type_ids`." + "reason": "Stale indexer cache crash in GlmMoeDsa is unrelated to Gemma 4 mm token type defaults." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Incorrect tied-weights warning in Qwen3-Next is unrelated to the `patchtsmixer` post_init policy issue." + "reason": "Incorrect tied-weights warning and patchtsmixer post_init policy are different concerns." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Different multimodal model failures: missing `fpn_position_embeddings` vs video input processing error." + "reason": "SAM 3 Video missing attribute and Qwen3OmniMoe video input validation are different bugs." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Serialization inconsistency for tied weights is a different issue from `resize_token_embeddings` not updating decoder embeddings." + "reason": "Both mention tied/resized embeddings, but 44466 is a serialization inconsistency while 45335 is a resize propagation bug." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Stale indexer cache and flash-attention `s_aux=None` crash are unrelated." + "reason": "Second-forward stale cache crash and flash-attention `s_aux=None` crash are unrelated." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "Tokenizer special-token registration bug is unrelated to ESM2 being broken." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage do not share a code path." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Warnings about tied embeddings in Qwen3.5 fine-tuning are not the same as a flash-attention generation failure." + "reason": "A warning about `tie_word_embeddings` is not the same as a flash-attention generation failure." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "A stale warning in paged generation is unrelated to the Qwen3.5 tied-embeddings warning." + "reason": "Different generation warnings: tied-embedding config vs stale `num_return_sequences` warning." } ] }, @@ -11010,238 +10971,238 @@ "issue:43475|issue:44661", "issue:42673|issue:43704", "issue:45310|issue:45375", - "issue:43232|issue:44188", "issue:43873|issue:44945", + "issue:43232|issue:44188", "issue:41628|issue:44351", "issue:42831|issue:44222", "issue:44945|issue:45753" ], - "prepared_review_unit_hash": "ca4361975af60932dd98d84ccd4d0f07a6b813842f9668921cdd3b9c616ecbc1", + "prepared_review_unit_hash": "c8c7d5772c351d22eb551242e8f4f0efdd78bc4932154692a6b179267afcd3fa", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37329, - "estimated_input_tokens": 9333, - "estimated_eval_tokens": 18922 + "serialized_chars": 37327, + "estimated_input_tokens": 9332, + "estimated_eval_tokens": 18920 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated Transformers regressions/import/import-like breakages, but there are at least two duplicate families visible from titles alone: the Qwen3.5-Moe from_pretrained/config-field reports and the Qwen3ForCausalLM VRAM-leak reports (42673/43704). The rest are largely distinct model- or subsystem-specific bugs.", - "confidence": 0.76, - "canonical_issue_reason": "issue:45278 is the broadest umbrella representative for this cluster because it captures the generic post-upgrade breakage theme, even though the cluster contains multiple separate subfamilies.", + "summary": "Mostly heterogeneous, but there are two clear duplicate-style pockets: the repeated Qwen3ForCausalLM VRAM leak reports, and the repeated Qwen3.5 MoE from_pretrained error reports. The rest span unrelated import, generation, tokenizer, save/load, and model-specific regressions.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella item: it captures the general Transformers v5 import-regression theme better than the more model-specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best single representative of the overall cluster since it is the most general and least model-specific item.", + "best_issue_reason": "Issue 45278 is the most reusable representative for this cluster because it reflects the broad upgrade/import breakage pattern rather than a narrow model-specific failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Different failures: gemma-4 ZeRO3 from_pretrained vs deepspeed optimizer stepping twice." + "reason": "Different bugs: Zero3 from_pretrained failure vs deepspeed optimizer-step-count issue." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "Unrelated bugs: ZeRO3 deque pop vs auto_mappings referencing removed Sam3LiteText configs." + "reason": "DeepSpeed ZeRO3 deque error and auto_mappings CI breakage are unrelated failure modes." }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both involve generation/cache, but the concrete code paths and symptoms differ." + "reason": "Both involve generation internals, but the code paths differ: chunked_prefill cache handling vs /v1/responses tensor indexing." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Different model-specific issues: Qwen3-VL weight shape mismatch vs a Qwen3.5 fine-tuning warning." + "reason": "Qwen3-VL weight-shape mismatch is not the same as a Qwen3.5 LoRA warning about tie_word_embeddings." }, { "left": "issue:44933", "right": "issue:45459", "accept": false, - "reason": "Both are import/lookup related, but one is a missing image_utils import and the other is tokenizer error masking when protobuf is absent." + "reason": "Missing image_utils import and tokenizer error masking without protobuf are separate problems." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Different tokenizer bugs: Whisper word-timestamp crash vs Kimi-K2.5 codec/fix_mistral_regex regression." + "reason": "Whisper timestamp decode crash and Kimi tokenizer codec regression are different model/tokenizer bugs." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both touch generation internals, but the sync_gpus/update_model_kwargs path differs from chunked_prefill cache handling." + "reason": "Both touch generation state, but one is sync_gpus kwargs update and the other is chunked_prefill cache missing." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Separate save_pretrained problems: tokenizer metadata mismatch vs processor file omission." + "reason": "Tokenizer save metadata mismatch and processor.save_pretrained missing files are distinct serialization bugs." }, { "left": "issue:43673", "right": "issue:45278", "accept": false, - "reason": "Generation cache behavior is unrelated to the broad import-error regression described by 45278." + "reason": "Generation cache handling is unrelated to the broad import-error regression report." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Different breakages: custom model initialization vs ESM2-specific regression." + "reason": "Custom model initialization breakage is not the same as the ESM2 workflow regression." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "No shared bug: MLukeTokenizer task AttributeError vs a slowdown regression after upgrade." + "reason": "Tokenizer AttributeError and performance slowdown after upgrade are different issues." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Processor save_pretrained missing files is unrelated to the deepspeed double-step optimizer bug." + "reason": "Processor save_pretrained file omission does not match a deepspeed double-step optimizer bug." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Both involve quantized models, but saving failure and GGUF loading failure are different code paths and symptoms." + "reason": "Quantized model saving and GGUF loading failure are opposite directions and different bugs." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "ImportError due to missing PyTorch is not the same as Sentence Transformers hanging during load." + "reason": "Missing PyTorch import error is not the same as a loading hang in Sentence Transformers." }, { "left": "issue:44749", "right": "issue:45310", "accept": false, - "reason": "A slowdown regression and a Qwen3.5-Moe from_pretrained error are not the same bug." + "reason": "Upgrade slowdown and Qwen3.5 Moe from_pretrained failure are not the same code-path problem." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Both mention missing fields, but one is a SAM output attribute and the other is AutoConfig returning incomplete remote-code config data." + "reason": "SAM3 output attribute missing and AutoConfig trust_remote_code missing fields are unrelated." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 API guidance is unrelated to InternVL2 meta-initialization incompatibility." + "reason": "TF32 API-warning/deprecation and remote-code meta-init incompatibility are different topics." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "Different model-specific vision bugs: SAM 3 Video missing output field vs qwen2.5-vl wrong vision_position_ids." + "reason": "Both are vision/model regressions, but the specific missing attribute and wrong position_ids bugs are distinct." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but they refer to different missing symbols and different modules." + "reason": "Both are import errors, but they involve different missing symbols and different causes." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report vs a specific gemma-4 ZeRO3 loading regression." + "reason": "Generic bug report provides no evidence of the same underlying failure as the Gemma-4 Zero3 issue." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe Transformers v5 issue is unrelated to processor save_pretrained file loss." + "reason": "Qwen3.5-Moe transformers-v5 incompatibility and processor save_pretrained file loss are separate regressions." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Different AttributeErrors in different subsystems: Mistral tokenizer regex patch vs auto_docstring kwargs processing." + "reason": "Tokenizer regex crash and auto_docstring annotation crash are unrelated." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized model saving vs Llama4Config missing pad_token_id are different defects." + "reason": "Quantized save failure is unrelated to Llama4Config missing pad_token_id." }, { "left": "issue:45375", "right": "issue:45478", - "accept": true, - "reason": "Same Qwen3.5-Moe regression family: one report identifies the missing deepstack_visual_indexes field, the other reports the resulting from_pretrained failure." + "accept": false, + "reason": "Related Qwen3.5-Moe family, but one is a missing config field dropped by strict and the other is a from_pretrained failure; not the same concrete bug." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "SAM 3 Video output attribute issue is unrelated to add-new-model-like failing in TOKENIZER_MAPPING_NAMES." + "reason": "SAM3 vision output attribute missing and add-new-model-like tokenizer mapping failure are different problems." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact same title and same Qwen3ForCausalLM VRAM-leak symptom; these are the same underlying bug." + "reason": "Same title and same VRAM-leak symptom for Qwen3ForCausalLM in multiple dataloader threads; this looks like the same underlying bug." }, { "left": "issue:45310", "right": "issue:45375", - "accept": true, - "reason": "Same underlying Qwen3.5-Moe from_pretrained/config regression: 45310 reports the failure, 45375 points to the missing config field." - }, - { - "left": "issue:43232", - "right": "issue:44188", "accept": false, - "reason": "Both involve generation/runtime behavior, but sync_gpus update logic and compile-time attention-kernel divergence are distinct bugs." + "reason": "Both are Qwen3.5-Moe-related, but one is a from_pretrained error and the other is a missing vision config field; too different to merge." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading issues are unrelated to pipeline-parallelism output errors." + "reason": "Quantization/offloading behavior and pipeline-parallel output corruption are different bugs." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs syncing and divergent attention kernels under torch.compile are unrelated." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import failures, but AutoImageProcessor and HybridCache are different symbols with different causes." + "reason": "AutoImageProcessor import failure and HybridCache import failure are separate export issues." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FP8 accuracy regression and FP8 save_pretrained bug are different problems." + "reason": "FP8 accuracy regression and FP8 save_pretrained failure are distinct problems." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel output correctness issue is unrelated to the Qwen3_5 infinite-loop report." + "reason": "Pipeline-parallel output mismatch and image-specific infinite loop are not the same bug." } ] }, "evaluator_result": { "accept": false, - "feedback": "The cluster-level summary is mostly grounded, but it overstates one point: saying the only clear duplicate family is the Qwen3.5-Moe from_pretrained/config-field regression is not supported by the packet, since issue:42673 and issue:43704 have the exact same title about the Qwen3ForCausalLM VRAM leak and also look like a duplicate family. That makes the summary insufficiently precise. The rest of the soft-edge rejections are generally conservative, but this factual overclaim is enough to reject." + "feedback": "Mostly conservative, but the summary overstates heterogeneity by saying the only clear duplicate-style pair is Qwen3ForCausalLM VRAM leak. The packet also contains another obvious duplicate-style pair with the same title, issue:45310 and issue:45478, both about the Qwen3.5 MoE from_pretrained error. That omission makes the summary a bit too strong. The pairwise rejections themselves look generally conservative, and the 42673/43704 acceptance is plausible." }, "evaluator_used": true, "retried": true, @@ -11315,9 +11276,9 @@ "issue:42222|issue:45200", "issue:44222|issue:45310", "issue:42994|issue:43824", + "issue:36010|issue:41628", "issue:43232|issue:44792", "issue:42175|issue:45712", - "issue:36010|issue:41628", "issue:43377|issue:44242", "issue:43673|issue:44188", "issue:44038|issue:45310", @@ -11339,225 +11300,225 @@ "issue:44734|issue:45103", "issue:30333|issue:33290", "issue:43723|issue:44568", - "issue:44060|issue:45125", "issue:42898|issue:45356", + "issue:44060|issue:45125", "issue:42898|issue:44779", "issue:43874|issue:45092" ], - "prepared_review_unit_hash": "487a4e8ec20d7c8ca124bf34c557458b6806a427a3a6a40d76ee35bfc0738cd8", + "prepared_review_unit_hash": "c60c4989e89e544eb8fd3c3a564fe251961f1637a593ab92a32bc0dbf718e6b9", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37050, - "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": "This cluster is mostly a loose collection of unrelated Transformers issues with a few superficial v5/tokenizer/model-loading similarities. I did not find any soft pair that looks like the same underlying bug.", - "confidence": 0.31, - "canonical_issue_reason": "issue:42898 is the broadest user-facing regression here and best captures the tokenization/v5 behavior-change theme, though the cluster as a whole is heterogeneous.", + "summary": "The cluster is highly heterogeneous: a few items are tokenizer/model-loading regressions, but most are unrelated bugs across training, multimodal, generation, packaging, and test infrastructure. I would not collapse any soft pair as a duplicate, and there is no safe single canonical issue for the whole cluster.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:42898 is the most representative issue to anchor the cluster because it is a clearly scoped, high-signal v5 tokenizer regression; the rest are largely separate bugs.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "Different model families and failure modes; vitpose breakage is not the same as Gemma 4 multimodal token defaults." + "reason": "Different failures: vitpose model breakage vs Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Same broad Qwen/MoE area, but one is FP8 save_pretrained and the other is from_pretrained construction; different code paths." + "reason": "Both mention MoE/FP8, but one is save_pretrained and the other is from_pretrained loading; different code paths." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving failure vs import error for a model class; unrelated bugs." + "reason": "Quantized model saving and a missing Qwen2.5-VL import are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different missing symbols from different modules: GenerationMixin vs AutoImageProcessor." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwargs/sync_gpus issue is unrelated to the Janus image-generation test failure." + "reason": "Generation cache/sync_gpus bug vs a janus image-generation test failure; not the same defect." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/backend-extra availability is unrelated to leftover dummy classes leaking into dir()." - }, - { - "left": "issue:36010", - "right": "issue:41628", - "accept": false, - "reason": "Both are import errors, but for different symbols and different missing-export problems." + "reason": "Packaging/backend dependency issue vs leftover dummy classes in repo checks." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI padding-mask correctness is unrelated to MoE load-balancing loss being omitted." + "reason": "MIMI padding-mask batching bug vs router-logits load-balancing loss omission." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Chunked prefill cache issue and torch.compile attention-kernel branching are different defects." + "reason": "Both involve generation behavior, but one is cache/chunked_prefill and the other is attention kernel divergence under torch.compile." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both touch Qwen3-VL/Qwen3.5-MoE compatibility, but they are different model/version regressions and different symptoms." + "reason": "Different Qwen-related model compatibility errors; not the same concrete bug." }, { "left": "issue:43526", "right": "issue:45561", "accept": false, - "reason": "BEiT label reduction bug is unrelated to the xdist captured-info race." + "reason": "BeitImageProcessorFast label reduction bug vs pytest-xdist file race." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Unrecognized model/config loading failure is unrelated to the test infrastructure race." + "reason": "Unrecognized model/config loading vs test utility file contention." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque pop are unrelated." + "reason": "Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque error are unrelated." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Legacy-field BEiT loading bug is unrelated to the xdist file race." + "reason": "Legacy BEiT preset loading bug vs pytest-xdist race; no common code path." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they affect different behaviors and different model/tokenizer code paths." + "reason": "Tokenization cleanup behavior change vs mdeberta BOS/EOS token insertion; related area but different bug." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Backend install extras are unrelated to a missing import from image_utils." + "reason": "TensorFlow backend packaging vs missing image_utils import are distinct issues." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Default config initialization bug is unrelated to Whisper timestamp decoding crash." + "reason": "Default multimodal config initialization vs Whisper word timestamp decode crash." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Loading hang and post-init embedding reinitialization are different problems." + "reason": "Sentence-Transformers loading hang vs embedding reinit on resize; unrelated." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version compatibility issue is unrelated to the missing image_utils import." + "reason": "Python version compatibility issue vs missing import from image_utils." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Config-default handling and ModelOutput key assignment are separate bugs." + "reason": "Config-default bug vs ModelOutput key assignment bug; different objects and code paths." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are v5 tokenizer regressions, but one is a cleanup-space behavior change and the other is wrong Granite token IDs; not the same bug." + "reason": "Different tokenizer regressions with different models and symptoms; not one underlying bug." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen2.5-VL import error are unrelated import-path failures." + "reason": "Sentence-transformers import NameError vs Qwen2.5-VL import error; different missing imports." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Attention-kernel divergence under compile is unrelated to a docstring typo." + "reason": "Attention-kernel divergence vs docstring typo; not a duplicate." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Embedding dtype drift during eval is unrelated to the ZeRO3 deque error." + "reason": "Embedding dtype drift in evaluation vs DeepSpeed ZeRO3 deque error are unrelated." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing processor save files and Qwen3.5-MoE from_pretrained error are different save/load paths with different causes." + "reason": "processor.save_pretrained file omission vs Qwen3.5 Moe from_pretrained failure; different fixes." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache continuation crash is unrelated to the auto_docstring annotations AttributeError." + "reason": "KV-cache continuation crash in serve vs auto_docstring annotations crash; separate subsystems." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLFlow failure reporting and DeepSpeed OOM are unrelated training issues." + "reason": "MLflow job status reporting vs DeepSpeed/Adafactor OOM; unrelated training issues." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading failure in v5 and BOS/EOS not added for mdeberta-v3 are different tokenizer bugs." + "reason": "Tokenizer loading/import problem vs BOS/EOS token behavior change; different failure modes." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is cleanup-space behavior and the other is Kimi-K2.5 codec/regex handling." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are unrelated architecture/setup issues." - }, - { - "left": "issue:42898", - "right": "issue:45356", - "accept": false, - "reason": "Both are tokenizer regressions, but the Kimi-K2.5 codec/warning issue is distinct from clean_up_tokenization_spaces behavior." + "reason": "Incorrect tied-weights warning vs missing _tp_plan for tensor parallelism; different model-internal bugs." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Deepseek tokenizer output regression is a different tokenizer bug from the clean_up_tokenization_spaces change." + "reason": "Different tokenizer regressions: cleanup-space behavior vs DeepSeek token IDs." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "GLM46V missing image-patch helper and InternVL2 meta-init incompatibility are different multimodal model issues." + "reason": "Missing image patch-count method vs old InternVL2 meta-init incompatibility; not the same defect." } ] }, @@ -11668,250 +11629,250 @@ "issue:43479|issue:44913", "issue:42831|issue:45310" ], - "prepared_review_unit_hash": "35ce6dfd724ee4a765ee0b0fa24405d8738c0997988d751bcbaa1167017ef818", + "prepared_review_unit_hash": "8bed832a48df8e6be74bb8719648cf2095381fb441721627d94b3c0992eedb61", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38264, - "estimated_input_tokens": 9566, - "estimated_eval_tokens": 19388 + "serialized_chars": 38260, + "estimated_input_tokens": 9565, + "estimated_eval_tokens": 19386 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous and does not contain a safe duplicate group; the soft pairs mostly share broad themes like imports, tokenizer loading, save_pretrained, or v5 regressions, but not the same underlying bug.", - "confidence": 0.89, - "canonical_issue_reason": "No safe canonical issue: the items span unrelated model-, tokenizer-, config-, save_pretrained-, generation-, and training-path bugs.", + "summary": "The cluster is highly heterogeneous: it mixes unrelated import errors, tokenizer/model regressions, serialization bugs, training issues, and docs-style complaints. No PRs are present, and none of the soft pairs look like the same concrete underlying bug/change.", + "confidence": 0.73, + "canonical_issue_reason": "issue:44297 is the most self-contained, clearly reproducible bug in the set, with a narrow serialization mismatch (`tokenizer_class` in `tokenizer_config.json`) and a straightforward fix path.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the closest representative because it is a concrete v5 tokenizer-loading regression, but it still does not cover the rest of the mixed cluster.", + "best_issue_reason": "issue:44297 is the strongest representative issue because it is concrete, isolated, and easy to validate compared with the broader regression or discussion-style reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "Unrelated: one is a PEFT version pin/release issue, the other is leftover dummy-class cleanup and torch-less exports." + "reason": "Unrelated problems: a version pin to unreleased PEFT versus leftover dummy classes leaking into `dir(transformers)`." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Both mention v5 output changes, but one is tokenization cleanup behavior and the other is a model-specific generation discrepancy." + "reason": "Both mention v5-era behavior changes, but one is token cleanup semantics and the other is full model output drift; not the same concrete bug." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Both are import/export problems, but they concern different symbols and different root causes." + "reason": "Image-utils import/export issue versus dummy PT object cleanup; different code paths and failure modes." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Different code paths: position_ids buffer access vs deepspeed causing a double optimizer step." + "reason": "Registering/reading `position_ids` as a buffer is unrelated to a deepspeed double optimizer step." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "Both involve save_pretrained, but one is MoE FP8 saving and the other is tokenizer metadata mismatch." + "reason": "FP8 `save_pretrained` behavior for MoE models is a different serialization bug than tokenizer config mismatch." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save metadata bug is unrelated to optimizer stepping with deepspeed." + "reason": "Tokenizer save/load metadata mismatch is unrelated to optimizer stepping under deepspeed." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Both are initialization/config-related, but one is Phi4Multimodal defaults and the other is non-persistent buffer initialization." + "reason": "Different configuration bugs: default vision/audio config initialization vs non-persistent buffer junk values." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Different areas entirely: feature-extractor regression vs auto-docstring crash with future annotations." + "reason": "Unsupported `center` arg in `LasrFeatureExtractor` is unrelated to the auto-docstring annotations crash." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Both mention save_pretrained, but processor file saving and MoE FP8 saving are not the same bug." + "reason": "MoE FP8 save bug and processor save_pretrained missing files are both save-related, but not the same artifact or code path." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "Accuracy regression in FineGrainedFP8 is unrelated to processor.save_pretrained missing files." + "reason": "FP8 accuracy regression and processor save_pretrained file omission are distinct issues." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "Both touch training, but MLflow reporting on failure and compute_loss signature mismatch are different failures." + "reason": "MLflow reporting on failed jobs is unrelated to `CustomTrainer.compute_loss` signature breakage." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Both are import errors, but for different missing modules and different causes." + "reason": "Import errors, but one is missing `transformers.onnx` and the other is missing PyTorch; different root causes." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different packages; no shared underlying defect." + "reason": "Missing `AutoImageProcessor` export is unrelated to `huggingface_hub.is_offline_mode` import failure." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLflow failure reporting is unrelated to a model-loading regression for Llama 3.2 vision." + "reason": "Training-job MLflow reporting and Llama 3.2 model loading regression do not share the same underlying bug." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Both involve generation-related code, but the reported failures are different and model-specific." + "reason": "Generation kwargs update bug and Kosmos2.5 OCR index error are unrelated." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading in v5 is not the same as Qwen3.5 flash-attention generation failure." + "reason": "Tokenizer loading in v5 and Qwen3.5 flash-attention generation problems are different code paths and symptoms." }, { "left": "issue:28282", "right": "issue:30333", "accept": false, - "reason": "Different symptoms and subsystems: PyTorch import missing vs MLflow tracking after failed training." + "reason": "PyTorch-not-found import error versus MLflow job state bug; unrelated." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Training-job reporting and embedding resize/reinit are unrelated defects." + "reason": "MLflow reporting issue and embedding resize/post_init reinitialization bug are different subsystems." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "save_pretrained FP8 MoE bug is unrelated to the max_shard_size default question." + "reason": "FP8 save_pretrained bug is unrelated to the `max_shard_size` default question." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer config save mismatch is not the same as Qwen3.5 MoE from_pretrained loading failure." + "reason": "Tokenizer config mismatch is not the same as Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "Both involve model internals, but the DSA indexer split issue is unrelated to InternVL2 meta-init compatibility." + "reason": "`torch.split()` return-value handling in GlmMoeDsaIndexer is unrelated to InternVL2 meta-initialization compatibility." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Same broad DSA area, but the torch.split return-values bug and missing ReLU are different code changes." + "reason": "Same DSA/indexer area, but one reports split return values and the other a missing ReLU; not the same bug." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "Different serving failures: chat/completions request handling vs Gemma4Processor missing _tokenizer." + "reason": "Chat completion endpoint rejection and Gemma4Processor missing `_tokenizer` are different failures." }, { "left": "issue:44514", "right": "issue:45103", "accept": false, - "reason": "Qwen2_5_VL batched chat-template crash is unrelated to the auto_docstring annotations crash." + "reason": "Batched `apply_chat_template` crash and auto-docstring annotation crash are unrelated." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Wrong docstring for position_ids is documentation-only; GLM-5 RoPE is an implementation bug." + "reason": "Docstring mismatch for `position_ids` is not the same as GLM-5 RoPE implementation discussion." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Buffer registration/read error is unrelated to processor.save_pretrained missing output files." + "reason": "Buffer registration/read error and processor save_pretrained missing files are separate serialization/runtime issues." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer load regression and eos_token_id inconsistency are separate problems." + "reason": "Tokenizer loading failure in v5 is unrelated to Gemma-3 `eos_token_id` inconsistency." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "MLA precision/config discussion is unrelated to InternVL2 meta-initialization compatibility." + "reason": "RMS norm epsilon precision discussion and InternVL2 meta-init incompatibility are different model bugs." }, { "left": "issue:43723", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading complaints, but for different models and likely different root causes." + "reason": "Both are tokenizer-loading complaints, but one is generic v5 loading and the other is a specific model; no evidence they share the same concrete cause." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Old remote-code meta-init incompatibility is distinct from saving incorrect visual encoder keys." + "reason": "Old InternVL2 meta-init incompatibility is distinct from Qwen3.5 visual encoder key saving regression." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence-transformers loading stall and SigLIP2 zero probabilities are not the same bug." + "reason": "Sentence Transformers loading stuck and SigLIP2 zero probabilities are different model/runtime problems." }, { "left": "issue:43723", "right": "issue:44442", "accept": false, - "reason": "AutoTokenizer loading failure for one model is unrelated to FastSpeech2ConformerTokenizer loading failure for another model." + "reason": "Generic tokenizer loading regression is unrelated to FastSpeech2ConformerTokenizer AutoTokenizer failure." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import-name errors, but they target different exported symbols with different fixes." + "reason": "Different import/export problems: missing `AutoImageProcessor` versus missing `Qwen2_5_VLForConditionalGeneration`." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Qwen3 offset_mapping correctness is unrelated to mdeberta add_special_tokens behavior." + "reason": "Qwen3 offset mapping bug and mDeBERTa special-token behavior regression are unrelated tokenizer issues." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Phi4Multimodal default subconfigs and GPTNeoX rotary_pct reload behavior are separate configuration bugs." + "reason": "Config initialization for multimodal defaults is unrelated to GPTNeoX `rotary_pct` persistence on reload." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy regression is unrelated to Qwen3.5 MoE from_pretrained loading failure." + "reason": "FP8 accuracy regression and Qwen3.5 MoE from_pretrained failure are different problem types." } ] }, @@ -11986,8 +11947,8 @@ ], "soft_pairs": [ "issue:42886|issue:42994", - "issue:44263|issue:44373", "issue:42757|issue:45278", + "issue:44263|issue:44373", "issue:43756|issue:44060", "issue:42831|issue:44038", "issue:43232|issue:45593", @@ -12010,21 +11971,21 @@ "issue:43122|issue:44514", "issue:42907|issue:43937", "issue:43408|issue:43475", - "issue:44261|issue:44373", "issue:44360|issue:44485", + "issue:44261|issue:44373", "issue:39401|issue:44521", "issue:44625|issue:44829", "issue:43116|issue:45200", "issue:29127|issue:30990" ], - "prepared_review_unit_hash": "51bb5a0f8cace2378c2ce6df79f57e539e5dfff78f4c8b232a3a2043aede74cf", + "prepared_review_unit_hash": "85fbc7435da8d72d5977ce784cb167d1fe0a76118fb68abaa777baeb5cd46abe", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, @@ -12039,198 +12000,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of unrelated Transformers issues spanning tokenizers, imports, generation, multimodal/chat templates, model-specific bugs, and documentation. There is no clear single duplicate cluster or canonical artifact; all soft edges are rejected.", + "summary": "This cluster is a mixed set of unrelated Transformers issues. The soft pairs share broad themes like tokenization, import regressions, or generation behavior, but they point to different concrete bugs and code paths, so none should be deduplicated.", "confidence": 0.93, - "canonical_issue_reason": null, + "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous; if one representative is needed, issue:42886 is the clearest concrete regression report.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:42886 is a strong representative because it has a specific offline/cache failure mode, version context, and reproducible behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "One is an offline tokenizer cache-loading failure; the other is a quantized model saving failure." + "reason": "Offline tokenizer cache loading vs quantized model saving are different failures." }, { - "left": "issue:44263", - "right": "issue:44373", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "torch.split return handling in a GLM MoE indexer is unrelated to a position_ids docstring error." + "reason": "Both mention imports, but one is a specific missing symbol and the other is a broad v5 upgrade regression." }, { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "Both involve imports, but one is a specific missing hub symbol and the other is broad upgrade-related import breakage." + "reason": "A torch.split indexing bug and a docstring issue are unrelated." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Different model-specific bugs: RoPE layer dropping versus a tied-weights warning issue." + "reason": "Different model-specific bugs: RoPE layer count vs tied-weights warning." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FineGrainedFP8 accuracy regression is unrelated to the Qwen3-VL-Moe bug." + "reason": "Accuracy regression in FineGrainedFP8 is unrelated to a Qwen3-VL-Moe bug." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation sync_gpus kwargs handling and D-FINE auxiliary loss behavior are different code paths." + "reason": "Generation cache/state update bug vs missing auxiliary losses in D-FINE are separate issues." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "A broken env CLI command is not the same as an import-time NameError in sentence-transformers." + "reason": "CLI/env command breakage is unrelated to a sentence-transformers import NameError." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "FP8 accuracy problems and a Qwen3_5 image-specific infinite loop are distinct issues." + "reason": "Different symptoms and models: FP8 accuracy issue vs Qwen3_5 infinite loop on one image." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure is unrelated to tokenizer decoder type changes in v5." + "reason": "Offline cache loading and tokenizer decoder type mismatch are different tokenizer problems." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Missing transformers.onnx import and extra space after added token are unrelated tokenizer/package issues." + "reason": "Missing module import and extra whitespace in tokenization are unrelated." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 configuration guidance is unrelated to a torch.split return-value bug." + "reason": "TF32 configuration guidance does not match a DSA indexer return-value bug." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Tokenization differences across versions and multimodal assistant mask generation are different bugs." + "reason": "Tokenizer output drift across versions is separate from multimodal assistant mask handling." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None handling and AutoProcessor kwargs forwarding affect different APIs." + "reason": "Tokenizer mapping None handling and AutoProcessor kwarg forwarding are different code paths." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Sentence-Transformers loading hangs and a MacOS bus error on CLIP are not the same failure mode." + "reason": "Loading hang for Sentence-Transformers and a MacOS bus error in CLIP are distinct failures." }, { "left": "issue:43232", "right": "issue:44743", "accept": false, - "reason": "Both mention generation/cache, but one is sync_gpus bookkeeping and the other is Qwen3.5 recurrent state reset." + "reason": "Generation kwargs update after sync_gpus is unrelated to recurrent-state reset in qwen3_5." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized models and pipeline-parallel incorrect output are separate problems." + "reason": "Quantized model save failure and pipeline-parallelism output corruption are different bugs." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "Single missing AutoImageProcessor import is not the same as broad import errors after a version upgrade." + "reason": "Cannot import one symbol is not the same as multiple import errors after a version update." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility and pipeline parallelism output errors are unrelated." + "reason": "Python version compatibility and pipeline-parallelism output are unrelated." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Passing kwargs to cached_file and crashing on assistant tool-call messages are different apply_chat_template/processor paths." + "reason": "Cached-file kwargs forwarding and chat-template crashes with tool calls are separate behaviors." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Wrong tokenizer offset mapping and failure to load a specific model are unrelated tokenizer/model-loading issues." + "reason": "Wrong offset mapping for Qwen3 tokenizer is unrelated to loading a specific model failing." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading registered position_ids buffers and tokenizer_class mismatch on save_pretrained are different bugs." + "reason": "Buffer read error on position_ids and tokenizer_class mismatch on save_pretrained are different bugs." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "A missing nn import in sentence-transformers is unrelated to stale Sam3LiteText config references." + "reason": "Import NameError in sentence-transformers is unrelated to stale Sam3LiteText references in auto_mappings." }, { "left": "issue:43122", "right": "issue:44514", "accept": false, - "reason": "General tokenization differences and a batched chat-template padding crash are distinct issues." + "reason": "Tokenizer behavior change and batched chat-template crash are not the same defect." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Same general area, but different failure modes and fixes." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Dequantized model saving and invalid GenerationConfig for GLM-5 are not the same failure." + "reason": "Saving dequantized models and GenerationConfig validation are unrelated." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "The model-type warning and the missing fpn_position_embeddings attribute are separate SAM 3 problems." + "reason": "Model-type warning and missing attribute on vision encoder output are different SAM3 issues." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:44360", + "right": "issue:44485", "accept": false, - "reason": "A numerical precision/config issue is unrelated to a wrong docstring." + "reason": "A missing ReLU in DSA indexer and GLM-5 RoPE implementation are unrelated." }, { - "left": "issue:44360", - "right": "issue:44485", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "DSA indexer activation choice and GLM-5 RoPE implementation are different model components." + "reason": "Precision/config issue in MLA q_a_layernorm and a docstring correction are not duplicates." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset mapping and multimodal assistant masks do not describe the same bug." + "reason": "Tokenizer offset mapping bug and multimodal assistant mask bug affect different paths." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Qwen3.5 num_labels propagation and flash_attention_3 degenerate training are unrelated." - }, - { - "left": "issue:43116", - "right": "issue:45200", - "accept": false, - "reason": "Example script multi-label behavior and Gemma 4 token_type_ids defaults are different training issues." + "reason": "num_labels propagation and flash_attention_3 training degeneration are distinct issues." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "LayoutLMv3 error clarity and Sentence Transformers loading hangs are unrelated." + "reason": "LayoutLMv3 error-message clarity and Sentence-Transformers loading stall are unrelated." } ] }, @@ -12339,238 +12300,238 @@ "issue:39401|issue:44779", "issue:33453|issue:39401" ], - "prepared_review_unit_hash": "e31a2f4be78e378c9b2c6df8112d7b8d641d976e79b1c58db2f37356597311f9", + "prepared_review_unit_hash": "50d59ce4c42ba72670b847d6091f0120c89dc63c13e0acc17574da7e83f326ad", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37376, - "estimated_input_tokens": 9344, - "estimated_eval_tokens": 18944 + "serialized_chars": 37377, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 }, "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 v5 regressions, import errors, tokenizer changes, and model-specific bugs. The soft pairs share broad themes in a few places, but none are strong enough to treat as the same underlying defect.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 45278 is the broadest representative of the cluster: it explicitly aggregates multiple import errors after upgrading to 5.5.0, which matches the cluster\u2019s dominant regression/import-breakage theme.", + "summary": "This cluster is a loose mix of Transformers v5 regressions (imports, tokenizers, save/load, model-specific bugs) rather than one shared duplicate bug. None of the soft edges look like the same concrete issue, so they should all be rejected. The broadest representative issue is the general v5 import-regression report.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45278 is the most umbrella-like report and best represents the broad v5 breakage theme across the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the best single issue to anchor the cluster because it is the most umbrella-like and likely to subsume nearby import-regression reports.", + "best_issue_reason": "issue:45278 has the widest scope and is the most suitable cluster representative for triage, even though the underlying reports are not duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they affect different models and describe different failure modes; not enough evidence of one shared code-path bug." + "reason": "Both are tokenizer regressions in v5, but they describe different models and failure details; not clearly the same bug." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "One is about saving dequantized Ministral/Devstral models, the other about loading Phi-4 multimodal; different model families and operations." + "reason": "Different components and symptoms: dequantized save failure vs model load failure." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe load/regression versus tokenizer_config save_pretrained mismatch; unrelated symptoms and code paths." + "reason": "Qwen3-VL-Moe v5 bug vs tokenizer save_pretrained metadata mismatch; different code paths." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "A model loading/regression issue versus a position_ids buffer read error; not the same bug." + "reason": "Different failures: model-specific v5 breakage vs position_ids buffer access." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Saving a dequantized model and broken gguf loading are both from_pretrained/save-pretrained adjacent, but the concrete failures are different." + "reason": "Saving dequantized weights and gguf loading are unrelated behaviors." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bugs and multi-label example-script output issues are unrelated." + "reason": "Tensor-parallel training bug is unrelated to the classification example script issue." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols and likely different packaging/export regressions." + "reason": "Both are import errors, but for different missing symbols and different packages." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Different model-specific save/load bugs; no clear shared failing code path." + "reason": "One is Qwen3-VL-Moe runtime breakage; the other is FP8 save_pretrained on moe models." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Missing is_offline_mode and missing HybridCache are separate import/export breakages." + "reason": "Different missing imports from different modules; no shared underlying defect." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5 and v5.4+, but one is a bad saved checkpoint and the other is a from_pretrained error; not enough to unify them." + "reason": "Both concern Qwen3.5 v5 regressions, but one is save_pretrained checkpoint corruption and the other is from_pretrained load failure; not clearly the same bug." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Wrong tokenizer decoder type and Deepseek tokenization mismatch are distinct tokenizer regressions." + "reason": "Wrong decoder type and wrong tokenization are related broadly, but not the same concrete tokenizer bug." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "A dtype drift between train/eval and incorrect output under pipeline parallelism are different training/runtime issues." + "reason": "Embedding dtype mismatch in eval is unrelated to pipeline-parallel wrong output." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "SigLIP2 implementation mismatch and a Qwen3_5 image infinite loop are unrelated model bugs." + "reason": "SigLIP2 implementation discrepancy and Qwen3_5 infinite loop are different model-specific problems." }, { "left": "issue:43122", "right": "issue:44977", "accept": false, - "reason": "Tokenizer regression versus flash-attention generation failure; different subsystems and symptoms." + "reason": "Both involve tokenization/flash-attention regressions, but they affect different models and behaviors." }, { "left": "issue:42907", "right": "issue:44222", "accept": false, - "reason": "Both mention saving, but one is dequantized model saving and the other is FP8 MoE save_pretrained; different mechanisms." + "reason": "Saving dequantized models and FP8 save_pretrained are separate serialization issues." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Both touch from_pretrained flows, but one is missing weights in UMT5Encoder and the other is kwargs not reaching cached_file; not the same bug." + "reason": "Missing embed_tokens loading and cached_file kwargs forwarding are unrelated load-path bugs." }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "Sentence-transformers import failure and a PyTorch version check bug in flex_attention are unrelated." + "reason": "Sentence-transformers import breakage and an AuxRequest version-check bug are unrelated." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "A specific missing import symbol and a broad upgrade-wide import regression are not the same concrete defect." + "reason": "Specific missing symbol import and broad import-regression umbrella are not the same underlying issue." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leakage in Qwen3ForCausalLM multi-threaded use and bfloat16 dtype mismatches in CI are unrelated." + "reason": "VRAM leak in Qwen3ForCausalLM and dtype mismatch CI failure are different bugs." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GenerationConfig validation failure and GLM-5 RoPE discussion are different GLM issues." + "reason": "GenerationConfig validation error and RoPE implementation discussion are different GLM-5 issues." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "An import path failure for transformers.onnx and a LayoutLMv3 error-message clarity issue do not match." + "reason": "ONNX import failure and LayoutLMv3 error-message clarity are unrelated." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "VitPose model breakage and Ministral-3 FlashAttention failure are unrelated model-specific bugs." + "reason": "VitPose breakage and Ministral FlashAttention failure are different model-specific regressions." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 API guidance and a missing image patch helper are different concerns." + "reason": "TF32 control warning and missing image-patch helper are unrelated." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption and chunked_prefill cache missing are separate training/runtime bugs." + "reason": "Optimizer state corruption in trainer/FSDP2 is unrelated to chunked_prefill cache missing." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version support regression and an AuxRequest import version-check bug are different compatibility issues." + "reason": "Python version compatibility issue and AuxRequest import version check are different problems." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example-script multi-label output failure and SigLIP2 implementation mismatch are unrelated." + "reason": "Classification script behavior and SigLIP2 implementation mismatch are unrelated." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "SGD args ignored and trainer.evaluate failing after train are different trainer bugs." + "reason": "Optimizer SGD args not used is unrelated to trainer.evaluate() failing after train()." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "An example-script multi-label bug and a pipeline() error-message issue are not the same underlying problem." + "reason": "Example script multi-label behavior and pipeline translation warning are different features." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5/Qwen-MoE adjacent, but they report different failures in different APIs; too weak to merge." + "reason": "Both mention Qwen3.5/Qwen3-Moe on v5, but one is Qwen3-VL-Moe and the other is a from_pretrained issue; not the same bug." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained MoE and Qwen3.5 from_pretrained error are distinct save/load failures." + "reason": "FP8 save_pretrained moe and Qwen3.5 from_pretrained error are distinct serialization/load failures." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "A Qwen3 Omni helper-function behavior bug and a flex_attention import version check are unrelated." + "reason": "Helper output-length logic in qwen3_omni_moe is unrelated to the PyTorch version-check import bug." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "Invalid GenerationConfig and a stale paged generate warning are different generation-path issues." + "reason": "GenerationConfig invalid and stale num_return_sequences warning are different generation issues." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and are described as different offset/tokenization failures." + "reason": "Both are tokenizer regressions, but for different model families and symptom types." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression and Qwen3 tokenizer offset_mapping bug are separate issues." + "reason": "General tokenizer loading regression and Qwen3 offset_mapping bug are not the same concrete failure." } ] }, @@ -12675,214 +12636,214 @@ "issue:43012|issue:43577", "issue:43673|issue:45341" ], - "prepared_review_unit_hash": "384aff5a5e749320365e8611e9cb567e5e544f30cc5cbde81531d58120287e8f", + "prepared_review_unit_hash": "2a48e2bca3bf7f67f6a7d826e1f208d7fe8e1caafcbc738d591c25e168d28163", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36487, - "estimated_input_tokens": 9122, - "estimated_eval_tokens": 18500 + "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 set is mostly a loose collection of unrelated Transformers issues; the only strong commonality is that several touch tokenizer/import/regression areas, but none of the soft pairs are the same underlying bug or change.", - "confidence": 0.62, - "canonical_issue_reason": "issue:45278 is the broadest umbrella report in the set (multiple import errors after upgrading 4.57.0\u21925.5.0) and is the best representative of the shared regression/import-breakage theme.", + "summary": "This cluster is heterogeneous: the soft pairs mostly share only broad subsystem labels (tokenizers, import errors, loading, optimizer, chat API) but not the same concrete bug. I would not merge any of the proposed soft links.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the best umbrella-style issue here because it is the broadest open report about a wave of import regressions after a version update, but it still does not subsume the specific failures in the soft pairs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most general, user-facing, and triage-worthy issue here; it can plausibly subsume nearby import-regression reports better than the narrower, model-specific bugs.", + "best_issue_reason": "issue:45278 is the most suitable representative issue for this set because it is broad, recent, and open, but the rest of the items are too diverse to treat as true duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Both are crashes, but one is a tokenizer mapping lookup problem and the other is a flash-attention path AttributeError; different code paths and fixes." + "reason": "Tokenizer mapping fallback vs flash-attention crash on sink-less models; unrelated code paths." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() and a flex_attention version-check bug are unrelated." + "reason": "Docs cleanup about pipeline removals vs a flex_attention import/version check bug." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Both involve Qwen3.5 generation, but one is invalid GenerationConfig loading and the other is flash-attention runtime behavior; not the same bug." + "reason": "Different failures: invalid GenerationConfig for GLM-5 vs flash-attention generation bug for Qwen3.5." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "A VRAM leak in dataloader threads and an isolated reproduction of another issue are not the same concrete problem." + "reason": "Qwen3 VRAM leak in multithreaded dataloading is not the same as an isolated reproduction of another issue." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A PEFT version pin problem and a PyTorch version check bug are unrelated." + "reason": "PEFT version pinning issue vs PyTorch version gating bug; unrelated." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Both are v5 regressions, but they hit different APIs and failure modes: init_empty_weights loading vs PretrainedConfig type checking." + "reason": "init_empty_weights argument mismatch vs PreTrainedConfig type-check regression; different surfaces." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Batched video processor shape handling and missing torchvision crash paths are distinct." + "reason": "Batched video shape bug vs missing-torchvision crash; both video-related but not the same defect." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "A generic testing_utils bug and a pytest-xdist race on captured_info.txt are not the same underlying defect." + "reason": "Generic testing_utils bug vs xdist race on captured_info.txt; insufficiently specific to be the same bug." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Both are tokenizer internals, but one is a mapping-name None case and the other is a Mistral regex patch crash." + "reason": "None-returning tokenizer mapping and Mistral regex patch crash are distinct tokenizer regressions." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Separate import errors for different symbols; no evidence they stem from one missing export." + "reason": "Different import errors for different symbols; same broad theme, not the same missing-export bug." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Offset-mapping correctness and runtime post-processor update behavior are different tokenizer bugs." + "reason": "Offset mapping bug vs runtime post-processor update bug; separate tokenizer behaviors." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Different models and different failures: missing projections in qwen3next versus Qwen3.5 from_pretrained load error." + "reason": "Layer projections missing in qwen3next vs Qwen3.5 from_pretrained load error; different model-loading problems." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Flash-attention training degeneration and save_pretrained checkpoint regression are unrelated." + "reason": "Flash-attention training degeneration vs save_pretrained checkpoint regression; unrelated failures." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "A multi-thread VRAM leak and an Apple Silicon TypeError in allocator warmup are unrelated." + "reason": "VRAM leak in Qwen3 threading vs Apple Silicon caching_allocator_warmup TypeError; different bugs." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "Chat/completions request validation and Qwen3.5 model loading are different problems." + "reason": "save_pretrained shard-size question vs Qwen3.5 MoE load error; not the same issue." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 control warnings and MLA rms_norm_eps precision issues are unrelated." + "reason": "TF32 configuration warning vs missing rms_norm_eps precision error; both numeric-settings related but not the same defect." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "Broken vitpose models and missing TensorFlow backend when only tensorboard is installed are different installation/runtime issues." + "reason": "Broken VitPose model loading vs missing TensorFlow backend when only tensorboard is installed; unrelated." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A SAM3 model-type warning and a MobileLLM key error do not share the same root cause." + "reason": "Model-type mismatch warning vs MobileLLM key error; different loading problems." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "Both mention imports, but one is a single missing image_utils export while the other is a broad post-upgrade import-breakage report." + "reason": "A specific missing image_utils import is not enough to call it the same as a broad 'many import errors' report." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "Not the same subsystem: 3d_parallel.py execution failure versus a testing_utils bug." + "reason": "3d_parallel.py runtime failure vs a testing_utils bug; different files and behaviors." }, { "left": "issue:42673", "right": "issue:43577", "accept": false, - "reason": "A VRAM leak in threaded loading and a BLIP2 dtype persistence bug are unrelated." + "reason": "Qwen3 threading VRAM leak vs BLIP2 dtype staying float32; separate model/runtime issues." }, { "left": "issue:42175", "right": "issue:42222", "accept": false, - "reason": "Missing TensorFlow from a torch-only install and vitpose model breakage are not the same issue." + "reason": "Dependency installation/backend packaging issue vs VitPose model breakage; not the same bug." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "A generation-config validation error and an unrecognized-model config.json error are separate load-time failures." + "reason": "Invalid GenerationConfig for a known model vs unrecognized model_type in config.json; different config-loading errors." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "Different import/utility failures: missing layer_type_validation versus an audio feature extraction issue." + "reason": "ImportError for a missing symbol vs audio feature extraction failure; unrelated." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD arguments ignored and optimizer.step() being called twice under deepspeed are distinct optimizer bugs." + "reason": "Unused SGD args and double optimizer-step under deepspeed are different optimizer bugs." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Both concern chat/completions, but one is endpoint acceptance and the other is streaming inference on a specific model." + "reason": "Chat/completions request rejection vs streaming inference failure on Qwen3.5-0.8B; different API breakages." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 settings warnings and FineGrainedFP8 accuracy loss are different precision-related problems." + "reason": "TF32 behavior control and FineGrainedFP8 accuracy issues are both precision-related but not the same change or bug." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions, but one is a decoder-type mismatch and the other is changed tokenization output across versions." + "reason": "Wrong decoder type and changed tokenization output are related to v5 tokenizer behavior but not clearly the same underlying bug." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "A bfloat16 warning during compilation and BLIP2 dtype loading behavior are not the same bug." + "reason": "PyTorch warning during bfloat16 compilation vs BLIP2 dtype not propagating; distinct dtype-handling issues." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache missing in chunked_prefill and a testing_utils bug are unrelated." + "reason": "Generation cache missing in chunked_prefill vs a testing_utils bug; completely different areas." } ] }, @@ -12995,238 +12956,262 @@ "issue:43824|issue:43986", "issue:43066|issue:43906" ], - "prepared_review_unit_hash": "7bbda4c4de4d25d90374b6ae9ab5c9ecbd59ff5d3fbbe3012fc5ea3793ebf2d9", + "prepared_review_unit_hash": "f3b62b55831eac313479b29eb450e0c9d09e41e1e640d10e76c6d6072b220ec0", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 38, - "serialized_chars": 39087, - "estimated_input_tokens": 9772, - "estimated_eval_tokens": 19800 + "serialized_chars": 39090, + "estimated_input_tokens": 9773, + "estimated_eval_tokens": 19802 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a collection of distinct Transformers bug reports, not a single duplicate cluster. None of the soft-pair candidates look like the same underlying bug/change, so all should be rejected.", - "confidence": 0.91, - "canonical_issue_reason": null, + "summary": "This cluster is thematically similar but not duplicate-tight: the items span unrelated import errors, model-loading regressions, config/runtime warnings, and training issues. None of the soft pairs appear to describe the same underlying bug or change.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45478 is the best representative of the cluster's recurring Transformers v5 model-loading regressions: it is concrete, versioned, and actionable.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue cleanly represents this heterogeneous set; the strongest candidates are all about different bugs (imports, model loading, config, FlashAttention, training).", + "best_issue_reason": "issue:45478 is the strongest overall issue candidate because it is specific, clearly reproducible, and closest to the cluster's main loading/regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen-related failures: missing layer projections in qwen3next vs a general from_pretrained error for Qwen3.5 MoE." + "reason": "Different Qwen-family problems: missing layer projections vs a from_pretrained failure in Qwen3.5 MoE." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Unrelated topics: shard-size configuration question vs model loading failure." + "reason": "Unrelated topics: shard-size documentation/question vs Qwen3.5 MoE loading error." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "One is an explicit API removal/regression warning; the other is a position-id key mismatch during model loading." + "reason": "Different regressions: removal of tie_embeddings_and_encoder_decoder vs unexpected position-id keys." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different GLM-family bugs: missing image-processor method vs missing RMS norm epsilon in config." + "reason": "Different subsystems: image-patch counting AttributeError vs MLA rms_norm_eps precision issue." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "Both GLM-related, but they describe different code paths: DSA indexer split handling vs RoPE implementation." + "reason": "Both GLM-related, but distinct code paths: torch.split behavior vs RoPE implementation." }, { "left": "issue:42222", "right": "issue:45459", "accept": false, - "reason": "VitPose model breakage and tokenizer/protobuf error handling are unrelated." + "reason": "Different bug classes: broken vitpose models vs tokenizer error masking when protobuf is absent." }, { "left": "issue:43012", "right": "issue:43408", "accept": false, - "reason": "A bfloat16 precision warning is not the same as an AutoModel type mismatch warning." + "reason": "A warning during bf16 compilation is unrelated to a model-type mismatch warning." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression are separate issues." + "reason": "Importing AutoImageProcessor is unrelated to PretrainedConfig type-checking changes." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both involve FlashAttention, but one is a Ministral-3 runtime failure and the other is degenerate training with flash_attention_3." + "reason": "Different FlashAttention failures in different models/tasks; not the same concrete bug." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "PretrainedConfig typing regression is unrelated to the TensorFlow/tensorboard backend error." + "reason": "PretrainedConfig typing and TensorFlow/tensorboard backend detection are unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Different Qwen model families and failure modes: shape mismatch on load vs missing projections in qwen3next." + "reason": "Different Qwen model failures: weight-shape mismatch vs missing attention projections." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a testing_utils bug." + "reason": "Unsafe sys.modules access is unrelated to a small testing_utils bug." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Both are Qwen VL loading issues, but one is a weight-shape mismatch and the other is a Qwen3-VL-Moe-specific bug." + "reason": "Both involve Qwen3-VL, but one is a weight mismatch and the other is a broader version/MoE incompatibility." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "A tokenizer mapping None regression is not the same as a specific model loading failure for physical-intelligence/fast." + "reason": "Tokenizer mapping returning None is not the same as failing to load a specific model repo." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Tie-embeddings API removal and gguf_file loading errors affect different code paths." + "reason": "Removal of a config field is unrelated to gguf_file breaking AutoTokenizer/AutoModelForCausalLM loading." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer creation bug is unrelated to config type checking." + "reason": "SwitchTransformers sparse-layer creation is unrelated to PretrainedConfig type checking." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Video-model AutoProcessor crash without torchvision and a testing_utils bug are unrelated." + "reason": "A processor crash without torchvision is unrelated to a testing_utils bug." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "Different warnings in different subsystems: model-type mismatch vs tie_word_embeddings guidance." + "reason": "These are different warning conditions with different model/config causes." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Model loading without torchvision is unrelated to a max_shard_size question." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Different GLM bugs in different components; not the same underlying defect." + "reason": "Different GLM-related failures: missing image-patch method vs torch.split return handling." }, { "left": "issue:43976", "right": "issue:45071", "accept": false, - "reason": "Python-version compatibility problem and PretrainedConfig type checking are separate regressions." + "reason": "Python version compatibility problem is unrelated to config type checking." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding-window bug has nothing to do with testing utilities." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to a testing_utils bug." }, { "left": "issue:44829", "right": "issue:45245", "accept": false, - "reason": "FlashAttention training degeneration and category-cardinality RuntimeError are unrelated." + "reason": "FlashAttention training degradation is unrelated to the categorical cardinality runtime error." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Both are loading-time failures, but one is torchvision availability and the other is a missing mistral_common import." + "reason": "Missing torchvision crash is unrelated to the ReasoningEffort import regression." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different missing symbols from different packages." + "reason": "Both are import errors, but they come from different packages and symbols." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Same model family, but one is a missing config epsilon issue and the other is a RoPE implementation discussion." + "reason": "Different GLM issues: layernorm epsilon precision vs RoPE implementation." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Training/optimizer-state corruption with FSDP2+PEFT is not the same as Liger Kernel not being applied with model_init." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to Liger Kernel not being applied." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Different model families and different problems: missing attention projections vs a config field being dropped." + "reason": "Missing attention projections in qwen3next is unrelated to a missing config field in Qwen3_5MoeVisionConfig." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Different import failures for different symbols; same general symptom, not the same bug." + "reason": "Different import failures from different modules and symbols." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "API removal warning and flash_attention s_aux=None crash are different code paths." + "reason": "Config-field removal is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Tokenizer AddedToken argument duplication and TOKENIZER_MAPPING_NAMES returning None are separate tokenizer regressions." + "reason": "Tokenizer AddedToken argument duplication is unrelated to TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Both touch Qwen models, but the concrete failures differ: Qwen3-VL-Moe vs qwen3next layer projections." + "reason": "Both touch Qwen variants, but one is a broader loading bug and the other is a missing-layer-0 projection issue." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "cached_file kwargs propagation and auto_docstring annotation handling are unrelated." + "reason": "Passing kwargs to cached_file is unrelated to auto_docstring crashes under future annotations." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Offline-loading network requests and UMT5Encoder missing embed_tokens.weight are different loading bugs." + "reason": "local_files_only network leakage and missing embed_tokens.weight are separate loading bugs." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Explicit API removal and quantization/offloading behavior are unrelated." + "reason": "Embedding-tie removal and offloading/quantization behavior are not the same bug." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision and tokenizer mapping None are different failure modes." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Qwen2.5-VL import failure and video-model processor crash are distinct model-loading problems." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Tokenizer decoder type regression is unrelated to the isolated reproduction of another issue." } ] }, @@ -13322,8 +13307,8 @@ "issue:43618|issue:43761", "issue:38617|issue:43502", "issue:44038|issue:44230", - "issue:43756|issue:44485", "issue:44483|issue:45290", + "issue:43756|issue:44485", "issue:43299|issue:43867", "issue:44483|issue:45381", "issue:43299|issue:45753", @@ -13339,262 +13324,262 @@ "issue:44220|issue:44655", "issue:43976|issue:44938" ], - "prepared_review_unit_hash": "02c9595ed667a7578a0ba8734cda07bda5e0087c7fed18d7467bac48c11387f6", + "prepared_review_unit_hash": "fd0e67524222420abf982f743f8d4a9a103830c10ddbbc45e73057dd0957da18", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 38, - "serialized_chars": 38783, - "estimated_input_tokens": 9696, - "estimated_eval_tokens": 19648 + "serialized_chars": 38790, + "estimated_input_tokens": 9698, + "estimated_eval_tokens": 19652 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: mostly separate Transformers regressions across v5 API removals, loading/saving, chat templates, vision models, and quantization. The soft links are largely keyword-level similarity rather than the same underlying bug.", - "confidence": 0.83, - "canonical_issue_reason": "issue:43097 is the broadest and most central v5 breakage report in the set, with inbound references and a clear API/regression angle; it is the closest thing to an umbrella issue, though the cluster is not a true duplicate set.", + "summary": "The cluster is heterogeneous: most items are distinct Transformers regressions or feature/compatibility issues that only share broad theme words like v5, loading, saving, or attention. None of the soft edges look like true duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "issue:44483 is the broadest user-facing v5 regression in the set, but the cluster is too mixed to treat it as a true duplicate canonical.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest representative of the overall cluster theme, but the surrounding issues are too diverse for it to be a clean duplicate target.", + "best_issue_reason": "If a representative issue must be chosen, issue:44483 is the most generic and broadly impactful, though still not a mergeable umbrella for the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43097", "right": "issue:43901", "accept": false, - "reason": "Removed encoder-decoder tie API vs stale docs for return_all_scores; unrelated symptoms and code paths." + "reason": "Removed API warning/deprecation vs docs mismatch; same release era, but not the same bug." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline translation-task messaging vs pipeline save_pretrained failure; different behavior and modules." + "reason": "Chat template/task support error vs pipeline serialization; different code paths and symptoms." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens weight in UMT5 encoder vs default max_shard_size choice; not the same bug." + "reason": "Missing model weight on load vs save_pretrained shard sizing question; unrelated save/load problems." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Flash Attention unsupported error in Nemotron vs GPT2 attention scaling ignored; different model/backend issues." + "reason": "Flash Attention support limitation vs GPT2 attention scaling being ignored; different attention bugs." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs vs Bamba compile+SDPA failure; no shared concrete code-path." + "reason": "xLSTM training blockers vs Bamba torch.compile/SDPA failure; different models and root causes." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "PyTorch warning on bfloat16 compile vs model failure under torch.compile; warning-only vs functional bug." + "reason": "Precision warning during compile vs actual compile failure under SDPA; not the same defect." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Compile-time float32 warning vs tie_word_embeddings warning; same general warning theme, different cause." + "reason": "bfloat16 compile warning vs tie_word_embeddings warning; both warnings, but different code paths." }, { "left": "issue:42886", "right": "issue:43381", "accept": false, - "reason": "Offline cache loading regression vs gradient checkpointing eval-mode restriction; unrelated." + "reason": "Offline cache loading failure vs gradient checkpointing in eval mode; unrelated behaviors." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Saving dequantized models vs Qwen3.5 infinite loop on an image; entirely different bugs." + "reason": "Quantized/dequantized save failure vs image-generation infinite loop; different models and failure modes." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Dequantized model save failure vs pipeline.save_pretrained failure; different save flows and failure modes." + "reason": "Dequantized model saving vs pipeline object serialization; both saving-related, but not the same bug." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention scaling under SDPA/FlashAttention vs single-label classification zero loss; no overlap in code path." + "reason": "Attention backend scaling bug vs degenerate single-label classification loss; unrelated subsystems." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Flash Attention support check vs Python version compatibility load failure; unrelated." + "reason": "Missing Flash Attention support vs Python version compatibility failure; no shared root cause." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2/PEFT optimizer state mismatch vs GPT2 attention scaling backend bug; different subsystems." + "reason": "FSDP2 optimizer-state corruption vs attention scaling ignored; different training vs inference issues." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training blockers vs BeitImageProcessorFast label reduction bug; not the same defect." + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; unrelated components." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size vs Apple Silicon allocator TypeError; unrelated runtime failures." + "reason": "Gemma3Config attribute missing vs Apple Silicon TypeError in allocator warmup; unrelated errors." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Chat/completions request rejection vs batched apply_chat_template crash; related to chat APIs but different concrete failure." + "reason": "Chat/completions request rejection vs batched apply_chat_template crash; both v5-facing, but different APIs and code paths." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "A warning during torch.compile vs flex_attention failure on torch>=2.9; different backend issues." + "reason": "Compile warning vs compiled flex_attention failure on newer torch; different failure severity and cause." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIPOutput attentions assignment regression vs GPT2 attention scaling backend bug; different models and symptoms." + "reason": "CLIPOutput attentions not assigned vs GPT2 attention scaling ignored; both mention attention, but not the same bug." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Missing hub utility import vs missing model-class import; both import errors, but unrelated symbols and fixes." + "reason": "Missing huggingface_hub symbol vs missing Transformers export; both import errors, but for different modules/symbols." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Attentions field no longer assigned in CLIPOutput vs hidden_states missing in CLIPVisionModel.forward; related model family, but distinct outputs and regressions." + "reason": "Missing attentions output vs hidden_states=None regression; same model family, different output fields and fixes." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "Missing configuration_utils symbol vs API requests despite local_files_only; one import regression and one network-behavior bug." + "reason": "Missing internal import vs unexpected network access when local_files_only=True; different offline-related problems." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Qwen3-VL-Moe loading/bug report vs fp8 support request; feature support and bug report are not the same issue." + "reason": "Loading bug in Qwen3-VL-Moe vs fp8 support request for Qwen3-VL/Qwen3.5 MoE; bug vs feature request." }, { - "left": "issue:43756", - "right": "issue:44485", + "left": "issue:44483", + "right": "issue:45290", "accept": false, - "reason": "Smollm3 RoPE-layer dropping bug vs GLM-5 RoPE implementation discussion; same broad topic, different models and concrete defects." + "reason": "Chat/completions endpoint rejection vs assistant tool-call template crash; different layers of the chat stack." }, { - "left": "issue:44483", - "right": "issue:45290", + "left": "issue:43756", + "right": "issue:44485", "accept": false, - "reason": "Request rejection at /v1/chat/completions vs assistant tool-call chat_template crash; different input shapes and failure points." + "reason": "Smollm3 RoPE layer drop bug vs GLM-5 RoPE discussion; both RoPE-themed, but not the same issue." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3.5-VL-Moe loading breakage vs state_dict-sorted load error; both loading related but not the same underlying bug." + "reason": "Qwen3VL-Moe loading break vs sorted-state_dict load error; both load failures, but different triggers and models." }, { "left": "issue:44483", "right": "issue:45381", "accept": false, - "reason": "Chat/completions request handling vs Qwen2.5-VL video vision_position_ids bug; unrelated paths." + "reason": "Chat/completions request handling vs Qwen2.5-VL video position-ids bug; unrelated failures." }, { "left": "issue:43299", "right": "issue:45753", "accept": false, - "reason": "Qwen3.5-VL-Moe loading failure vs Qwen3_5 infinite loop on a specific image; different model behavior and failure mode." + "reason": "Qwen3VL-Moe loading regression vs Qwen3.5 image infinite loop; different model families and symptoms." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "TextClassificationPipeline docs mismatch vs pipeline.save_pretrained inability; docs issue vs serialization bug." + "reason": "Docs mention stale return_all_scores vs pipeline save_pretrained failure; documentation drift is not the same as serialization bug." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Gemma3Config vocab_size attribute missing vs SAM3 vision encoder output missing fpn_position_embeddings; different model classes and attributes." + "reason": "Missing Gemma3Config attribute vs missing SAM3 vision encoder attribute; both are attribute errors, but unrelated models." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure vs pipeline object save_pretrained failure; different serialization targets." + "reason": "Quantized model save failure vs pipeline object save_pretrained; both save-related, but not the same code path." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Sorted state_dict load error vs Qwen3.5 checkpoint corruption after save_pretrained; both checkpoint/load related but not the same defect." + "reason": "Sorted state_dict load issue vs Qwen3.5 checkpoint regression in save_pretrained; different save/load directions and root causes." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "Flash Attention support error vs GenerationConfig invalid for GLM-5; different validation and model-init paths." + "reason": "Flash Attention support limitation vs invalid GenerationConfig for GLM-5; unrelated model-specific problems." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "patchtsmixer post_init policy issue vs GPT2 attention scaling backend bug; unrelated." + "reason": "patchtsmixer post_init policy issue vs GPT2 attention scaling bug; no shared underlying defect." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different modules ('GenerationMixin' vs 'is_offline_mode'); superficial import-error similarity only." + "reason": "Different missing imports from different modules; both are import errors, but not the same broken symbol." }, { "left": "issue:43066", "right": "issue:45381", "accept": false, - "reason": "Tokenizer decoder type regression vs Qwen2.5-VL video position-id bug; different subsystems." + "reason": "Tokenizer decoder-type regression vs Qwen2.5-VL video position_ids bug; separate v5 regressions." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3Config missing vocab_size vs patchtsmixer post_init optionality; no common bug." + "reason": "Gemma3 config missing vocab_size vs patchtsmixer post_init validation; unrelated model/config issues." }, { "left": "issue:44220", "right": "issue:44655", "accept": false, - "reason": "Audio feature extraction issue vs pipeline.save_pretrained failure; unrelated functionality." + "reason": "Audio feature extraction bug vs pipeline save_pretrained serialization; different subsystems entirely." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python 3.9+ incompatibility in 5.1.0 vs Python 3.14 load failure; different version constraints and causes." + "reason": "Different Python compatibility failures (3.9/3.10 vs 3.14); same general theme, but not the same bug." } ] }, @@ -13700,236 +13685,233 @@ "issue:42890|issue:45600", "issue:43502|issue:44062" ], - "prepared_review_unit_hash": "ef55e5bf3bc89460b8392af2c4f15b2c31d9f523fee63981fac14780880f577a", + "prepared_review_unit_hash": "72111cc5762b6f25e8019af46e247fd431d2e7ef024e31105664be3204d2aa37", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 32, - "serialized_chars": 36292, - "estimated_input_tokens": 9073, - "estimated_eval_tokens": 18402 + "serialized_chars": 36293, + "estimated_input_tokens": 9074, + "estimated_eval_tokens": 18404 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a mixed cluster of unrelated Transformers regressions, but it does contain more than one duplicate-style pocket: the import-time slowdown pair and a separate Qwen3.5 from_pretrained pair (45310/45478).", - "confidence": 0.36, - "canonical_issue_reason": "issue:45278 is still the broadest umbrella report, but it should be treated as one representative of a larger mixed cluster rather than the only strong duplicate theme.", + "summary": "The soft pairs are mostly same-subsystem but not the same underlying bug or change; I would reject all of them as duplicates. The broadest issue in the set is the open import-regression report (issue:45278), which is the best triage anchor, but there is no clear duplicate cluster here.", + "confidence": 0.82, + "canonical_issue_reason": "No soft pair is a strong duplicate; issue:45278 is the broadest umbrella report and the most useful stand-in canonical issue for triage.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best single cluster anchor because it captures the broad import/load regression theme that spans the clearest soft duplicate pocket.", + "best_issue_reason": "issue:45278 is open, broad, and phrased as a multi-model regression after upgrade, so it is the best single issue to anchor the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "Different bugs: a DSA indexer split behavior issue vs a DeepSeekV3MoE implementation divergence." + "reason": "Different models and failures: a GlmMoeDsaIndexer torch.split issue vs a DeepSeekV3MoE implementation mismatch." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Both are config-attribute errors, but for different models and different missing fields." + "reason": "Both are missing-config-attribute reports, but for different models and different fields; not the same bug." }, { "left": "issue:43352", "right": "issue:45593", "accept": false, - "reason": "Unrelated functionality: Flash Attention support for Nemotron vs auxiliary-loss handling in D-FINE." + "reason": "Unrelated bugs: FlashAttention support gating for Nemotron vs D-FINE auxiliary loss behavior." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "Different problems: CLIPOutput attentions assignment vs Qwen3.5 save_pretrained checkpoint serialization." + "reason": "CLIPOutput attentions assignment bug vs Qwen3.5 save_pretrained checkpoint regression; different code paths." }, { "left": "issue:44188", "right": "issue:44380", "accept": false, - "reason": "Both involve attention backends, but the concrete failure modes and affected code paths are different." + "reason": "Both involve attention internals, but one is a torch.compile branch divergence and the other is backend scaling config handling." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Unrelated: distributed optimizer-state bug vs a cache-strategy typo." + "reason": "Training/FSDP optimizer-state corruption is unrelated to a cache-strategy typo." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "Different areas: missing config attribute vs MoE load-balancing loss logic." + "reason": "Different model-loading/config issues: missing vocab_size vs MoE load-balancing loss omission." }, { "left": "issue:43352", "right": "issue:43931", "accept": false, - "reason": "Different model/backend issues: Flash Attention support vs Qwen3-VL weight shape mismatch." + "reason": "Nemotron FlashAttention support limitation is not the same as Qwen3-VL weight-shape mismatch." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "The second is a broad import/regression report; it is not the same concrete cache-update bug." + "reason": "Cache update no_grad decoration and broad import errors after upgrade are different regressions." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Both are from_pretrained/loading problems, but they affect different models and failure modes." + "reason": "UMT5Encoder missing embed_tokens.weight is unrelated to Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Both are model I/O issues, but one is a state_dict loading error and the other is a missing config field during save/load." + "reason": "A sorted-state_dict loading error is not the same as a missing Qwen3_5MoeVisionConfig field." }, { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Both point to the same underlying import-time performance regression: slow `import transformers`, with 41669 identifying the likely cause and 44246 reporting the symptom." + "accept": false, + "reason": "Both concern import performance, but one is a specific 'import *' cleanup issue and the other is a general slow-import report." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Different import/compile failures with different triggers: compiled Python files vs a torch.jit parsing issue on Python 3.13." + "reason": "Generic compiled-Python-file errors are not the same as a Python 3.13 @torch.jit.script parsing/IndentationError." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "No overlap beyond being bugs in Transformers; one is a config attribute issue and the other is distributed training state corruption." + "reason": "A missing config attribute in Gemma3 is unrelated to FSDP2/PEFT optimizer-state corruption." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Both are test instability reports, but one is seeding-related flakiness and the other is a pytest-xdist file race." + "reason": "A missing set_seed in a model test is not the same as a pytest-xdist file-write race." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Different regression classes: Python 3.14 import failure vs PretrainedConfig type-checking breakage." + "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are different compatibility issues." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Both involve loading/saving Qwen3.5-related artifacts, but the concrete bug paths differ." + "reason": "Both are model-loading/saving regressions, but they affect different models and different serialization paths." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "The second is a broad complaint about 5.0 breaking many models; it is not the same concrete cache-update bug." + "reason": "A specific cache-method change is not the same as a vague report that 5.0 breaks too many models." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Same general family of model-loading errors, but different models and different missing components." + "reason": "Different models and symptoms: UMT5 weight omission vs Qwen3.5 MoE from_pretrained error." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Different issues: cache update decoration vs Python-version compatibility failure." + "reason": "Cache no_grad decoration and Python version support for Transformers 5.1.0 are unrelated." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Different bugs: CLIP attentions output vs patchtsmixer post_init policy." + "reason": "CLIPOutput attentions assignment and PatchTSMixer post_init allowance are unrelated model API issues." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Both mention version-related failures, but one is a Python 3.14 import issue and the other is a Qwen2.5-VL video input positioning bug." + "reason": "Python 3.14 load failure is not the same as qwen2.5-vl video vision_position_ids being wrong." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "Unrelated DSA-indexer bug vs DeepSeekV3 MoE divergence." + "reason": "A missing ReLU in a DSA indexer is unrelated to DeepSeekV3MoE divergence from the remote implementation." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Both mention tensor parallelism, but one is a training bug and the other is an OOM during model loading." + "reason": "Tensor-parallel training bug and TP-related OOM on model load are different failure modes and likely different causes." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "Different symptoms and subsystems: missing PyTorch dependency vs tokenizer-loading regression." + "reason": "PyTorch-not-found ImportError and tokenizer-loading regression are unrelated." }, { "left": "issue:43867", "right": "issue:44484", "accept": false, - "reason": "Different model-loading bug vs a save_pretrained shard-size question." + "reason": "Model-loading failure is unrelated to the default max_shard_size question." }, { "left": "issue:43618", "right": "issue:44977", "accept": false, - "reason": "Different attention-related problems affecting different models and code paths." + "reason": "CLIPOutput attentions regression is unrelated to Qwen3.5 flash-attention generation issues." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "No meaningful overlap: CLIP attentions assignment vs an AMP/CUDA crash in LwDetrImageLoss." + "reason": "Different affected components: CLIPOutput attentions vs LwDetrImageLoss AMP crash." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Different areas: flaky test seeding vs a multimodal model loading regression." + "reason": "A flaky test due to missing seeds is unrelated to loading Phi-4 multimodal with latest transformers." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Both are CI/test issues, but they stem from different causes and are not the same bug." + "reason": "A seed-related test flake is not the same as a GraniteSpeech stale device override CI failure." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Different CI failures: one is test seeding/flakiness, the other is stale auto_mappings references." + "reason": "A flaky integration test is unrelated to auto_mappings.py referencing removed Sam3LiteText configs." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Completely different problems: local_files_only still triggering network calls vs AddedToken constructor argument duplication." + "reason": "Local-files-only network leakage and AddedToken keyword duplication are different tokenizer/loading bugs." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The soft-edge verdicts are mostly conservative, but the summary overstates the cluster shape: the packet contains at least one additional obvious duplicate-style pair (issue:45310 and issue:45478 share the same Qwen3.5 from_pretrained error title), so it is not grounded to say there is only one strong duplicate pair. Because of that overclaim, I would not accept this cluster evaluation." - }, - "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 }, @@ -13999,8 +13981,8 @@ "issue:42548|issue:44351", "issue:43010|issue:44075", "issue:44387|issue:44556", - "issue:42890|issue:45341", "issue:44222|issue:45125", + "issue:42890|issue:45341", "issue:43010|issue:43065", "issue:44355|issue:44749", "issue:43502|issue:44485", @@ -14036,14 +14018,14 @@ "issue:41628|issue:44355", "issue:36296|issue:43526" ], - "prepared_review_unit_hash": "0f9dba16886b7a4032c0b8d801684957eaa898eb4efa9662807a4c95f927bd3f", + "prepared_review_unit_hash": "a3b9149a31929f5810e68e6fed853491a05f05de43134eab493b1646a8d60a95", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 47, @@ -14058,168 +14040,168 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers issues. The soft edges are all false duplicates on title similarity alone; the only common thread is broad v5 regression chatter, not the same bug.", - "confidence": 0.89, - "canonical_issue_reason": "issue:44556 is the closest representative of the shared theme: a concrete v5 upgrade/regression loading failure. It is still not a true duplicate hub, but it is broader and more central than the others.", + "summary": "This cluster is mostly noisy: many issues share Transformer/v5/loading/import wording but appear to describe different bugs. I only see one soft edge that plausibly points to the same underlying model-loading problem.", + "confidence": 0.58, + "canonical_issue_reason": "issue:43576 is the broadest and most central v5 regression report in the set, with inbound references and cluster-wide similarity to other v5 breakage reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44556 best captures the main cluster theme of existing workflows breaking after upgrading to v5, even though the surrounding issues are not duplicates of it.", + "best_issue_reason": "issue:43576 is the best representative issue because it is the most umbrella-like and high-signal among otherwise mostly unrelated regression reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer loading and the other is checkpoint reload after upgrade; different failure points and likely different fixes." + "reason": "Both are loading failures, but one is tokenizer loading and the other is checkpoint reload after upgrade; different concrete bugs." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different symbols and different code paths." + "reason": "Both are import errors, but they concern different exported symbols and likely different missing re-exports." }, { "left": "issue:43010", "right": "issue:44075", "accept": false, - "reason": "Cache update no_grad behavior and SGD optimizer args are unrelated bugs." + "reason": "Cache update/no_grad and SGD args not used are unrelated code paths and symptoms." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 CUDA memory/OOM behavior is unrelated to model reload failure after upgrading." + "reason": "Int4 OOM from reserved memory and checkpoint reload failure are unrelated regression types." }, { - "left": "issue:42890", - "right": "issue:45341", + "left": "issue:44222", + "right": "issue:45125", "accept": false, - "reason": "Both mention testing, but one is missing set_seed in a specific integration test and the other is a generic testing_utils issue." + "reason": "FP8 save_pretrained bug and missing tensor-parallel plan for Qwen3.5 Moe are different problems." }, { - "left": "issue:44222", - "right": "issue:45125", + "left": "issue:42890", + "right": "issue:45341", "accept": false, - "reason": "FP8 save_pretrained behavior and tensor-parallel _tp_plan support are different concrete problems." + "reason": "Both touch testing, but one is flaky integration tests and the other is a generic testing_utils bug; not the same issue." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "No shared bug: one concerns cache/layer update semantics, the other a dummy Conv2d in Sam3PixelDecoder." + "reason": "Cache update semantics and a dummy Conv2d in Sam3PixelDecoder are unrelated." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and slower filtering after upgrade are unrelated." + "reason": "Compiled Python file errors and a data-loading slowdown after upgrade are different issues." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "local_files_only network leakage and GLM-5 RoPE implementation are different issues." + "reason": "local_files_only network leakage and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:45125", "right": "issue:45478", - "accept": false, - "reason": "Both are Qwen3.5 MoE loading complaints, but one is tensor-parallel plan support and the other is a broader from_pretrained regression; too little evidence they are the same bug." + "accept": true, + "reason": "Both point to the same Qwen3.5 Moe loading regression around from_pretrained / tensor-parallel support in v5.4.x." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "SentenceTransformer loading hang and forward-doc confusion are unrelated." + "reason": "Stuck loading Sentence Transformers and docs confusion for model.forward are unrelated." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Different generation/cache bugs with different symptoms and code paths." + "reason": "Different generation/cache bugs with different failing code paths." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Missing deepstack_visual_indexes in config is unrelated to missing _tp_plan for tensor parallelism." + "reason": "Tensor-parallel Qwen3.5 Moe loading and missing deepstack_visual_indexes in vision config are different config issues." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLFlow reporting on failed training and tokenizer-loading regression are separate issues." + "reason": "MLflow reporting failure and tokenizer-loading regression are unrelated." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "One is offline/local file behavior, the other is FP8 save_pretrained for MoE models." + "reason": "Network request suppression and FP8 save_pretrained are different bugs." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but they affect different models and likely different causes." + "reason": "Qwen3VL Moe loading regression and checkpoint reload after version upgrade are not the same concrete failure." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger kernel application and a DSA indexer ReLU discussion are unrelated." + "reason": "Liger Kernel application and DSA indexer missing ReLU are unrelated." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "A network/files-only bug and a nonsensical issue title are not duplicates." + "reason": "local_files_only behavior and an unrelated issue titled 'racoon' do not match." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FP8 accuracy and a missing image-processor method are unrelated." + "reason": "FP8 accuracy and missing image-patch count method are different defects." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "Training-job MLFlow reporting and tensor-parallel training are different bugs." + "reason": "MLflow failure reporting and tensor-parallel training bug are unrelated." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-Python loading errors and PretrainedConfig type checking are unrelated." + "reason": "Compiled file loading failure and PretrainedConfig type-checking regression are different problems." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both mention FP8/MoE, but one is save_pretrained and the other inference support; not the same concrete bug." + "reason": "FP8 save_pretrained and FP8 inference support are related by theme, but not the same bug or code path." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled file errors and inability to save Pipeline objects are distinct issues." + "reason": "Compiled-file errors and Pipeline.save_pretrained failure are separate save/load issues." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline and an unrelated placeholder issue are not duplicates." + "reason": "Docs still referencing pipeline() and an unrelated issue titled 'racoon' are not duplicates." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading regressions, but they concern different models and different breakages." + "reason": "Tokenizer-loading regression and Llama 3.2 vision model loading breakage are different model families and failures." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing huggingface_hub import and broken transformers env command are different failures." + "reason": "Missing huggingface_hub import and broken transformers env command are different surface areas." }, { "left": "issue:33453", @@ -14231,7 +14213,7 @@ "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env command and PretrainedConfig type-checking regression are unrelated." + "reason": "Broken env CLI and PretrainedConfig type checking are different regressions." }, { "left": "issue:43576", @@ -14243,74 +14225,77 @@ "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing imports: AutoImageProcessor versus PreTrainedModel." + "reason": "AutoImageProcessor and PreTrainedModel import failures are different missing exports." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "A documentation clarification and an import error are not the same bug." + "reason": "Documentation confusion and missing layer_type_validation import are unrelated." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No meaningful overlap: placeholder title versus Gemma-4 positional encoding issue." + "reason": "No meaningful relation between the two reports." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Missing layer_type_validation import and broken env command are separate issues." + "reason": "Missing configuration_utils import and broken env command are unrelated." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Generation cache missing and a cache-strategy typo are unrelated." + "reason": "Missing cache during chunked_prefill and a typo in cache strategies are different issues." }, { "left": "issue:43576", "right": "issue:44938", "accept": false, - "reason": "Broken env command and Python 3.14 load failure are different compatibility problems." + "reason": "Broken env command and Python 3.14 load failure are not the same bug." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs handling and an unrelated placeholder issue are not duplicates." + "reason": "Scheduler kwargs bug and unrelated issue titled 'racoon' do not match." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation and an unrelated placeholder issue are not the same." + "reason": "GLM-5 RoPE implementation and unrelated 'racoon' issue are unrelated." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training bug and text-classification docs mention are unrelated." + "reason": "Tensor-parallel training bug and TextClassificationPipeline docs mismatch are different." }, { "left": "issue:41628", "right": "issue:44355", "accept": false, - "reason": "Missing import and compiled-file runtime errors are different failure modes." + "reason": "AutoImageProcessor import error and compiled-file loading errors are different problems." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor-parallel training and BeitImageProcessorFast label reduction are unrelated." + "reason": "Tensor-parallel training bug and BeitImageProcessorFast reduce_labels bug are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is cautious and stays within what the packet supports: issue:43576 does have inbound references and several similarity links, and the soft-edge verdicts are generally conservative given only titles/body metadata. The one accepted Qwen3.5 MoE pair also looks plausibly like the same bug family." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -14392,8 +14377,8 @@ "issue:44492|issue:45850", "issue:43572|issue:45850", "issue:42371|issue:44492", - "issue:42890|issue:44908", "issue:45278|issue:45850", + "issue:42890|issue:44908", "issue:44704|issue:44908", "issue:43408|issue:44315", "issue:44246|issue:44273", @@ -14419,280 +14404,214 @@ "issue:44273|issue:44908", "issue:45230|issue:45310" ], - "prepared_review_unit_hash": "34db44a2d8f976b971a8d915f55b5c271bb56d25a8413c28783a647122522380", + "prepared_review_unit_hash": "77a47171854c84aabf4b7dba102d6edbf57c68a096abc230bd058f54e6ad7bdf", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39361, - "estimated_input_tokens": 9841, - "estimated_eval_tokens": 19938 + "serialized_chars": 39356, + "estimated_input_tokens": 9839, + "estimated_eval_tokens": 19934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous mix of Transformers v5 regression reports, model-specific bugs, and unrelated API/documentation issues. The soft links mostly reflect broad lexical similarity around \u201cv5 breakage\u201d rather than the same underlying defect.", - "confidence": 0.92, - "canonical_issue_reason": "issue:45850 is the broadest umbrella complaint in the set and best matches the cluster\u2019s overall theme of widespread v5 regressions, though the cluster is too mixed to treat as one concrete duplicate.", + "summary": "This is a loose mix of Transformers v5 regression/model-specific issues, not a tight duplicate cluster. The soft pairs mostly share broad keywords like v5, import, cache, or model names, but they describe different code paths and failure modes.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set (\u201cSince 5.0 version it breaks too many models\u201d), so it best matches the cluster\u2019s overall regression theme, even though it is very vague.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the most representative issue for the whole cluster because it captures the general v5 breakage theme and is broad enough to subsume several related reports.", + "best_issue_reason": "issue:45278 is the clearest actionable representative of the main theme: upgrade-to-v5 import breakage after moving from 4.57.0 to 5.5.0.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "Different bugs: Qwen3.5 chat crash vs a generic \u2018many models break\u2019 report." + "reason": "Specific Qwen3.5 chat crash vs a broad complaint about many models breaking; not the same concrete bug." }, { "left": "issue:44261", "right": "issue:44360", "accept": false, - "reason": "MLA epsilon precision issue and DSA indexer ReLU issue are unrelated code paths." + "reason": "MLA q_a_layernorm precision/config issue vs a DSA indexer ReLU issue; unrelated subsystems and symptoms." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Specific processor save_pretrained missing-files bug vs a generic placeholder bug report." + "reason": "processor.save_pretrained missing files is concrete; the other is just a generic bug report with no matching details." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading failure and TF backend/tensorboard error are distinct problems." + "reason": "Lazy loading failure vs undefined TensorFlow backend error; different features and code paths." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading bug vs unsafe sys.modules access in modeling_utils; not the same defect." + "reason": "Lazy loading behavior vs unsafe sys.modules access in modeling_utils; not the same bug." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Different Qwen3_5/VL symptoms: 2D box hallucination vs infinite loop on one image." + "reason": "2D bounding-box hallucination/error vs infinite loop on a specific image; different failure mode." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Same model family but different concrete issues: from_pretrained failure vs missing tp plan." + "reason": "Generic Qwen3-VL-Moe bug vs missing _tp_plan for tensor parallelism; related model family but different concrete issue." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken transformers env command and checkpoint reload regression are separate features." + "reason": "Broken `transformers env` command vs checkpoint reload incompatibility after upgrading; different functionality." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Model output bug vs PretrainedConfig type-checking breakage; unrelated." + "reason": "Qwen3-VL bounding-box output issue vs PretrainedConfig type-checking breakage; unrelated." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Llama4Config pad_token_id attribute bug is model-specific and not the generic v5 breakage report." + "reason": "Llama4Config pad_token_id AttributeError is a specific config bug, not the same as a broad v5 breakage complaint." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are upgrade-related, but one is checkpoint reload and the other is remote-code meta init incompatibility." + "reason": "Both mention v5 migration, but one is checkpoint reload failure and the other is remote-code/meta-init incompatibility; not the same code-path problem." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-star performance issue and chunked_prefill cache regression are different code paths." + "reason": "Import-star performance issue vs missing GenerationMixin cache during chunked_prefill; unrelated." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "Cache-strategy typo is a narrow docs/wording issue, not the same as broad model breakage." + "reason": "Typo in cache strategies vs vague broad regression claim; no shared concrete bug." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is distinct from the generic v5 complaint." + "reason": "StableLmConfig pad_token_idx missing is a specific config regression, not the same as a generic v5 breakage report." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and cache-strategy typo are unrelated." + "reason": "TF32 API settings migration vs a cache-strategy typo; unrelated issues." }, { - "left": "issue:42890", - "right": "issue:44908", + "left": "issue:45278", + "right": "issue:45850", "accept": false, - "reason": "Missing set_seed in tests and inverse_sqrt scheduler kwargs are unrelated issues." + "reason": "Both are broad v5 breakage reports, but one is import errors and the other is a generic \u201ctoo many models\u201d complaint; too vague to call the same duplicate." }, { - "left": "issue:45278", - "right": "issue:45850", + "left": "issue:42890", + "right": "issue:44908", "accept": false, - "reason": "Both mention broad v5 breakage, but one is import errors and the other is a generic \u2018many models\u2019 complaint; not one concrete bug." + "reason": "Missing set_seed in tests vs inverse_sqrt scheduler kwargs being ignored; different areas." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs forwarding and scheduler kwargs handling are different subsystems." + "reason": "AutoProcessor kwargs forwarding vs scheduler kwargs handling; unrelated code paths." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "sam3 tracker/type warning and Liger kernel application are unrelated." + "reason": "Model-type warning for sam3_tracker vs Liger Kernel not applied with model_init; unrelated." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Import latency and lazy loading correctness are different problems." + "reason": "Slow import sometimes vs lazy loading not working properly; related to imports only at a high level, but not the same bug." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility and Qwen3-VL bounding-box output are unrelated." + "reason": "Python version compatibility vs Qwen3-VL bounding-box output error; unrelated." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 support error and Qwen3-VL hallucination are different bugs." + "reason": "Flash Attention 2 support limitation vs Qwen3-VL output hallucination/error; different problem." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken env command and scheduler kwargs regression are unrelated." + "reason": "Broken env command vs scheduler kwargs ignored; no shared code path." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer class mismatch on save_pretrained is not the same as a generic bug report." + "reason": "tokenizer_class mismatch on save_pretrained vs a generic bug report; insufficient evidence of duplication." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs ignored and PretrainedConfig type-checking breakage are different defects." + "reason": "Scheduler kwargs bug vs PretrainedConfig type checking breakage; unrelated." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star slowdown and HybridCache import failure are unrelated." + "reason": "Import-star performance issue vs cannot import HybridCache; both involve imports, but different concrete failures." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Position_ids buffer access error is specific and unrelated to the generic bug report." - }, - { - "left": "issue:41669", - "right": "issue:45341", - "accept": false, - "reason": "Import-star performance issue and testing_utils typo/bug are unrelated." - }, - { - "left": "issue:44492", - "right": "issue:44945", - "accept": false, - "reason": "Cache strategy typo and pipeline parallelism output bug are different issues." - }, - { - "left": "issue:43874", - "right": "issue:45850", - "accept": false, - "reason": "GLM46V image-processor method missing is model-specific, not the generic v5 breakage report." - }, - { - "left": "issue:44393", - "right": "issue:45478", - "accept": false, - "reason": "Different Qwen3.5/VL bugs: bounding-box hallucination vs from_pretrained error." - }, - { - "left": "issue:44315", - "right": "issue:44829", - "accept": false, - "reason": "Liger-kernel application issue and flash_attention_3 training degeneration are unrelated." - }, - { - "left": "issue:44077", - "right": "issue:44861", - "accept": false, - "reason": "Optional post_init validation and tied-weight-key crash are distinct code paths." - }, - { - "left": "issue:43827", - "right": "issue:44908", - "accept": false, - "reason": "Docs still referencing pipeline() and scheduler kwargs handling are unrelated." - }, - { - "left": "issue:31795", - "right": "issue:33453", - "accept": false, - "reason": "Documentation confusion about input_ids/past_key_values is unrelated to tokenizer loading regression." - }, - { - "left": "issue:41669", - "right": "issue:44908", - "accept": false, - "reason": "Import-star slowdown and scheduler kwargs bug are unrelated." - }, - { - "left": "issue:43576", - "right": "issue:44273", - "accept": false, - "reason": "Broken env command and lazy loading failure are different features." - }, - { - "left": "issue:44655", - "right": "issue:44861", - "accept": false, - "reason": "Pipeline save_pretrained failure and tied-weight-key AttributeError are unrelated." + "reason": "position_ids buffer read error vs generic bug report; no basis to merge." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer ReLU complaint and Gemma4AudioRelPositionalEncoding bug are unrelated." + "reason": "DSA ReLU discussion vs Gemma4AudioRelPositionalEncoding issue; unrelated." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading bug and inverse_sqrt scheduler kwargs bug are unrelated." + "reason": "Lazy loading issue vs scheduler kwargs bug; different functionality." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report and Qwen3.5 MoE from_pretrained error are not the same issue." + "reason": "Generic bug report vs Qwen3.5 MoE from_pretrained error; too little overlap to treat as duplicates." } ] }, @@ -14787,154 +14706,154 @@ "issue:31515|issue:33453", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "c4f2abe84f2d7be12527413783f8821d7d12d9720800ac90bf9cc94eb99d60df", + "prepared_review_unit_hash": "4b66ea23f01b572d0b3cc11af7df8b8f2e45910b7564723019f4bc3956d1ab52", "component_budget": { "node_count": 393, "item_count": 393, "soft_pair_count": 1479, - "serialized_chars": 761577, - "estimated_input_tokens": 190395, - "estimated_eval_tokens": 381046 + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 }, "budget": { "node_count": 36, "item_count": 36, "soft_pair_count": 32, - "serialized_chars": 29798, - "estimated_input_tokens": 7450, - "estimated_eval_tokens": 15156 + "serialized_chars": 29795, + "estimated_input_tokens": 7449, + "estimated_eval_tokens": 15154 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly loose and heterogeneous. A few items orbit Transformer v5 loading/regression issues, but most soft pairs are different bugs or unrelated topics, so I would not merge any of them as duplicates.", - "confidence": 0.36, - "canonical_issue_reason": "issue:45092 is the strongest representative of the broad v5 compatibility/loading-breakage theme in this cluster (meta initialization incompatibility with older checkpoints). It is more representative than the very specific or unrelated issues, though the cluster is not a tight duplicate set.", + "summary": "This cluster is very heterogeneous: a few v5/tokenizer/model-loading regressions are superficially similar, but most items are unrelated bug reports. I did not find any soft pair that looks like the same concrete fixable bug, so no duplicate merges are justified.", + "confidence": 0.28, + "canonical_issue_reason": "issue:45850 is the broadest umbrella for the loose 'Transformers 5.0 breaks models' theme, but the cluster is too mixed for a strong canonical issue.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45092 is the best issue to anchor the cluster because it clearly captures a concrete and widely relevant Transformers v5 loading regression, which is the closest common thread among the more similar-looking issues.", + "best_issue_reason": "issue:45850 is the closest thing to a cluster-wide summary, but it is too vague to serve as a high-quality duplicate anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Different failures in different model components: Qwen3VL timestamp calculation vs GLM-5 RoPE implementation." + "reason": "Different model bugs: Qwen3VL timestamp calculation vs GLM-5 RoPE implementation." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Unrelated bugs: processor timestamp math vs scheduler kwargs handling." + "reason": "Unrelated code paths: processor timestamp logic vs inverse_sqrt scheduler kwargs handling." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import star cleanup/perf concern is not the same as a Qwen3ForCausalLM VRAM leak." + "reason": "One is about import-* cleanup/perf, the other is a VRAM leak in Qwen3ForCausalLM." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion about forward args is unrelated to a model loading regression." + "reason": "Documentation confusion vs a concrete model-loading regression." }, { "left": "issue:31795", "right": "issue:37428", "accept": false, - "reason": "Doc wording issue is not the same as an ImportError in flash attention utilities." + "reason": "Documentation issue vs ImportError from flash attention utils." }, { "left": "issue:44315", "right": "issue:45092", "accept": false, - "reason": "Both involve model initialization, but they are different concrete problems: Liger kernel application vs old remote-code meta-init incompatibility." + "reason": "Different failures: Liger kernel application in model_init vs remote-code/meta-init incompatibility." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Different model-specific implementation bugs." + "reason": "Different architectures and bugs: GLM-5 RoPE vs Gemma-4 audio positional encoding." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak under multi-threaded dataloading is unrelated to RoPE math." + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to GLM-5 RoPE behavior." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings migration is not the same bug as a memory leak." + "reason": "TF32 settings API change is unrelated to a dataloader-thread VRAM leak." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network access bug is unrelated to timestamp calculation." + "reason": "local_files_only network access bug is unrelated to Qwen3VL timestamp calculation." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Different processor bugs: timestamp computation vs save_pretrained missing files." + "reason": "Processor timestamp math and processor.save_pretrained missing files are different bugs." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Different model paths and symptoms." + "reason": "Timestamp calculation bug vs VRAM leak; no shared code-path problem." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return handling in one model is unrelated to Gemma-4 audio positional encoding." + "reason": "torch.split return-value issue in GlmMoeDsaIndexer is unrelated to Gemma audio positional encoding." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is a distinct issue from audio positional encoding math." + "reason": "CPU offload device mismatch is not the same bug as Gemma audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "Precision/config epsilon discussion is not the same as Gemma-4 audio positional encoding." + "reason": "MLA q_a_layernorm epsilon/config precision issue is unrelated to Gemma audio encoding." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer-v5 related, but one is broad behavior drift and the other is a specific decoder-type bug; not enough to treat as the same underlying change." + "reason": "Both mention tokenizer v5, but one is broad behavior regression and the other is decoder-type mismatch; not the same concrete bug." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug is unrelated to the non-technical unemployment issue." + "reason": "Scheduler kwargs bug is unrelated to the placeholder unemployment issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Different crashes in different code paths: tied weight key handling vs chat template tool-call handling." + "reason": "Tied-weights helper crash is unrelated to chat template tool-call handling." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Typo in cache strategies is unrelated to the unrelated/invalid unemployment issue." + "reason": "A typo in cache strategies is unrelated to the unemployment issue." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "One is invalid/off-topic, the other is a broad complaint about many v5 model breakages." + "reason": "Generic 'many models break' complaint does not match the unrelated unemployment report." }, { "left": "issue:43618", @@ -14946,55 +14865,67 @@ "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "No relation between the invalid unemployment issue and a testing_utils bug." + "reason": "Testing_utils bug and unemployment are unrelated." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id on Llama4Config is a separate config regression from tied weight key handling." + "reason": "Missing pad_token_id on Llama4Config is unrelated to _get_tied_weight_keys on list inputs." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is different from the tied-weight AttributeError." + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key handling." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading is not the same as an ImportError for layer_type_validation." + "reason": "Slow from_pretrained checkpoint loading is unrelated to a configuration_utils ImportError." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support request is unrelated to missing processor save files." + "reason": "FP8 support work is unrelated to processor.save_pretrained missing files." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on training failure is unrelated to slow from_pretrained checkpoint loading." + "reason": "MLFlow reporting after failed training and slow checkpoint loading are different problems." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowness is not the same as a tensor-parallel training bug." + "reason": "from_pretrained performance regression is unrelated to tensor-parallel training bugs." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "A docs clarification issue is unrelated to checkpoint loading performance." + "reason": "Loading slowdown and documentation confusion are not the same bug." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 support/inference request is different from a from_pretrained error on Qwen3.5 Moe." + "reason": "FP8 inference support request is unrelated to the Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowdown and tokenizer loading regression are different issues." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching audio-model support and qwen fp8 inference support are unrelated features/bugs." } ] }, @@ -15024,12 +14955,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "b8efcd7f07f942162804152f02dbaadc1797db2efec15b2f0c6f3ca1f3ff62b1", + "prepared_review_unit_hash": "21e3a71f0c0bd65c709cadf9728c56b6bb17e51d6de26c8457dde6dc6c8133d6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -15037,7 +14968,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -15046,18 +14977,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate feature requests in different areas: Flash Attention 2 support for more architectures vs GGUF support for more architectures. They share the \u201ccommunity contribution\u201d framing but not the same underlying change.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the requests target different features and code paths (Flash Attention 2 vs GGUF), so they are not duplicates.", + "summary": "These are two different community feature requests in the same general area (model/runtime support) but for unrelated backends: Flash Attention 2 vs GGUF. They should not be deduplicated.", + "confidence": 0.95, + "canonical_issue_reason": "Issue #26350 is the more central representative of this cluster because it has substantially higher discussion activity and is the older, broader feature request.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good duplicate canonical; they are distinct enhancement requests rather than the same problem.", + "best_issue_reason": "#26350 is the best issue representative here because it has the most engagement and the topic is a concrete support-extension request, but it is not a duplicate of #33260.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different feature requests affecting different subsystems; same broad theme of architecture support is too weak to treat as the same underlying issue." + "reason": "Different underlying changes: Flash Attention 2 support and GGUF support are separate architecture/backend requests with different code paths and implementation goals." } ] }, @@ -15094,80 +15025,77 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "53dfd80092d04b74820013db058a329d9e34729804d8c77e188731d35023584a", + "prepared_review_unit_hash": "ec7a7fe249462e7a0f339892d8b7cf27e3df5c7231b6e7b35397688de8e12f27", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4452, - "estimated_input_tokens": 1113, - "estimated_eval_tokens": 2482 + "serialized_chars": 4453, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4452, - "estimated_input_tokens": 1113, - "estimated_eval_tokens": 2482 + "serialized_chars": 4453, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues are near-identical placeholder/template duplicates of an i18n docs-translation request; the Turkish issue is concrete and should not be merged with the placeholder set.", - "confidence": 0.67, - "canonical_issue_reason": "issue:43834 is the oldest of the near-identical placeholder issues and best represents that duplicate subset; issue:27088 is a separate, concrete Turkish-specific request.", + "summary": "These issues all use the same documentation-translation template, but they refer to different languages, so they are not the same underlying task/bug. The Turkish issue is the only concrete one; the others look like placeholder or separate language requests.", + "confidence": 0.9, + "canonical_issue_reason": "issue:27088 is the only concrete, language-specific request ([i18n-TR] Turkish). The others are generic placeholders for different languages, so they should not be merged as duplicates of the Turkish task.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43834 has the most obviously duplicate title/body pattern among the placeholder issues and is the best dedupe target for that group.", + "best_issue_reason": "issue:27088 is the best representative because it is specific and actionable, unlike the placeholder/template issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "27088 is explicitly Turkish; 44370 is only a generic placeholder template, so they do not clearly describe the same underlying request." + "reason": "Same translation template, but not the same underlying request: Turkish vs another language placeholder." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "27088 is a concrete Turkish translation request, while 44365 is a generic placeholder issue with no concrete language target." + "reason": "Same template wording, but they represent different languages and therefore different tasks." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "The Turkish-specific issue and the generic placeholder issue are too different to merge as the same underlying task." + "reason": "Both are docs-translation requests, but the language target differs, so they are not duplicates." }, { "left": "issue:43834", "right": "issue:44370", - "accept": true, - "reason": "Same placeholder title pattern and nearly identical body size/content indicate the same underlying template issue." + "accept": false, + "reason": "Generic language placeholders; no evidence they refer to the same specific language or work item." }, { "left": "issue:43834", "right": "issue:44365", - "accept": true, - "reason": "These are effectively indistinguishable placeholder translation-request issues with the same title pattern and very similar bodies." + "accept": false, + "reason": "Template-matched requests only; different language placeholders, so not the same issue." }, { "left": "issue:44365", "right": "issue:44370", - "accept": true, - "reason": "Same generic i18n docs-translation template, with near-identical text and no distinguishing target details." + "accept": false, + "reason": "Both are placeholder translation issues, but for different languages, so they should remain separate." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: 43834/44365/44370 share the same placeholder title and extremely close body sizes, while 27088 is clearly the Turkish-specific request. The soft-edge rejections for 27088 are conservative, and the accepted placeholder-to-placeholder matches are supported by the strong textual similarity available here." - }, - "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 }, @@ -15237,7 +15165,7 @@ "issue:37015|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "e416590c57cf708eed69e954156aadd4b864ef1e92ceabfa0e82c7f22467187c", + "prepared_review_unit_hash": "e673687f895e0935c0f9afcf686ae3d52b17c7b25d21532482a32d08c4a52732", "component_budget": { "node_count": 15, "item_count": 15, @@ -15259,228 +15187,216 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are all model-request issues, but they span different concrete models and capabilities (vision, OCR, audio/TTS/ASR, BERT variants, sparse transformers). They are only loosely similar by theme, so none of the soft links should be treated as duplicates.", + "summary": "No duplicate cluster here: the items are all separate model-request issues spanning different models, modalities, and architectures. The high soft similarity is driven by templated wording, not the same underlying change.", "confidence": 0.97, - "canonical_issue_reason": "Issue 43837 is the best representative anchor because it is a concrete request with the highest apparent linkage density in this cluster, but the set is too heterogeneous for true deduplication.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43837 is the most central and representative item among these model-support requests, even though it is still specific to Qwen3-ASR rather than the whole cluster.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different requested models/architectures (EuroBert config vs DEIMv2), not the same change." + "reason": "Different model requests (EuroBert config vs DEIMv2 support); same request template only." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config and Qwen3-TTS are unrelated model requests in different modalities." + "reason": "EuroBert config and Qwen3-TTS are unrelated model additions." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "Different model families and domains (EuroBert vs RF-DETR)." + "reason": "EuroBert config and RF-DETR are different models and changes." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "Both are model-addition requests, but for distinct detection models and not one concrete duplicate." + "reason": "RF-DETR and DEIMv2 are distinct model requests." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR and LFM2.5 Audio target different model types and code paths." + "reason": "RF-DETR and LFM2.5 Audio are unrelated models/modalities." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Object detection model request vs sparse transformer request; not the same underlying change." + "reason": "RF-DETR and the weight-sparse transformer request are different underlying changes." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR and Qwen3-ASR are different model families with different functionality." + "reason": "RF-DETR and Qwen3-ASR are unrelated model additions." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR vs Qwen3-TTS are unrelated domains (vision detection vs speech synthesis)." + "reason": "RF-DETR and Qwen3-TTS are different models and tasks." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are different model families and capabilities." + "reason": "NeoBERT and Qwen3-ASR are unrelated model requests." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT request vs OpenAI sparse transformer request; different concrete additions." + "reason": "NeoBERT and the weight-sparse transformer request are different model families." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "BERT-style NLP model vs audio model; not the same change." + "reason": "NeoBERT and LFM2.5 Audio are unrelated additions." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are different models and different subsystems." + "reason": "RF-DETR and NeoBERT are different models and code paths." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are distinct model requests, not a duplicate." + "reason": "NeoBERT and ModernVBERT are distinct model additions." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support is different from Qwen3-ASR support." + "reason": "MiniCPM3 architecture support and Qwen3-ASR are unrelated requests." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 and the sparse transformer request are different model families." + "reason": "MiniCPM3 architecture support and the sparse transformer request are different changes." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are separate concrete model additions." + "reason": "MiniCPM3 and ModernVBERT are distinct model requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS are different model types and tasks." + "reason": "MiniCPM3 architecture support and Voxtral TTS are unrelated models." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + "reason": "DEIMv2 and Qwen3-TTS are different model additions." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are different concrete model requests." + "reason": "ModernVBERT and Qwen3-ASR are unrelated model requests." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are different models and modalities." + "reason": "ModernVBERT and Qwen3-TTS are different models/tasks." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "A specific model request (ModernVBERT) is not the same as a broad RoPE-family request." + "reason": "ModernVBERT and BERT-like models with RoPE are not the same change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "Broad BERT-like RoPE discussion vs Voxtral TTS; different change scopes." + "reason": "Generic RoPE support for BERT-like models and Voxtral TTS are unrelated." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are related but distinct capabilities; not one mergeable duplicate." + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct modalities and requests." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Different audio models and different tasks (TTS vs Voxtral TTS implementation request)." + "reason": "Qwen3-TTS and Voxtral TTS are different models; not one code-path fix." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Speech synthesis request vs DeepSeek-OCR2; unrelated model additions." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS and the sparse transformer request are different model families." + "reason": "Qwen3-TTS and the sparse transformer request are different underlying changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model requests, not the same code path." + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model additions." }, { "left": "issue:43837", - "right": "issue:43909", + "right": "issue:44205", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are both audio-related but request different models/capabilities." + "reason": "Qwen3-ASR and SAM3-LiteText are unrelated models." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse transformer support and EUPE are different model requests." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Speech recognition request vs EUPE perception-encoder request; different concrete changes." + "reason": "Qwen3-ASR and EUPE are unrelated additions." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support and OpenAI sparse transformer support are unrelated additions." + "reason": "Qwen3-ASR and the sparse transformer request are different model families." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "ASR and TTS are different capabilities; these are not the same underlying fix." + "reason": "Qwen3-ASR and Voxtral TTS are different models and modalities." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS are different model families and tasks." + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and code paths." - }, - { - "left": "issue:43909", - "right": "issue:44205", - "accept": false, - "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are unrelated model requests." - }, - { - "left": "issue:44121", - "right": "issue:45322", - "accept": false, - "reason": "OpenAI weight-sparse transformer and EUPE are distinct model proposals." - }, - { - "left": "issue:43837", - "right": "issue:45027", - "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different speech tasks, not one duplicate change." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model additions." } ] }, @@ -15521,12 +15437,12 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "4b24148a5f3108808b4f3d4a0c5df606864abbe4ee76a6641e140aead1e10c5f", + "prepared_review_unit_hash": "90fda2da84ff3715f67316d45749179f936f3ac170881cc4e0ad93361a1b1414", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6290, + "serialized_chars": 6291, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -15534,7 +15450,7 @@ "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6290, + "serialized_chars": 6291, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -15543,60 +15459,60 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates of one another; the cluster is a loose set of unrelated bug reports and typo/config issues across different models and tokenizer behavior.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the items span unrelated problems (norm implementation, ignore_value handling, kwargs propagation, type hints, typo, tokenizer warning).", + "summary": "These issues only share superficial \u201csmall bug in model/config handling\u201d similarity; they describe different models, parameters, and failure modes, so none of the soft links look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43668 is the most broadly representative of the cluster\u2019s loose theme (a concrete API/config correctness bug), but the set is still too heterogeneous to justify a stronger canonical duplicate grouping.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue: none of these issues is a plausible representative for the others because they do not describe the same underlying bug.", + "best_issue_reason": "issue:43668 is the clearest, most self-contained problem statement and the most central by similarity, even though it should not absorb the others as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Different problem types and model areas: Llama4 norm implementation vs ModernBERT config type hint." + "reason": "Different models and different bugs: Llama4TextL2Norm vs ModernBERTConfig type hint; only generic norm-related wording overlaps." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value handling is unrelated to passing kwargs into a loss function." + "reason": "One is Mask2Former ignore_value not being used; the other is kwargs not forwarded to a loss function. Same broad area, different code paths." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Different domains: Mask2Former loss handling vs tokenizer regex warning regression." + "reason": "Mask2Former ignore_value handling is unrelated to tokenizer regex warning behavior." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Unrelated issues: Mask2Former model behavior vs a typo in a quantization attribute name." + "reason": "Different subsystems and failure modes; a typo in one quantization attribute is not the same as an unused ignore_value setting." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "No shared underlying bug; one concerns model loss inputs, the other a config annotation." + "reason": "Both are parameter-related bugs, but they affect different models and different code paths; not the same underlying issue." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and a config type hint bug are distinct and not mergeable into one fix." + "reason": "Forwarding kwargs to loss calculation is unrelated to ModernBERTConfig norm_eps typing." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Different symptoms and code paths: config typing vs tokenizer regex warning." + "reason": "Incorrect type hint in a model config and tokenizer regex warnings are unrelated problems." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "A type-hint correction and a typo in a quantization attribute are separate issues." + "reason": "A config type-hint typo is not the same bug as a quantization attribute typo." } ] }, @@ -15626,7 +15542,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "988e716cded2fb633e137b3a6d502c7bc5d59d1b707fc9eac52286da92680d33", + "prepared_review_unit_hash": "d7537a911c297ab83e9b32b938ff362639e2ad3647a9d572e95c93d8a308a23c", "component_budget": { "node_count": 2, "item_count": 2, @@ -15648,18 +15564,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are about different requests: one is a decoding behavior change, the other is support for a specific model version. They do not look like duplicates or the same underlying bug.", - "confidence": 0.98, - "canonical_issue_reason": "No clear canonical issue: 'Set Block Decoding' and 'DeepSeek v3.2 support' describe different feature requests in different parts of the product.", + "summary": "The two issues are not duplicates: one is about block decoding mechanics, the other is a request for DeepSeek v3.2 support. They refer to different user problems and different likely code paths.", + "confidence": 0.97, + "canonical_issue_reason": "If a representative must be chosen, issue #41084 is the more concrete technical issue; #43037 is a broad model-support request and does not overlap in substance.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a strong duplicate anchor for the other; they should be tracked separately.", + "best_issue_reason": "Issue #41084 is the better cluster representative because it describes a specific decoding feature/problem rather than a generic support request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying requests: block decoding configuration vs model support for DeepSeek v3.2. The titles and implied scopes do not match closely enough to treat as the same issue." + "reason": "Different topics and problem domains: block decoding vs DeepSeek v3.2 support. No evidence they describe the same underlying bug or change." } ] }, @@ -15689,12 +15605,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "9cec6a5d7d1e0673526a9951ec25bb8afe89891503b957d5a78f7e669451c447", + "prepared_review_unit_hash": "0fb68af1bb2a9b173c9617c778fa1891a4e9cfa52804cd03f448444d66eff91f", "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 }, @@ -15702,7 +15618,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -15711,18 +15627,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are related only at a very high level (custom trainers/Trainer extensibility), but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPs estimation for custom trainers. They should not be clustered as duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "#43598 is the broader Trainer extensibility issue and is more likely to subsume a general fix around custom trainer behavior.", + "summary": "These are related to Trainer custom behavior, but they describe different problems: one is about `remove_unused_column` configurability, the other is about FLOPS estimation for custom trainers. They are not the same underlying bug.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue fits well here because the two issues target different code paths and user problems.", "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the better representative issue because it targets a core Trainer customization path and is more general than the FLOPs-estimation-specific report.", + "best_issue_reason": "Neither issue is a true duplicate; if forced to pick a representative, `Revisit remove_unused_column in Trainer for better customizability` is the clearer broader Trainer customization request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying bugs and code paths: one concerns column removal/customizability, the other FLOPs accounting for custom trainers. Shared theme is too broad to treat as the same issue." + "reason": "Both mention custom trainers, but one concerns unused-column handling and the other concerns FLOPS estimation. Different behavior, different fix." } ] }, @@ -15752,7 +15668,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "16e12f63f3552b614a51d2daa312744ca5e5811716516272939c532ed72d04f6", + "prepared_review_unit_hash": "40e0d97fb77b8a0b2489c3aa8f9267e07315879f51947c1e7d0b79ad531e428d", "component_budget": { "node_count": 2, "item_count": 2, @@ -15774,18 +15690,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are model-support requests in the Qwen family, but they target different concrete additions: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration versus Qwen3.5 sequence-classification support. They are related, but not the same underlying change.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43684 is the broader API registration request and is the more general representative of the cluster, but it does not subsume the separate Qwen3.5 sequence-classification request.", + "summary": "These are related only at a very high model-family level, but they request different capabilities for different Qwen variants: one is about registering Qwen3-Omni with generic auto classes, the other is about adding sequence classification support for Qwen3.5. Not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more central request: it targets AutoModel/AutoModelForConditionalGeneration registration for Qwen3-Omni, which is a foundational compatibility change.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the best single issue to anchor this set because it asks for a core model-registration capability, making it the most broadly representative support request.", + "best_issue_reason": "#43684 is the better canonical issue because it describes a general registration gap rather than a narrower task-specific head support request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different model families/tasks and different code changes: Qwen3-Omni AutoModel/AutoModelForConditionalGeneration registration is not the same as Qwen3.5 AutoModelForSequenceClassification support." + "reason": "Different underlying changes: model registration for Qwen3-Omni auto classes versus sequence classification support for Qwen3.5. They concern different model variants and different code paths, so they should not be merged as duplicates." } ] }, @@ -15822,76 +15738,76 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "4136e76a0afa3794ba7ad7da107aa4d272aec044dd9b8ce81dba3ead925a7835", + "prepared_review_unit_hash": "15efe102201ef0e23434d739dcf5af12fba357aecb051706e79084645039133c", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4260, - "estimated_input_tokens": 1065, - "estimated_eval_tokens": 2386 + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4260, - "estimated_input_tokens": 1065, - "estimated_eval_tokens": 2386 + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues share the exact \"huggingface\" title and look like duplicate reports; \"Fett\" appears unrelated and should stay separate.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 43845 is the earliest of the repeated \"huggingface\" reports and matches the same title pattern as the other likely duplicates.", + "summary": "Three closed issues share the exact title \"huggingface \" and look like duplicates; issue #43835 titled \"Fett\" appears unrelated. I would keep the earliest huggingface report as the canonical issue.", + "confidence": 0.84, + "canonical_issue_reason": "Issue #43845 is the earliest of the three near-identical \"huggingface \" reports, so it is the best canonical issue among the likely duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the best representative because it is earliest among the likely duplicate set and has no evidence suggesting it is distinct from 43846/43859.", + "best_issue_reason": "No richer issue content is available, but #43845 is the earliest clearly matching report and is the best single issue to retain for the duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different title and no shared issue target or body evidence; likely unrelated." + "reason": "Different titles and no evidence they describe the same problem; only weak similarity." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different title and no concrete evidence they describe the same bug." + "reason": "Different titles and no shared target or file evidence; likely unrelated." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different title and no concrete evidence they describe the same bug." + "reason": "\"Fett\" does not match the \"huggingface\" reports and there is no substantive overlap." }, { "left": "issue:43845", "right": "issue:43846", "accept": true, - "reason": "Exact same title and very similar metadata; likely duplicate reports of the same issue." + "reason": "Same exact title and similar minimal metadata; these look like duplicate reports of the same issue." }, { "left": "issue:43845", "right": "issue:43859", "accept": true, - "reason": "Exact same title as the other \"huggingface\" issues, so this likely refers to the same report." + "reason": "Same exact title and same apparent topic; likely another duplicate report." }, { "left": "issue:43846", "right": "issue:43859", "accept": true, - "reason": "Exact same title and no distinguishing evidence; likely the same underlying issue." + "reason": "Same exact title and no distinguishing evidence; likely the same underlying issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall: 43835 is distinct by title, and 43845/43846/43859 share the exact title with only soft-similarity evidence, so treating them as likely duplicates is conservative enough. The canonical-choice rationale is also supported by the earliest timestamp among that cluster." + "feedback": "The summary is grounded in the packet: #43845, #43846, and #43859 share the exact title \"huggingface \" and are the only plausible duplicate cluster, while #43835 has a different title and no evidence tying it in. The canonical choice of #43845 is conservative because it is simply the earliest of the matching trio. The pairwise verdicts are also suitably cautious, relying only on identical titles and the absence of distinguishing evidence." }, "evaluator_used": true, "retried": false, @@ -15929,88 +15845,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "f80930aa0d0db318047521004b221fa35ff7e8db8c05cf25e6e0c96e127b9dd6", + "prepared_review_unit_hash": "13c6327423b956499c64887bbae323eec5630302f837b284486869ae99214b03", "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": 5779, + "estimated_input_tokens": 1445, + "estimated_eval_tokens": 3146 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5779, + "estimated_input_tokens": 1445, + "estimated_eval_tokens": 3146 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are low-signal, likely spam or placeholder issues with no substantive content. None look like the same underlying bug or request, so I reject all soft duplicate links.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "This cluster looks like unrelated low-information issue noise/spam, not true duplicates. There isn\u2019t enough substance to merge any pair as the same underlying bug or request.", + "confidence": 0.58, + "canonical_issue_reason": "issue:44418 is the only item with a non-placeholder title, so it is the most usable anchor if one must be chosen; the others are generic spam-like placeholders.", "canonical_pr_reason": null, - "best_issue_reason": "No issue here is a good canonical duplicate target; issue 44418 is the only one with a non-spam title, but it is still too vague to represent a real cluster.", + "best_issue_reason": "issue:44418 is the most descriptive issue in an otherwise content-poor cluster, but it should still be treated cautiously because the cluster does not show a shared concrete problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are generic spam placeholders with no body detail; no evidence of the same bug or change." + "reason": "Same placeholder-style title, but no body/context to show a shared underlying bug or change." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are generic spam placeholders with no body detail; not the same underlying issue." + "reason": "Identical spam-like titles are not enough to establish the same issue; no concrete evidence ties them together." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam, the other is a vague documentation request; these are not the same report." + "reason": "One is a placeholder/spam-like issue and the other is a generic documentation suggestion; no clear shared problem." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. unrelated title about \"Google LLC\"; no shared concrete problem." + "reason": "Different titles and no supporting details; cannot treat them as the same issue." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A vague documentation suggestion is unrelated to a spam placeholder; no duplicate signal." + "reason": "Generic documentation suggestion versus placeholder spam; no evidence of the same underlying topic." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "A vague documentation suggestion is unrelated to a spam placeholder; not duplicates." + "reason": "Generic documentation suggestion versus placeholder spam; too little information to merge." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but there is no concrete bug/change description tying them to the same underlying issue." + "reason": "Same placeholder title, but identical titles alone are insufficient without any issue content showing the same bug/request." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. unrelated \"Google LLC\" title; no evidence they refer to the same problem." + "reason": "No shared concrete subject matter is visible; the titles do not indicate the same underlying issue." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. unrelated \"Google LLC\" title; no shared underlying bug or request." + "reason": "No concrete overlap beyond being closed issues; not enough to call them duplicates." } ] }, @@ -16040,12 +15956,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "2ec4ff2a77b7b7af38e78015e7835051d58ba19caf98c0e28699e854add2fa9f", + "prepared_review_unit_hash": "9eb1b482dc991bc2be0dc9c26039713098674912f853eefdefe756a6dd20ee80", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -16053,7 +15969,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -16062,28 +15978,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues look like a probable duplicate pair: they have the same title and highly similar bodies, but the only explicit pair signal is soft_similarity and there is no deterministic confirmation, so the match should be treated as likely rather than certain.", - "confidence": 0.86, - "canonical_issue_reason": "Issue #45145 is a reasonable canonical anchor because it has the same title/body as #45151 and slightly more inbound reference activity, but the duplicate call is only probabilistic.", + "summary": "The pair is only weakly similar: same title and comparable body length, but no direct body comparison or other hard signals. With deterministic_accept=false, this should be treated conservatively as a non-duplicate soft match rather than an accepted duplicate.", + "confidence": 0.87, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #45145 is the best issue to keep as the canonical representative since it appears first and has the small edge in linkage/activity, though the evidence does not make the duplication fully deterministic.", + "best_issue_reason": "Issue #45145 is the best single representative if one must be chosen: it is earlier and has the only inbound reference, but this does not by itself justify merging the pair as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": true, - "reason": "Probable duplicate: same title and highly similar long-form body, with no distinguishing targets or code artifacts. The evidence supports merging them as the same report, but only with moderate confidence rather than certainty." + "accept": false, + "reason": "Identical title alone plus soft similarity is not enough here; there is no body-level comparison, target, filename, or other deterministic evidence that these are the same underlying issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "The packet supports a likely duplicate pair, but the analyst overstates the certainty. The two issues share the same title and have highly similar bodies, yet the only pair evidence is soft_similarity and the conservative deterministic flag is false. Phrases like \"same underlying issue\" and \"effectively duplicates\" are stronger than the packet warrants. Please downgrade the verdict or qualify it as a probable duplicate rather than a definite one." + "feedback": "The duplicate call is too strong for the packet. The evidence only supports soft similarity: identical title and similar body length, but no direct body comparison, links, targets, filenames, or deterministic accept signal. The summary also overstates by saying the bodies are likely the same and that there are no distinguishing signals. Canonicalization to #45145 is fine only if the pair itself is accepted, but here the soft-edge verdict should remain conservative." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -16160,40 +16076,40 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "ba5080f4747b33c35dd36dc04246b03a1c309547ec9689f54e7fb9034fbbeb7e", + "prepared_review_unit_hash": "8bb320c344d835e1306d73607cde92e67523e8eff5967b31517c17f3b666f9a0", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two different 4-bit inference bugs: one is a Gemma4 audio-module uint8/torch.finfo TypeError, the other is a BitsAndBytesConfig skip-module handling regression that removes the default lm_head exclusion and triggers an AssertionError. They are related only at a broad quantization level, not the same underlying issue.", - "confidence": 0.92, - "canonical_issue_reason": "Issue #45674 is the better canonical issue because it is more general, has clearer configuration-level root cause, and appears to have more discussion/inbound references.", + "summary": "These are two different 4-bit inference bugs: one is a Gemma4/audio-module uint8 dtype TypeError from `torch.finfo()`, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an AssertionError. They are related only at a broad subsystem level, not the same underlying issue.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #45674 is the better anchor because it describes a specific, reusable configuration regression with a clear failure mode and reproduction path. #45672 appears to be a separate Gemma4/audio-module dtype bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the best representative of the cluster since it describes a broader, actionable 4-bit inference regression and is likely to attract the right fix/triage attention.", + "best_issue_reason": "#45674 is the strongest representative issue in this cluster because it is narrowly defined and actionable; however, the cluster itself looks heterogeneous rather than duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both concern quantized inference, but the failure modes differ: a uint8/torch.finfo TypeError in Gemma4 audio modules versus an llm_int8_skip_modules bug clearing lm_head exclusion and causing an AssertionError. They do not look like the same underlying bug." + "reason": "Different root causes and failure paths: #45672 is a Gemma4 audio-module `torch.finfo()` TypeError on uint8 weights, while #45674 is a BitsAndBytesConfig `llm_int8_skip_modules` / `lm_head` exclusion regression causing AssertionError. Same broad area, but not the same bug." } ] }, @@ -16223,7 +16139,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "d2d5685e3893caf5423ef26f9a68190d09d6fff57f508fa462ee050bad4038ff", + "prepared_review_unit_hash": "313c85982daf129cf286dc5af6243be86df942233c1215e9ee6683019e0ae5ed", "component_budget": { "node_count": 2, "item_count": 2, @@ -16245,18 +16161,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues only share a loose textual similarity in the titles; the names and dates differ and there is no body, linkage, or shared context indicating the same underlying bug or task.", - "confidence": 0.97, + "summary": "Two closed issues with empty bodies and unrelated personal-name titles; they do not appear to describe the same underlying bug or change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative: both are low-signal, unrelated entries with no description, targets, or evidence of a shared problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Reject: the issues appear to be distinct personal-name/date entries with no evidence of a shared bug, request, or duplicate context." + "reason": "Reject: the titles refer to different people and dates, and there is no body or other evidence indicating the same underlying issue." } ] }, @@ -16286,7 +16202,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "360b47d594381223b95296ffdefc5c1f4ed2d06180c3365f83f093fbee818a39", + "prepared_review_unit_hash": "39e5d44a72d7ed6f82f415b83f1ee459873bc381a21267e8f4b558e71fe31d2a", "component_budget": { "node_count": 2, "item_count": 2, @@ -16308,8 +16224,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are both just bare personal-name-and-date titles with no body or technical content, and there is no clear evidence they describe the same bug or request. I would not cluster them as duplicates.", - "confidence": 0.94, + "summary": "These two closed issues only share a loose text similarity in their titles and do not describe the same bug, request, or code change. They should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -16319,7 +16235,7 @@ "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Only superficial similarity in format (a name plus date); no shared problem statement, code path, or concrete duplicate signal." + "reason": "Unrelated personal-name/date titles with no shared problem statement, code context, or issue target; not the same underlying bug or change." } ] }, @@ -16359,91 +16275,91 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "ea22fba3dd65336de89ca1d09e67ae78eddfb3ff6b9fa93fe81d425570671fbe", + "prepared_review_unit_hash": "203baefa561837a9b95002ae37334d4facbfe89f7517ecbfa88cacd8f9f60b84", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5729, - "estimated_input_tokens": 1433, - "estimated_eval_tokens": 3122 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5729, - "estimated_input_tokens": 1433, - "estimated_eval_tokens": 3122 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The set splits into two highly likely duplicate issue pairs, 45870/45871 and 45872/45873, while 45869 appears separate and should remain distinct.", - "confidence": 0.89, - "canonical_issue_reason": "45870 is the best canonical representative for the highly likely duplicate pair with 45871 because it was created earlier, and the two issues have identical titles and matching body lengths.", + "summary": "The cluster contains two clear duplicate pairs: 45870/45871 and 45872/45873. Issue 45869 is about a dataset offering and does not match the example/scraping-service requests, so it should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue covers the whole cluster; the items split into two separate duplicate groups. If one representative must be chosen, 45872 is the strongest because it is the more specific, concrete request and has an exact duplicate in 45873.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the best overall issue representative because it anchors the earliest highly similar pair in the cluster, while 45869 looks separate and 45872/45873 form a different duplicate pair.", + "best_issue_reason": "45872 is the best representative issue overall: it is specific, concrete, and clearly has an exact duplicate partner (45873).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different request types: a dataset bundle/sample listing versus an example for financial time series prediction." + "reason": "Different requests: dataset bundle advertisement vs a financial time-series example. Same general finance theme, but not the same underlying issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different request types: a dataset bundle/sample listing versus an example for financial time series prediction." + "reason": "Different requests: dataset bundle advertisement vs a financial time-series example. Not the same bug/change." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Highly likely duplicate: identical title and matching body length, with close timestamps." + "reason": "Same title and same example request; these are duplicates describing the same underlying issue." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Same broad finance/DataForge theme, but the requested change is different: prediction example versus scraping-service request." + "reason": "Both are finance-related examples, but 45872 adds a scraping-service component and is a different request from the example issue." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same broad finance/DataForge theme, but the requested change is different: prediction example versus scraping-service request." + "reason": "Different request content: example time-series prediction vs time-series plus scraping service. Related topic, not the same issue." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Same broad finance/DataForge theme, but the requested change is different: prediction example versus scraping-service request." + "reason": "Same reasoning as 45870/45872: overlapping finance theme, but different concrete request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Same broad finance/DataForge theme, but the requested change is different: prediction example versus scraping-service request." + "reason": "Different request content: example time-series prediction vs time-series plus scraping service. Not a duplicate." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Highly likely duplicate: identical title and matching body length, with close timestamps." + "reason": "Exact title match and matching request content; these are duplicates of the same underlying issue." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The overall grouping is plausible, but several statements overstate the packet evidence. The data supports strong similarity for 45870/45871 and 45872/45873, but not definitive \"exact duplicate\" claims or that 45870 is the \"cleaner\" canonical copy. Also, describing 45869 as a \"dataset-promo request\" goes a bit beyond the metadata shown. Please soften the duplicate language to \"highly likely\" and keep the canonical rationale grounded in the earlier timestamp/title/body-length match only." + "accept": true, + "feedback": "The summary is well grounded in the packet: 45870/45871 and 45872/45873 are the only clearly identical pairs by title/body similarity, and 45869 is visibly different in topic. The soft-edge rejections stay conservative, and the overall \u2018no single canonical issue\u2019 conclusion fits the split cluster structure." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -16451,7 +16367,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16459,7 +16375,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-39895-14", + "cluster_id": "cluster-39895-13", "nodes": [ "pull_request:39895", "pull_request:41855", @@ -16473,20 +16389,18 @@ "pull_request:45170", "pull_request:45213", "pull_request:45655", - "pull_request:45795", - "pull_request:45884" + "pull_request:45795" ], "soft_pairs": [ "pull_request:39895|pull_request:43424", - "pull_request:41855|pull_request:43424", "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:45884", "pull_request:43823|pull_request:43995", "pull_request:43995|pull_request:44827", "pull_request:43995|pull_request:45213", @@ -16498,160 +16412,154 @@ "pull_request:43823|pull_request:45795", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "ef225cc2f66010105bfe5c23594d2a0cbbca964325de9ebe07b8be30d5ec32ed", + "prepared_review_unit_hash": "a5980b02fed89ceb74128a8df1753ec43bb8a1abd42b4998de4dd3a887294399", "component_budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 20, - "serialized_chars": 37461, - "estimated_input_tokens": 9366, - "estimated_eval_tokens": 18988 + "node_count": 13, + "item_count": 13, + "soft_pair_count": 19, + "serialized_chars": 35206, + "estimated_input_tokens": 8802, + "estimated_eval_tokens": 17860 }, "budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 20, - "serialized_chars": 37461, - "estimated_input_tokens": 9366, - "estimated_eval_tokens": 18988 + "node_count": 13, + "item_count": 13, + "soft_pair_count": 19, + "serialized_chars": 35206, + "estimated_input_tokens": 8802, + "estimated_eval_tokens": 17860 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly unrelated PRs grouped by broad lexical similarity; only the two `fixed_cross_entropy` PRs are the same underlying change and should be consolidated.", - "confidence": 0.96, + "summary": "This cluster is mostly a set of unrelated PRs across docs, new model adds, and small refactors. The only true duplicate pair is the two `fixed_cross_entropy` PRs that both add `weight`/`label_smoothing` support for the same helper in `loss_utils.py` and target the same issue. Everything else is only superficially similar by subsystem or broad theme.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the better canonical representative: it explicitly links issue 43240 and makes the underlying `fixed_cross_entropy` API fix in one place, while 43254 is essentially the same change with a looser title.", + "canonical_pr_reason": "PR 43251 is the best canonical PR for the loss-utils duplicate pair: it directly targets issue 43240, makes the same concrete `fixed_cross_entropy` change, and is the more explicit fix rather than the alternative phrasing in 43254.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the strongest cluster representative because it directly addresses the tracked bug, has the clearest intent, and overlaps almost exactly with the competing PR in the same code path.", + "best_pr_reason": "PR 43251 is the strongest representative artifact in this cluster because it has the clearest problem statement, explicit issue linkage, and the most directly scoped code change. The rest of the cluster does not form a single cohesive fix.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated model-addition vs export test; only broad 'adding features' similarity, not the same bug/change." + "reason": "Both are unrelated feature/test PRs; they do not fix the same code-path problem and cannot plausibly be merged as one change." }, { "left": "pull_request:41855", - "right": "pull_request:43424", + "right": "pull_request:43823", "accept": false, - "reason": "Tokenizer utility fix and executorch export test are different code paths with no concrete overlap." + "reason": "One is tokenizer API cleanup, the other is adding a new model; shared model-adjacent wording is not enough for duplication." }, { "left": "pull_request:41855", - "right": "pull_request:43823", + "right": "pull_request:43424", "accept": false, - "reason": "Missing tokenizer methods vs adding a new model; not the same underlying change." + "reason": "A tokenizer method addition and an executorch export test are unrelated changes." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Both are processor/model infrastructure work, but they target different models and different problems." + "reason": "Both touch processor/model auto plumbing in a broad sense, but they address different models and different bugs." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy kwargs fix is unrelated to adding MobileLLM." + "reason": "Loss helper kwargs support and adding a new model are unrelated." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss utility change and Mistral4 test/model updates are distinct fixes." + "reason": "These are different fixes in different areas; the overlap is only that both are small code changes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch export tests and Mistral4 changes do not share a concrete bug." + "reason": "An executorch dynamic-shape test and Mistral4 test fixes are not the same underlying change." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Different models and different functionality; similarity is only generic PR scope." + "reason": "A test for exportability and a new model addition are not duplicates." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New model implementation vs class annotation-order fix; unrelated code paths." - }, - { - "left": "pull_request:43823", - "right": "pull_request:45884", - "accept": false, - "reason": "New model implementation vs masking primitive cleanup; no shared underlying change." + "reason": "A new model implementation and a class-annotation resolution fix are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New model addition and Falcon refactor are not the same change." + "reason": "Adding MobileLLM and refactoring Falcon output collection are different model work, not one bug fix." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor and Mistral4 test fixes are different model-specific efforts." + "reason": "Both touch model internals, but they fix different models and different code paths." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor is unrelated to the model-creation skill PR." + "reason": "Falcon refactoring and a model-creation skill are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 fixes and layrnorm/layernorm renaming are different bugs." + "reason": "Mistral4 test fixes and a CLIP-like weight renaming fix are different changes with no shared concrete defect." }, { "left": "pull_request:44827", "right": "pull_request:45795", "accept": false, - "reason": "Mistral4 test/model work is unrelated to docs for audio/video processors." + "reason": "Mistral4 fixes and documentation for audio/video processors are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 changes and config-resolution order fix are unrelated." + "reason": "Different bugs in different subsystems; no evidence they are the same underlying issue." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 work and model-creation skill are not the same underlying change." + "reason": "A model test fix and a draft model-creation skill are not the same change." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition and the model-creation skill draft are not duplicates." + "reason": "Adding MobileLLM is not the same as creating an internal model-creation skill." }, { "left": "pull_request:43823", "right": "pull_request:45795", "accept": false, - "reason": "MobileLLM model addition vs docs for audio/video processors is unrelated." + "reason": "New model code versus documentation updates for processors are not duplicates." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs patch `src/transformers/loss/loss_utils.py` to add supported kwargs to `fixed_cross_entropy` for issue 43240; they fix the same concrete function and could be merged into one." + "reason": "Both PRs modify `src/transformers/loss/loss_utils.py` to pass the same extra `cross_entropy` kwargs (`weight` and `label_smoothing`) through `fixed_cross_entropy`; they fix the same concrete bug and are merge-equivalent." } ] }, "evaluator_result": { "accept": true, - "feedback": "The judgments are conservative and grounded in the packet. The only accepted pair, PR 43251 and PR 43254, clearly touch the same file (`src/transformers/loss/loss_utils.py`) and the same issue target (43240), so they look mergeable into one bugfix PR. The remaining pairwise rejects stay within what the titles/diff previews support and do not overstate overlap." + "feedback": "The summary is grounded in the packet and stays conservative. The lone accepted duplicate pair (43251/43254) is well supported by the shared `loss_utils.py` edit and shared issue target 43240, and the other soft-edge rejections do not overstate similarity. The cluster characterization as mostly unrelated PRs is also supported by the listed diffs." }, "evaluator_used": true, "retried": false, @@ -16662,7 +16570,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16678,7 +16586,7 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "e68719cb8967f73201a54babc3e2378c30ce5a3bfe489fd5d468df1618f1df07", + "prepared_review_unit_hash": "7f37e71834fcb93ae91f3ab81798b0252fabb5dfea0cefccce65ad1df0abd942", "component_budget": { "node_count": 2, "item_count": 2, @@ -16700,18 +16608,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about VibeVoice, but they target different scopes: the first implements the base VibeVoice integration, while the second is a draft adding VibeVoice Realtime and the acoustic tokenizer. They overlap on shared framework plumbing, not the same concrete change.", - "confidence": 0.89, + "summary": "These are related VibeVoice PRs, but they are not the same change: one adds the base VibeVoice integration, while the other adds VibeVoice Realtime and additional tokenizer/feature-extraction pieces. Shared filenames reflect the same subsystem, not duplicate work.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the core, non-draft base VibeVoice implementation and is the more central representative of this cluster.", + "canonical_pr_reason": "PR 40546 is the broader foundational VibeVoice implementation and is the better representative of the cluster than the later realtime-focused extension.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 is the strongest standalone candidate because it is the primary full model integration with substantial review activity, whereas PR 42781 is a draft extension for realtime/tokenizer support.", + "best_pr_reason": "PR 40546 is the best single PR to represent the cluster because it introduces the core model integration, docs, and auto-registry wiring.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Shared VibeVoice files and auto-registry updates do not mean the PRs fix the same bug/change; one adds the base model, the other extends it for realtime and acoustic tokenizer support. They would not plausibly merge as one PR without mixing distinct features." + "reason": "Different underlying change: 40546 implements VibeVoice, while 42781 adds VibeVoice Realtime and related tokenizer/feature-extraction support. They touch overlapping files but not the same concrete bug or mergeable patch." } ] }, @@ -16725,7 +16633,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16741,7 +16649,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "947620b71b419823cb7cbec5968a1139ea7c48654fa27566c2a9832ecf57ca56", + "prepared_review_unit_hash": "4ad4ae34e24f6dee3e60e33f812b00bff48f542c737e2be1fd1d6cb667559c01", "component_budget": { "node_count": 2, "item_count": 2, @@ -16763,18 +16671,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are unrelated: one adds TimesFM covariate forecasting support, while the other adds DINOv3 image classification support. They do not appear to fix the same code path or be mergeable into a single PR.", - "confidence": 0.99, + "summary": "The two pull requests are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They should not be deduplicated or clustered as the same change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR is appropriate because the PRs target different models and different features in separate code paths.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Neither PR is a better duplicate representative; both are distinct feature additions with no meaningful overlap beyond being documentation/model updates.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different subsystems and different user-facing changes: TimesFM covariate forecasting versus DINOv3 image classification. Not the same underlying bug/change." + "reason": "Different underlying changes and code paths: one adds covariate-aware forecasting to TimesFM, the other adds DINOv3 image classification support. They do not look mergeable into one PR and are not the same bug/fix." } ] }, @@ -16788,7 +16696,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16851,7 +16759,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16869,7 +16777,7 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "cb9d51b3e12681ecf661df4ef2bb13365813179f03ca376cb858c273be143c81", + "prepared_review_unit_hash": "771256700e3448515692b8cc95d84cec66d98e727adbe10dbfe99534234a9c7b", "component_budget": { "node_count": 3, "item_count": 3, @@ -16891,24 +16799,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three pull requests are separate model-addition feature PRs (DeepSeek OCR, Isaac, Kimi2-6). They share transformer integration boilerplate, but they do not fix the same underlying code path or represent the same change, so there are no duplicate/merge candidates here.", - "confidence": 0.99, + "summary": "All three PRs are separate model-addition efforts (DeepSeek OCR, Isaac, and Kimi2-6). They share Transformers integration boilerplate, but they do not describe the same bug or change, so no duplicate/merge is warranted.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: each PR introduces a different model family with distinct configs, tokenizers/processors, and tests; the overlap is just framework registration/documentation scaffolding.", + "canonical_pr_reason": "No canonical PR among these: each PR introduces a different model integration and changes different model-specific files.", "best_issue_reason": null, - "best_pr_reason": "If a single representative is needed, 45630 is the broadest/most mature of the three feature PRs, but it is not a duplicate of the others and should not be merged as such.", + "best_pr_reason": "PR #45630 is the most complete and actively reviewed model-addition PR in the set, so it is the best representative if one must be chosen, but it is still not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared auto-registration/docs files are generic template changes, not the same bug or code-path fix." + "reason": "DeepSeek OCR vs Isaac are different model integrations with different model-specific files and code paths; shared auto-registration/docs boilerplate is not enough." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model additions (Isaac vs Kimi2-6). They touch similar integration surfaces, but they implement unrelated models and cannot plausibly be merged as one PR." + "reason": "Isaac vs Kimi2-6 are distinct model additions; they only overlap on common Transformers wiring, not the same underlying change." } ] }, @@ -16922,7 +16830,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16938,7 +16846,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "e8be9af2b24fecec61d93daaadc4cf9a8c380ea7e8cb3fee8c2575047d79e742", + "prepared_review_unit_hash": "9d18f724c64066a3188ebcf35a73cd036a231fc1112bd21d0d57c789cc01c39d", "component_budget": { "node_count": 2, "item_count": 2, @@ -16960,18 +16868,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both workflow dependency-bump PRs, but they address different change sets: #42975 is a broad GitHub Actions runtime compatibility upgrade, while #42976 is a narrower refresh of Docker build actions. They overlap in files but not in the underlying change or failure mode.", + "summary": "Two independent workflow-maintenance PRs: one upgrades checkout/cache/upload-artifact for Node 24 compatibility, the other upgrades docker/build-push-action and related workflow actions. They overlap in GitHub Actions files but address different action families and are not the same underlying change.", "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42975 is the broader, more cohesive upgrade effort, touching many workflows to address Node 24 compatibility across actions versions.", + "canonical_pr_reason": "PR #42975 is the broader, more representative workflow-tooling upgrade and explicitly targets Node 24 compatibility across many workflows.", "best_issue_reason": null, - "best_pr_reason": "PR #42975 is the better cluster representative because it has a clearer single objective and wider scope than the narrower Docker-action version bump in #42976.", + "best_pr_reason": "PR #42975 is the best cluster representative because it spans more workflow files and captures the general GH Actions runtime compatibility migration.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Reject: both are workflow maintenance PRs, but #42975 upgrades several GitHub Actions for Node 24 compatibility, while #42976 only bumps docker/build-push-action to v6. They share files but not the same concrete bug or mergeable change." + "reason": "These are different upgrades: #42975 updates checkout/cache/upload-artifact, while #42976 updates docker/build-push-action. Same subsystem, but not the same concrete fix and not something that clearly belongs in one merged PR." } ] }, @@ -16985,7 +16893,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17014,12 +16922,12 @@ "pull_request:44601|pull_request:45838", "pull_request:45490|pull_request:45838" ], - "prepared_review_unit_hash": "7ae7a7ece2da5a6991a5665eacfed0e27ffba26576e245e93c5cb1ae23e94e08", + "prepared_review_unit_hash": "b28d53845c1daa33beba3ad11df7e18b78edf60589bd0113b937c067de5a341b", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 23514, + "serialized_chars": 23515, "estimated_input_tokens": 5879, "estimated_eval_tokens": 12014 }, @@ -17027,7 +16935,7 @@ "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 23514, + "serialized_chars": 23515, "estimated_input_tokens": 5879, "estimated_eval_tokens": 12014 }, @@ -17036,66 +16944,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated feature PRs: several add different models, and one adds native pipeline-parallel loading. They share a few auto/doc files, but not the same underlying bug or code-path change, so there are no true duplicates here.", - "confidence": 0.98, + "summary": "This cluster is a set of mostly unrelated model-addition PRs plus one distributed loading PR. The shared filenames and similarity scores are driven by repository templates and auto-mapping boilerplate, not the same underlying bug or change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the cluster contains distinct model-addition PRs plus an unrelated distributed-loading feature.", + "canonical_pr_reason": "No single canonical PR: each PR adds a different model or subsystem change (ViT NEPA, Molmo, Molmo2, lfm2 audio, CTSM, PP-OCRv6, and native PP loading), so they are not duplicates of one another.", "best_issue_reason": null, - "best_pr_reason": "No global best PR; none subsumes the others or addresses the same concrete change.", + "best_pr_reason": "There is no strong global best PR for deduping because the items do not describe one common change; if forced to pick a representative, any choice would be arbitrary.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "ViT NEPA and Molmo are different model additions with separate model code; shared docs/auto wiring is just boilerplate." + "reason": "Different models (ViT NEPA vs Molmo); similar docs/auto-registry edits are template overlap, not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "ViT NEPA vs Molmo2 are distinct model families and codepaths, not the same bug or change." + "reason": "Different models (ViT NEPA vs Molmo2) with distinct code paths and files; not the same underlying work." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are unrelated model integrations; overlap is only in generic registration/docs files." + "reason": "ViT NEPA model addition is unrelated to CTSM time-series model addition." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are related in name but are separate model-support PRs with different implementations and targets." + "reason": "Molmo and Molmo2 are separate model integrations with different implementations; they are related by naming, not by being the same PR." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM are different model families; no shared concrete code-path problem." + "reason": "Molmo and CTSM are different model families; overlap is only in shared framework boilerplate." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are separate model additions, not a single mergeable change." + "reason": "Molmo2 and CTSM are distinct model additions and do not fix or add the same code-path problem." }, { "left": "pull_request:43973", "right": "pull_request:45838", "accept": false, - "reason": "Lfm2 audio support and PP-OCRv6 text recognition support are unrelated model additions across different modalities." + "reason": "Lfm2 audio support and PP-OCRv6 text recognition are different model integrations with no shared concrete fix." }, { "left": "pull_request:44601", "right": "pull_request:45838", "accept": false, - "reason": "Native pipeline-parallel loading is a distributed-runtime feature, not the same change as PP-OCRv6 model support." + "reason": "Native pipeline-parallel loading infrastructure is unrelated to PP-OCRv6 model support; not mergeable as one PR." }, { "left": "pull_request:45490", "right": "pull_request:45838", "accept": false, - "reason": "CTSM and PP-OCRv6 are different model families; shared auto/doc files do not indicate duplicate functionality." + "reason": "CTSM model addition and PP-OCRv6 recognition models are different changes in different subsystems." } ] }, @@ -17109,7 +17017,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17117,63 +17025,85 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-43291-3", + "cluster_id": "cluster-43291-4", "nodes": [ "pull_request:43291", "pull_request:43665", - "pull_request:44815" + "pull_request:44815", + "pull_request:45892" ], "soft_pairs": [ "pull_request:43291|pull_request:43665", "pull_request:43291|pull_request:44815", - "pull_request:43665|pull_request:44815" + "pull_request:43291|pull_request:45892", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:45892", + "pull_request:44815|pull_request:45892" ], - "prepared_review_unit_hash": "00913baf0be3bf2befddbbfc2c5c9e8e59d82cba85558038c98cfc085ad2c181", + "prepared_review_unit_hash": "a7144b309434a0cab1768e797298f5b8c7cc8f62ca61249e9241515cf6ce4ad5", "component_budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 3, - "serialized_chars": 7093, - "estimated_input_tokens": 1774, - "estimated_eval_tokens": 3804 + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10131, + "estimated_input_tokens": 2533, + "estimated_eval_tokens": 5322 }, "budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 3, - "serialized_chars": 7093, - "estimated_input_tokens": 1774, - "estimated_eval_tokens": 3804 + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10131, + "estimated_input_tokens": 2533, + "estimated_eval_tokens": 5322 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated PRs: a Whisper tokenizer/test fix, a synthetic CLIP/ViT test-only change, and an FP8 dequant/loading fix. The only substantive bugfix is the Whisper PR; the others do not appear to address the same underlying issue or be mergeable into one change.", - "confidence": 0.94, + "summary": "These are mostly unrelated PRs that all use generic \u201cfix\u201d wording. Two are test-only changes, one is a Whisper tokenizer/test change, one is a FP8 dequantization change, and one is a DeepSeek v4 attention fix. They do not look like the same underlying bug or change cluster.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43291 is the clearest real code fix: it changes Whisper decode/timestamp handling in the tokenizer and pairs it with targeted tests. The other PRs are either test-only or a different quantization/loading area.", + "canonical_pr_reason": "No true duplicate PR stands out here; if one representative must be chosen, 44815 is the most substantive code fix and the broadest patch in the set, while the others are either test-only or different model-family bug fixes.", "best_issue_reason": null, - "best_pr_reason": "PR 43291 is the best representative of the cluster because it contains the most concrete, self-contained bugfix with directly related code and tests.", + "best_pr_reason": "44815 is the best representative because it contains a concrete implementation fix across core loading and FP8 integration, whereas 43665 is test noise and 43291/45892 are narrower model-specific changes.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer decoding vs. CLIP/ViT test files are unrelated code paths; this looks like an incidental test patch, not the same bug or change." + "reason": "Whisper tokenizer/test changes vs unrelated CLIP/ViT test edits; no shared underlying bug or code path." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decode behavior and FP8 dequant/model-loading logic are different subsystems and different fixes; no shared concrete code-path problem." + "reason": "Whisper decoding bug/test work is unrelated to FP8 dequantization and model loading." + }, + { + "left": "pull_request:43291", + "right": "pull_request:45892", + "accept": false, + "reason": "Different subsystems and models: Whisper tokenizer vs DeepSeek v4 attention logic." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "One is synthetic test edits for CLIP/ViT, the other is a dequant/loading fix for Mistral/FP8; they do not plausibly merge into one PR or share the same underlying bug." + "reason": "43665 is clearly test-only/assert noise; 44815 is a real FP8 dequant fix, so they are not the same change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:45892", + "accept": false, + "reason": "Test-only CLIP/ViT edits are unrelated to the DeepSeek v4 model bug fix." + }, + { + "left": "pull_request:44815", + "right": "pull_request:45892", + "accept": false, + "reason": "Both are real fixes, but they target different code paths and model families; not mergeable as one duplicate PR." } ] }, @@ -17187,7 +17117,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17203,7 +17133,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "061af280bc08819e4697e4dc24ea7453e6df68250603493966544c5c83300e5d", + "prepared_review_unit_hash": "312df79426bba27e475205f3e76e6a7fa33b4f4d511d4932e1c6f19cea78a485", "component_budget": { "node_count": 2, "item_count": 2, @@ -17225,18 +17155,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one is a setup/version/dependency tweak, while the other rewires multiple CI/workflow and testing-related files. They overlap only in a shared Vit test file and both being repo-test style PRs, which is not enough to treat them as the same change.", - "confidence": 0.96, + "summary": "The two PRs are unrelated: one is a setup/version/dependency tweak, the other is a CI/workflow and testing-utilities change. The shared Vit test file looks incidental, not evidence of the same fix.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43532 is the better representative PR because it is the broader, more substantive change set, touching multiple workflows and testing utilities rather than a narrow setup/version edit.", + "canonical_pr_reason": "PR 43532 is the more representative change set: it spans CI workflow and testing utility code, while 43488 is a small setup/dependency bump plus a test touch. Neither is a duplicate of the other.", "best_issue_reason": null, - "best_pr_reason": "PR #43532 is the strongest standalone PR in the cluster: it spans several files and a concrete CI/testing flow change, whereas #43488 is mostly a small dependency/version adjustment.", + "best_pr_reason": "PR 43532 is the better cluster representative because it contains the broader substantive code-path change; 43488 appears to be a narrow bot/setup-oriented edit.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Reject: the diffs target different code paths and maintenance areas. #43488 changes setup dependencies/versioning and a Vit test, while #43532 updates GitHub Actions workflows plus testing utilities. Shared test-file presence alone does not indicate the same underlying change." + "reason": "Different underlying changes. 43488 updates setup/versioning and adds a pytest dependency, while 43532 rewires CI/workflow behavior and touches testing utilities/notifications. The shared Vit test file is not enough to treat them as the same PR." } ] }, @@ -17250,7 +17180,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17266,12 +17196,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "5ff90a4a54ba4544bac56b4183f4467008db65c7ecf950e164f54db65c2c6c32", + "prepared_review_unit_hash": "4f60e83b520119d58447bddf8120a30a53cc5926ac8b4e471c535f11ce084ac7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -17279,7 +17209,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -17288,18 +17218,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related pipeline-addition PRs, but they implement different features: promptable concept segmentation vs promptable visual segmentation. The superficial overlap is mostly in docs and shared pipeline plumbing, not the same bug/change.", - "confidence": 0.97, + "summary": "These are distinct feature PRs: one adds Promptable Concept Segmentation and the other adds Promptable Visual Segmentation. They overlap on shared pipeline/docs scaffolding, but they implement different pipelines, different model integrations, and different test suites, so they should not be treated as duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR for this cluster: 43612 and 43613 add different pipelines for different tasks/models.", + "canonical_pr_reason": "No issue is present. If a representative PR is needed, #43613 is the broader/more complete feature addition, with more model integrations and a larger test surface.", "best_issue_reason": null, - "best_pr_reason": "If one PR must be chosen as the representative, 43613 is slightly broader and more integrated, but it is still a separate feature from 43612.", + "best_pr_reason": "No issue is present. #43613 is the better cluster representative because it spans the larger concrete change set and central pipeline registration work, but it is not a duplicate of #43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different concrete features and code paths: concept segmentation vs visual segmentation. Shared docs/pipeline scaffolding is not enough to treat them as duplicates or mergeable into one PR." + "reason": "Reject: they add different pipelines for different tasks (concept vs visual segmentation). Shared docs/registry files are just common scaffolding, and the implementation/test/model files are distinct, so they are not the same underlying change." } ] }, @@ -17313,7 +17243,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17329,40 +17259,40 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "dd31559c7ac1a964d2ee4a8176a24ad211ff729e4301f33609a25050380ee765", + "prepared_review_unit_hash": "f95d398d7589f5073a33ef9c7d9ce2644d8bad37299db0481a9cf9d69edb973c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate trainer.py feature PRs: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers. They share a file but not the same underlying bug or change.", + "summary": "These are two separate Trainer extension PRs touching different code paths in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT. They are not duplicates.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No single canonical PR: the two changes address different behaviors and should not be merged as one duplicate candidate.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other, so there is no global best PR for this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both modify trainer.py, but they address different code paths and concerns: loss scaling in training_step vs. column filtering/signature inspection. They are not the same fix and would not plausibly merge into one PR." + "reason": "Different underlying changes: 43651 alters training loss scaling during gradient accumulation, while 43663 adds an overridable helper for signature-column filtering. Shared file alone is not enough to treat them as the same bug or fix." } ] }, @@ -17376,7 +17306,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17394,55 +17324,55 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "f6eb8ac409f771c6ac0ff75d001f7fb0a3766c3e26c5ca0bde76c9effc863856", + "prepared_review_unit_hash": "9635ed14b57f971fc1bc28589d67d341ab2c2172268a37be941ddc014316d5c5", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5993, - "estimated_input_tokens": 1499, - "estimated_eval_tokens": 3254 + "serialized_chars": 5991, + "estimated_input_tokens": 1498, + "estimated_eval_tokens": 3252 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5993, - "estimated_input_tokens": 1499, - "estimated_eval_tokens": 3254 + "serialized_chars": 5991, + "estimated_input_tokens": 1498, + "estimated_eval_tokens": 3252 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three open PRs cluster around TypeAdapter/annotation runtime issues in `src/transformers/cli/serve.py`. Two are closely related fixes for the same optional-import/annotation problem; the third appears to be a separate annotation cleanup on a different issue thread.", - "confidence": 0.74, - "canonical_issue_reason": null, - "canonical_pr_reason": "`pull_request:43842` is the most focused and explicit fix for the TypeAdapter NameError, with a minimal change that avoids importing `TypeAdapter` at runtime when pydantic is absent and string-annotates the validator parameter.", - "best_issue_reason": null, - "best_pr_reason": "`pull_request:43842` is the best representative PR because it directly targets the reported failure mode, keeps the scope small, and is the clearest concrete fix in the cluster.", + "summary": "The cluster is centered on a `serve.py` annotation-related `TypeAdapter` NameError. `43836` and `43842` are tightly aligned on the same tracker (`43824`) and the same concrete fix area, while `43656` is related but broader and aimed at a different issue target (`43576`).", + "confidence": 0.84, + "canonical_issue_reason": "`43824` is the clearest shared issue proxy because it is explicitly targeted by both `43836` and `43842`, and it matches the concrete `TypeAdapter` NameError in `serve.py`.", + "canonical_pr_reason": "`43842` is the best canonical PR because it is the most focused fix for the `TypeAdapter` NameError in `serve.py`, with a minimal change and an explicit link to the shared issue.", + "best_issue_reason": "`43824` is the best issue representative for the cluster because it is the common tracker for the most tightly matching PR pair and corresponds to the specific runtime failure being fixed.", + "best_pr_reason": "`43842` is the strongest representative PR: it directly and narrowly addresses the `TypeAdapter` import-time NameError in `serve.py` without adding the broader annotation cleanup present in `43656`.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Both touch the same file and mention TypeAdapter, but they appear to address different underlying problems: `43656` is a broader annotations/future-annotations cleanup tied to a different issue, while `43842` specifically fixes pydantic-optional runtime evaluation. Not the same concrete bug fix." + "reason": "Related in the same file and both touch annotation evaluation, but `43656` targets a different issue (`43576`) and applies a broader `from __future__ import annotations`/annotation cleanup fix rather than the focused `TypeAdapter`-specific bug in `43842`. Not clearly the same underlying change." }, { "left": "pull_request:43836", "right": "pull_request:43842", "accept": true, - "reason": "Both PRs target the same TypeAdapter NameError path in `serve.py` and the same issue thread, and both change how `TypeAdapter` is referenced to avoid runtime annotation evaluation. They look like alternate fixes for the same bug and could plausibly be merged into one PR." + "reason": "Both PRs target the same issue (`43824`) and address the same concrete `TypeAdapter` NameError in `serve.py`, differing mainly in how narrowly they express the fix. They are plausibly mergeable into one bugfix PR." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: the three PRs all touch `src/transformers/cli/serve.py`, and the issue targets show that `43656` is tied to a different thread than `43836`/`43842`. The best-PR choice is also supported by the diff preview, since `43842` is the smallest explicit fix for the `TypeAdapter` runtime-name problem. The soft-edge decision between `43836` and `43842` is conservative enough: both appear to address the same bug class in the same file and could plausibly be merged into one PR, while `43656` is reasonably treated as a separate annotation cleanup." + "accept": false, + "feedback": "The summary is a bit too strong on the soft edge for 43656 vs 43842. 43656 targets a different issue (`43576`), and its fix is broader (`from __future__ import annotations` plus wider annotation cleanup) than the focused `TypeAdapter`-specific change in 43842. The packet supports saying they are in the same file and plausibly related, but not that they clearly fix the same bug or are comfortably mergeable into one PR for the same bugfix. The 43836/43842 pairing is better grounded because they share issue `43824`, but the 43656/43842 accept is overstated." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -17450,7 +17380,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17516,7 +17446,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -17578,12 +17508,12 @@ "pull_request:44071|pull_request:44722", "pull_request:44018|pull_request:44086", "pull_request:44018|pull_request:44085", - "pull_request:44068|pull_request:44086", "pull_request:44071|pull_request:44085", + "pull_request:44068|pull_request:44086", "pull_request:44059|pull_request:44086", "pull_request:44085|pull_request:44086", - "pull_request:44059|pull_request:44722", "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", "pull_request:44018|pull_request:44071", "pull_request:44017|pull_request:44161", "pull_request:44017|pull_request:44722", @@ -17596,8 +17526,8 @@ "pull_request:44059|pull_request:44071", "pull_request:44073|pull_request:44161", "pull_request:44056|pull_request:44101", - "pull_request:44059|pull_request:44068", "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", @@ -17610,8 +17540,8 @@ "pull_request:44018|pull_request:44161", "pull_request:43997|pull_request:44071", "pull_request:44010|pull_request:44071", - "pull_request:44010|pull_request:44073", "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", "pull_request:43997|pull_request:44074", "pull_request:43997|pull_request:44129", "pull_request:44071|pull_request:44129", @@ -17628,340 +17558,364 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "dc05490d44424914b9e86e91426604d95f7bec6460d1b4c2cb04e28f350f4a53", + "prepared_review_unit_hash": "34a2a01bc9aad95611920483ea6b3b9ff708b495fa3e77d064cfd7d0a1ab79a5", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84898, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84895, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75094, - "estimated_input_tokens": 18774, - "estimated_eval_tokens": 37804 + "serialized_chars": 75092, + "estimated_input_tokens": 18773, + "estimated_eval_tokens": 37802 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly an umbrella output-tracing migration, with a few true duplicate subgroups inside GPT-J/CodeGen and GPT-Neo. Most other pairings are only thematically similar per-model refactors and should stay separate.", - "confidence": 0.76, - "canonical_issue_reason": "Issue #43979 is the common umbrella tracking issue for this standardized output-tracing migration; several PRs in the cluster explicitly target it.", - "canonical_pr_reason": "PR #44722 is the strongest PR anchor because it is the broadest concrete GPT-J/CodeGen output-tracing refactor and overlaps the GPT-J-specific variants most directly.", - "best_issue_reason": "#43979 is the best global issue anchor because it cleanly captures the shared refactor theme across the cluster.", - "best_pr_reason": "#44722 is the best global PR anchor because it represents the broadest concrete change in the strongest duplicate subgroup.", + "summary": "This cluster is mostly a set of model-specific output-tracing refactors around issue #43979. The only true duplicate subset is GPT-J/CodeGen: 44066, 44085, and 44722 overlap on the same GPT-J output-tracing code path. The GPT-Neo pair is only thematically similar and should not be treated as a duplicate.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best canonical PR for the duplicate subset because it is explicit about the shared issue, spans the shared GPT-J/CodeGen refactor, and is the broadest representative of the overlapping patch set.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the strongest representative overall: it directly covers the shared GPT-J/CodeGen output-tracing change and cleanly anchors the actual duplicate cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; the shared umbrella issue is not enough." + "reason": "Different model families; only the general output-tracing theme overlaps." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different model families; same refactor pattern, but not the same change." + "reason": "Different model implementations and different concrete refactors." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Same GPT-J output-tracing refactor on the same file; 44066 is just the broader GPT-J/CodeGen sibling." + "reason": "Same GPT-J output-tracing code path with overlapping file changes; these could be merged into one PR." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "Different model families and separate per-model refactors." + "reason": "RegNet and EfficientNet are separate model code paths; same pattern, not the same change." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "Different model families; only the high-level output-tracing theme matches." + "reason": "EfficientNet and GPT-J/CodeGen are unrelated concrete changes." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen vs EfficientNet are unrelated code paths." + "reason": "Different models and files; only a broad refactor theme matches." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are different model implementations." + "reason": "MobileNetV2 and DeBERTaV2 are unrelated implementations." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen and MGP-STR are separate model-specific refactors." + "reason": "Same output-tracing pattern, but different model code paths." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are different model code paths." + "reason": "GPT-J/CodeGen versus MPT is only a theme match, not the same bug or patch." }, { "left": "pull_request:44018", "right": "pull_request:44068", - "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors on the same file/code path." + "accept": false, + "reason": "GPT-Neo pairing is only soft similarity; no shared issue target or concrete duplicate signal." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different models." + "reason": "GPT-J/CodeGen and GPT-Neo are different model refactors." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different concrete changes." + "reason": "Different model scope; the overlap is just the standardized output-tracing pattern." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated model refactors." + "reason": "ResNet and GPT-J/CodeGen are unrelated model implementations." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen touch different code paths." + "reason": "MGP-STR and GPT-J/CodeGen are different concrete code paths." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are unrelated model-specific changes." + "reason": "ResNet and GPT-J are unrelated model changes." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different concrete code paths." + "reason": "GPT-Neo and ResNet share only the broad refactor style." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are separate model refactors." + "reason": "ResNet and MPT are separate model code paths." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are different model families." + "reason": "GPT-Neo and Flaubert/XLM are different model changes." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model code paths." + "reason": "GPT-Neo refactor is separate from the GPT-J/CodeGen duplicate set." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are unrelated per-model refactors." + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." }, { - "left": "pull_request:44018", - "right": "pull_request:44086", + "left": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and MGP-STR do not fix the same concrete problem." + "reason": "MPT and GPT-J/CodeGen are different concrete changes." }, { "left": "pull_request:44018", - "right": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and GPT-J are different model implementations." + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." }, { - "left": "pull_request:44068", - "right": "pull_request:44086", + "left": "pull_request:44018", + "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and MGP-STR are different model families/code paths." + "reason": "GPT-Neo and GPT-J are separate model refactors." }, { "left": "pull_request:44071", - "right": "pull_request:44722", + "right": "pull_request:44086", "accept": false, - "reason": "MPT and GPT-J/CodeGen are separate model refactors." + "reason": "MPT and MGP-STR are not the same underlying change." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated changes." + "reason": "GPT-2 and MGP-STR are different model implementations." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are different model code paths." + "reason": "GPT-J and MGP-STR are unrelated concrete changes." }, { - "left": "pull_request:44059", - "right": "pull_request:44722", + "left": "pull_request:44098", + "right": "pull_request:44101", "accept": false, - "reason": "GPT2 and GPT-J/CodeGen are separate refactors, even though they share the umbrella theme." + "reason": "ViLT and Flaubert/XLM are different model code paths." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "ViLT and Flaubert/XLM are different model families." + "reason": "GPT-2 refactor is unrelated to GPT-J/CodeGen output tracing." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "VisualBert and Flaubert/XLM are unrelated implementations." + "reason": "GPT-Neo and MPT only share the high-level modernization pattern." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are unrelated model-specific refactors." + "reason": "SegFormer and LongT5 are different model refactors." }, { - "left": "pull_request:43997", - "right": "pull_request:44017", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "RegNet and SegFormer are separate model changes." + "reason": "SegFormer and GPT-J/CodeGen are unrelated code paths." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:44017", + "right": "pull_request:44073", "accept": false, - "reason": "RegNet and MPNet are unrelated per-model refactors." + "reason": "SegFormer and VisualBert are different model implementations." }, { - "left": "pull_request:44071", - "right": "pull_request:44129", + "left": "pull_request:44073", + "right": "pull_request:44101", "accept": false, - "reason": "MPT and SpeechT5 are different code paths and model families." + "reason": "VisualBert and Flaubert/XLM are separate model changes." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "SqueezeBert and MPT are unrelated model-specific changes." + "reason": "TextNet and Flaubert/XLM are unrelated model code paths." }, { - "left": "pull_request:44010", + "left": "pull_request:43997", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are different concrete refactors." + "reason": "RegNet and VisualBert share only the refactor theme." }, { - "left": "pull_request:43997", + "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are separate model changes." + "reason": "MPNet and LongT5 are different concrete implementations." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", - "accept": false, - "reason": "TextNet and Flaubert/XLM are unrelated implementations." - }, - { - "left": "pull_request:44056", + "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are different model-specific refactors." + "reason": "CVT/FNet and LongT5 are not the same underlying change." }, { - "left": "pull_request:43997", - "right": "pull_request:44018", + "left": "pull_request:44010", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated concrete changes." + "reason": "SqueezeBERT and MPT are unrelated model paths." }, { "left": "pull_request:44010", - "right": "pull_request:44018", + "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and GPT-Neo are different model code paths." + "reason": "SqueezeBERT and VisualBert are different implementations." }, { - "left": "pull_request:44010", - "right": "pull_request:44044", + "left": "pull_request:44044", + "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and DeBERTa are separate per-model refactors." + "reason": "DeBERTaV2 and SpeechT5 are different model refactors." }, { "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTaV2 and ViTS are not the same change." + }, + { + "left": "pull_request:44017", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa and TextNet are unrelated changes." + "reason": "SegFormer and TextNet are separate model code paths." }, { - "left": "pull_request:44066", - "right": "pull_request:44068", + "left": "pull_request:44071", + "right": "pull_request:44129", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different models." + "reason": "MPT and SpeechT5 are unrelated changes." }, { - "left": "pull_request:44068", - "right": "pull_request:44071", + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and SegFormer are only loosely similar refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", "accept": false, - "reason": "GPT-Neo and MPT are different code paths." + "reason": "RegNet and MPNet are different model code paths." }, { "left": "pull_request:44071", "right": "pull_request:44076", "accept": false, - "reason": "MPT and ImageGPT are separate model refactors." + "reason": "MPT and ImageGPT are unrelated implementations." }, { - "left": "pull_request:44072", - "right": "pull_request:44161", + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing refactor on the same file; these overlap enough to be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", "accept": false, - "reason": "EfficientNet and LongT5 are unrelated model-specific changes." + "reason": "GPT-J and SpeechT5 are different models and changes." }, { - "left": "pull_request:44073", - "right": "pull_request:44098", + "left": "pull_request:43996", + "right": "pull_request:44007", "accept": false, - "reason": "VisualBert and ViLT are different model families." + "reason": "CVT/FNet and RegNet are different model refactors." }, { - "left": "pull_request:44101", - "right": "pull_request:44161", + "left": "pull_request:43996", + "right": "pull_request:44129", "accept": false, - "reason": "Flaubert/XLM and LongT5 are unrelated code paths." + "reason": "CVT/FNet and SpeechT5 are unrelated model paths." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44044", + "right": "pull_request:44129", "accept": false, - "reason": "TextNet and Flaubert/XLM are separate model refactors." + "reason": "DeBERTaV2 and SpeechT5 are different concrete changes." }, { - "left": "pull_request:44154", - "right": "pull_request:44044", + "left": "pull_request:44044", + "right": "pull_request:44074", "accept": false, - "reason": "ViTS and DeBERTa are unrelated model implementations." + "reason": "DeBERTaV2 and TextNet are unrelated model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen standardized output-tracing refactor with direct file overlap; this is a true duplicate/overlap." } ] }, "evaluator_result": { "accept": false, - "feedback": "The cluster-level interpretation is mostly grounded, but the canonical issue reason overstates the evidence: #43979 is not explicitly targeted by most PRs in the packet, only by a minority of them. Please soften that language to something like \u201cseveral PRs explicitly target it\u201d or \u201cit serves as a common umbrella issue.\u201d The accepted soft-edge pairs are plausible, but the writeup should stay more conservative about how much direct issue overlap exists." + "feedback": "The summary is mostly grounded, but it overstates the GPT-Neo pairing. In the packet, pull_request:44018 vs pull_request:44068 is only a soft-similarity match; there is no deterministic accept signal or shared issue target, so calling them a clear duplicate subset is too strong. The GPT-J/CodeGen duplicate claims are well supported, but the GPT-Neo verdict should be more conservative." }, "evaluator_used": true, "retried": true, @@ -17972,7 +17926,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 26, + "component_count": 27, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -17999,70 +17953,70 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "f36ab84f6073bdee94ea4d4ef57c780baa25a4a25f32bd06a65240baf0de026e", + "prepared_review_unit_hash": "f1e620e090e23fda40c46d4d4eecda9d49a3128a880cb4ed434e7d2317f46c33", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84898, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84895, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17125, - "estimated_input_tokens": 4282, - "estimated_eval_tokens": 8820 + "serialized_chars": 17124, + "estimated_input_tokens": 4281, + "estimated_eval_tokens": 8818 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These pull requests are all separate, model-specific migrations to the new standardized output-capture/tracing utilities. They share a refactor theme, but they are not the same underlying code-path change, so the soft duplicate edges should be rejected.", - "confidence": 0.94, + "summary": "These PRs are all variants of a standardized output-tracing refactor, but they touch different model implementations and code paths rather than the same concrete fix.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "No true duplicate PR cluster here; if a representative is needed, PR 44073 is a good canonical example because it cleanly applies the standardized output-capture pattern in one model with explicit _can_record_outputs and can_return_tuple usage.", + "canonical_pr_reason": "PR 44073 is the clearest representative: it explicitly targets the standardized output-capture refactor, adds the new tracing hooks, and shows the core pattern used across the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 44073 is the best single representative of the cluster\u2019s theme: it applies the standardized output-tracing refactor directly and explicitly wires the capture targets, making it a clear exemplar of the change pattern.", + "best_pr_reason": "PR 44073 is the best single exemplar because its title and diff most directly reflect the shared refactor theme and it includes both the decorator import changes and output-capture hook setup.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Same broad refactor theme, but FocalNet and ImageGPT are different model implementations and code paths; not mergeable as one duplicate PR." + "reason": "Different model families (FocalNet vs ImageGPT) and separate implementations; same refactor style, but not the same underlying bug or mergeable fix." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Both are output-tracing refactors, but VisualBert and TextNet are unrelated model code paths; too model-specific to treat as the same change." + "reason": "VisualBert vs TextNet; both migrate output tracing, but they modify different model code paths and are not one concrete change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet vs VisualBert are distinct model internals; shared standardized-output machinery is not enough to call them the same underlying fix." + "reason": "MPNet and VisualBert are unrelated model implementations; these are parallel refactors, not duplicate fixes." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These are separate model migrations with similar utility usage, not the same concrete bug or one PR that could plausibly absorb the other." + "reason": "Both are output-tracing refactors, but for different architectures and files, so they should not be merged as one PR." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated model paths; similarity is only at the refactor pattern level." + "reason": "RegNet and SqueezeBert address different model internals; shared decorator imports do not indicate the same fix." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are different architectures, and 44129 is additionally tied to its own explicit tracking issue, so this is not a duplicate change." + "reason": "SqueezeBert and SpeechT5 are distinct code paths; the similarity is only the output-capture migration pattern, not a duplicate change." } ] }, @@ -18076,7 +18030,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18142,274 +18096,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "2ad3e5db4e78e2f14a9157e0dfcce8920a0b29976a8a6fe5224959eb58131c5d", + "prepared_review_unit_hash": "71cf12ecd24f9b0a0d704cceddddfbfcfbd0f5be6d298703ce86f7bed31080ba", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40178, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40162, + "estimated_input_tokens": 10041, + "estimated_eval_tokens": 20338 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40178, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40162, + "estimated_input_tokens": 10041, + "estimated_eval_tokens": 20338 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are all model-specific output-tracing refactor PRs that share a naming pattern and some helper usage, but they affect different code paths and should not be deduplicated as one change.", - "confidence": 0.82, + "summary": "This cluster is a set of similar-but-independent output-tracing refactors across many unrelated model implementations. They share a naming pattern and plumbing changes, but not the same concrete bug or change.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue artifacts are present. If one PR must represent the cluster, 44003 is the most central because it spans both mamba and falcon_mamba and shows the core tracing-helper integration.", + "canonical_pr_reason": "PR 44003 is the best representative: it makes the most substantial tracing-related plumbing changes, adds output-capturing support, and spans two closely related model files (mamba/falcon_mamba).", "best_issue_reason": null, - "best_pr_reason": "44003 is the strongest representative of the theme, but only as a cluster exemplar: it is broader than most others, yet still a model-family-specific refactor rather than the same change as the rest.", + "best_pr_reason": "PR 44003 is the strongest single representative of the cluster because it combines the core tracing infrastructure change with model-specific updates; the other PRs are narrower, mostly single-model refactors.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different models and files (timm_backbone vs upernet); same refactor theme only." + "reason": "Different model families and code paths; only the refactor theme is shared." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different model families (timm_backbone vs vision_text_dual_encoder); not the same code-path fix." + "reason": "Both are output-tracing refactors, but for unrelated model architectures and distinct forward paths." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different targets (timm_backbone vs univnet); shared helper usage is not enough to merge." + "reason": "Same general plumbing theme, but not the same underlying bug or mergeable change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Both are tracing refactors, but they touch unrelated model implementations and changes." + "reason": "Timm backbone and mamba/falcon_mamba are unrelated code paths; shared tracing terminology is not enough." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different architectures (timm_backbone vs rwkv); no shared concrete bug/change." + "reason": "Different models and mechanisms; no common concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated model paths; only the refactor pattern matches." + "reason": "Different models; both change tracing behavior, but not the same underlying issue." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different models (mobilenet_v1 vs upernet); not the same underlying change." + "reason": "Separate model implementations with distinct forward logic." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families and files; both alter tracing, but not one mergeable fix." + "reason": "Different model families; similar refactor pattern only." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "MobileNetV1 vs VisionEncoderDecoder are separate code paths; same utility addition only." + "reason": "MobileNet V1 and vision encoder-decoder wrappers are unrelated code paths." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 and SpeechEncoderDecoder are unrelated; no duplicated bug or change." + "reason": "Different architectures; no evidence of one concrete shared bug." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-Text Dual Encoder vs UnivNet are different model stacks; refactor theme is too broad." + "reason": "Unrelated models; same refactor motif but not the same fix." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different targets (vision_text_dual_encoder vs upernet); not the same concrete code path." + "reason": "Different code paths and model types." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Shared output-tracing terminology, but different model implementations and edits." + "reason": "Vision-text dual encoder vs mamba/falcon_mamba are unrelated implementations." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (vision_text_dual_encoder vs depth_anything); not one duplicate fix." + "reason": "Different models with no shared concrete bug." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated components; same helper refactor only." + "reason": "Unrelated model families; only the refactor style matches." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different code paths in different model families; not plausibly one PR." + "reason": "No common code path; one is UnivNet, the other is Mamba-family tracing." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "UnivNet vs Depth Anything are separate changes, despite similar output-tracing cleanup." + "reason": "Different models and different output-tracing implementations." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are unrelated models; no shared underlying bug." + "reason": "Different model families and forward paths." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model implementations and files; only the generic tracing refactor overlaps." + "reason": "UPerNet and Mamba-family changes are not the same underlying change." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Mamba-family tracing refactor vs SuperPoint output change; different code paths and behavior." + "reason": "Mamba-family tracing and SuperPoint tracing are unrelated implementations." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/RWKV-style tracing work is unrelated to CodeGen's decoder changes." + "reason": "Different models and different refactor scope; no mergeable shared fix." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone and CodeGen are unrelated implementations; same refactor motif only." + "reason": "Timm backbone and CodeGen are different code paths; same theme only." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen touch different domains and code paths; not a duplicate." + "reason": "UnivNet and CodeGen are unrelated models." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are different model families; no shared concrete fix." + "reason": "Distinct models with distinct forward semantics." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Vision-text dual encoder vs CodeGen are unrelated; same helper pattern is insufficient." + "reason": "Vision-text dual encoder and CodeGen are unrelated; not the same bug." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything vs VisionEncoderDecoder are distinct model code paths." + "reason": "Different wrapper models and different call chains; only both are tracing refactors." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Depth Anything and SpeechEncoderDecoder are unrelated; not the same bug/change." + "reason": "Depth Anything and Speech Encoder-Decoder are unrelated model families." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything vs SuperPoint are different model implementations; no duplicate fix." + "reason": "No shared concrete problem; different model implementations." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Depth Anything and RWKV are separate model families; only the refactor pattern matches." + "reason": "Different models and different output-capturing changes." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything vs DPR are unrelated code paths; not mergeable as one fix." + "reason": "Depth-related tracing and DPR tracing are separate changes." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44027", "accept": false, - "reason": "VisionEncoderDecoder and RWKV are different subsystems; no shared underlying issue." + "reason": "Vision and speech encoder-decoder wrappers are parallel but distinct code paths." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44028", "accept": false, - "reason": "VisionEncoderDecoder vs SpeechEncoderDecoder are analogous but distinct implementations." + "reason": "Different models with no common underlying bug." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44029", "accept": false, - "reason": "Different model families (vision encoder-decoder vs superpoint); not the same change." + "reason": "Distinct architectures; shared tracing theme is too broad." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "VisionEncoderDecoder and DPR touch unrelated paths; shared pattern only." + "reason": "Different model families and implementation details." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "SpeechEncoderDecoder and SuperPoint are unrelated models; not a duplicate." + "reason": "Speech encoder-decoder and SuperPoint are unrelated." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "SpeechEncoderDecoder vs RWKV are separate code paths with different behavior." + "reason": "Different code paths and model families." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "SpeechEncoderDecoder and DPR are unrelated; same refactor motif only." + "reason": "Separate models; not the same fix." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV differ in architecture and implementation; no same bug." + "reason": "SuperPoint and RWKV have unrelated forward logic." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint vs DPR are distinct models; not one mergeable PR." + "reason": "Different models; no evidence of a shared concrete issue." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are unrelated model families; the similarity is only in output-tracing cleanup." + "reason": "RWKV and DPR are unrelated implementations." } ] }, @@ -18423,7 +18377,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18442,7 +18396,7 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "946e530f0b646b136cde5204ca87d76279471899a6f2689fd37f998cf4367129", + "prepared_review_unit_hash": "3f16e3926c01faf250a3a8aed226b646b1b464af9a03ececae777f6006bb1efb", "component_budget": { "node_count": 3, "item_count": 3, @@ -18464,30 +18418,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-addition PRs (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS) that share Transformers boilerplate and auto-registry updates, but they do not fix or implement the same underlying model or code path.", - "confidence": 0.97, + "summary": "These are three separate model-addition PRs, not duplicates. They share Transformers boilerplate (docs, auto registries, model package wiring), but each targets a different model family and code path: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #45075 is the best standalone representative of the cluster because it is a non-draft, substantial model integration with a dedicated conversion script and broad test coverage; however, it does not subsume the other PRs.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR #45075 is the strongest overall example of a complete model-implementation PR in this set, but it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model implementations for different modalities/domains (vision-language vs TTS). Shared docs/auto-registration boilerplate is not enough to treat them as the same change." + "reason": "Different models and different concrete functionality: PenguinVL vs Voxtral TTS. The overlap is only shared framework wiring and docs, which is not enough to treat them as the same change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "They add unrelated models (PenguinVL vs DeepSeek-OCR-2) with distinct configs, modeling, processing, and tests; only the generic new-model scaffolding overlaps." + "reason": "Both are model additions, but they implement different model families and code paths (PenguinVL vs DeepSeek-OCR-2). Shared auto-registration and docs are boilerplate, not duplicate evidence." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "OCR model support and TTS model support are separate features with different code paths and artifacts; no evidence they could be merged into one concrete fix." + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model integrations with different modalities and implementation details. They are not plausibly mergeable into one PR as the same fix/change." } ] }, @@ -18501,7 +18455,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18517,12 +18471,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "46fec4e767fbe8413a8952dd3163fcb07f116505cd91dbb7d7308ca2dc3bcd25", + "prepared_review_unit_hash": "04945a5ba076074e47e1e356f76ff4936fa82738d4af99bdbae2655ff9ca8f40", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4406, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -18530,7 +18484,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4406, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -18539,18 +18493,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related by the broad theme of fusion, but they implement different features on different code paths: one adds a new low-level Module Fusion API, the other wires up n-to-1 kernel fusion through KernelConfig and model/integration changes. They do not look like duplicates.", - "confidence": 0.92, + "summary": "These PRs are conceptually related around module/kernel fusion, but they address different layers: one adds a generic Module Fusion API, the other implements n-to-1 kernel fusion via KernelConfig and model-loading integration.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45363 is the better cluster representative because it implements the more concrete end-to-end fusion path (`KernelConfig` plus modeling/integration plumbing) rather than only introducing a lower-level API layer.", + "canonical_pr_reason": "PR #45363 is the more concrete, end-user-facing implementation: it wires fusion through KernelConfig and model loading across multiple integration files, whereas #44979 is a lower-level API scaffold for module fusion.", "best_issue_reason": null, - "best_pr_reason": "PR #45363 is the strongest overall artifact in this set: it has a clearer user-facing feature, touches the actual runtime integration points, and includes an example demonstrating the intended fused workflow.", + "best_pr_reason": "PR #45363 is the best representative of the cluster because it targets a specific fusion mechanism and concrete code paths; #44979 is broader infrastructure and not the same change.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Not the same underlying change: #44979 adds a new standalone Module Fusion API in `src/transformers/module_fusion.py`, while #45363 implements n-to-1 kernel fusion via `KernelConfig` across different integration/modeling files. They are adjacent concepts, but not the same code-path fix or mergeable into one PR as duplicates." + "reason": "Related subsystem, but not the same underlying change: #44979 adds a general Module Fusion API, while #45363 implements n-to-1 kernel fusion through KernelConfig and model integration. They would not plausibly be merged as one PR." } ] }, @@ -18564,7 +18518,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18580,7 +18534,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "c87e5c55606dc12c4818673b6c908aa7531bb5417b1b69414825778a76e0c9bd", + "prepared_review_unit_hash": "68467cf68eee70c106b636b9f24be93bb1bbefd9027d5d1e1d1f8556f01d9d00", "component_budget": { "node_count": 2, "item_count": 2, @@ -18602,18 +18556,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests fix the same Mistral regex pre-tokenizer AttributeError by switching from `backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer` in the same code path. The second PR is a fuller version of the same change, adding tests, so they look like duplicates rather than distinct fixes.", - "confidence": 0.98, + "summary": "Both PRs fix the same Mistral regex patch AttributeError in the same file and target the same issue; 45317 is the stronger representative because it includes tests and a fuller patch.", + "confidence": 0.96, "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 and a clearer, more complete patch.", + "canonical_pr_reason": "pull_request:45317 is the better canonical PR because it covers the same fix as 45086 but adds test coverage and appears to be the more complete implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best overall PR for this cluster because it most completely addresses the bug and is the most merge-ready version of the shared fix.", + "best_pr_reason": "pull_request:45317 is the best PR representative since it addresses the same concrete bug, is more complete, and is more suitable as the merged duplicate target.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in the same function and file; 45317 is essentially a superset with tests, so these are duplicate PRs for the same change." + "reason": "Accept: both PRs patch the same AttributeError in _patch_mistral_regex, touch the same code path/file, and are plausibly mergeable into one PR; 45317 is just the more complete version with tests." } ] }, @@ -18627,7 +18581,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18643,7 +18597,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "870c5b3a753297973d2bc76e3f38cbb3c4073f623eba6608a3e73d0269f8daf3", + "prepared_review_unit_hash": "28bd75a0e7a00f69a785681104862b0b7acfdebcd6bea13f8389a5a0ffe9b692", "component_budget": { "node_count": 2, "item_count": 2, @@ -18665,18 +18619,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about Nemotron H model work, but they target different changes: PR 45115 is a focused refactor/conversion-mapping update for NemotronH inheritance, while PR 45569 is a broader split/addition of NemotronH dense/sparse variants with docs, auto-mappings, and tests. They are related but not the same mergeable fix.", + "summary": "Two open PRs touch NemotronH internals, but they are not duplicates: one is a refactor/conversion-mapping update for GraniteMoeHybrid compatibility, while the other introduces dedicated NemotronHDense/NemotronHSparse model variants, docs, and tests.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the broader, more complete implementation of the NemotronH family restructuring, covering new dense/sparse model packages plus docs and test updates; it looks like the main umbrella change in this cluster.", + "canonical_pr_reason": "PR #45569 is the broader, more representative change: it adds the new dense/sparse model split, documentation, auto-mapping updates, and tests, whereas #45115 is a narrower compatibility/refactor patch.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the best representative because it includes the user-facing API and documentation changes along with the new model variants, making it the most central and comprehensive artifact here.", + "best_pr_reason": "PR #45569 is the best standalone representative for the NemotronH work because it covers the actual feature/model surface change rather than only the conversion and inheritance refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "They overlap on NemotronH internals, but 45115 is a narrow refactor/conversion-mapping change and 45569 is a larger feature split into dense/sparse variants; they do not look like the same concrete bug or a single PR merge target." + "reason": "They overlap on NemotronH files, but the concrete changes differ: #45115 focuses on conversion-key renaming and GraniteMoeHybrid-style inheritance, while #45569 adds separate dense/sparse model variants plus docs/tests. Same subsystem, not the same underlying change." } ] }, @@ -18690,7 +18644,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18706,7 +18660,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "0624654766ab27c9f2e4bf26c5b6e8639f732772ed1eab900e4769cc452098fc", + "prepared_review_unit_hash": "fe64f2824af359f9044892150d7fa09c31366e290213da6c119c29dc143c9316", "component_budget": { "node_count": 2, "item_count": 2, @@ -18728,18 +18682,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in the same heterogeneous-model feature area, but they appear to be staged, complementary changes rather than duplicates: one adds per-layer config support, the other adds per-layer modeling/cache support on top of that.", - "confidence": 0.76, + "summary": "These are related feature PRs for heterogeneous model/config support, but they are not duplicates: one adds per-layer config plumbing, the other adds broader modeling/runtime support on top of that.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader umbrella change: it covers heterogeneous modeling plus config-related integration and touches more of the runtime path, so it best represents the cluster.", + "canonical_pr_reason": "PR #45332 is the broader, more representative change: it covers heterogeneous modeling/runtime integration across cache, masking, model utils, and GPT-OSS code, not just config serialization/plumbing.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the strongest representative because it spans the larger end-to-end feature and is the more likely parent change for the surrounding heterogeneous support work.", + "best_pr_reason": "PR #45332 is the better cluster representative because it spans the end-to-end heterogeneous support path and subsumes the narrower config-only groundwork in #45333.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Related feature work, but not the same concrete change: #45333 adds heterogeneous per-layer config support, while #45332 extends modeling/cache behavior using that support. They are adjacent pieces that could be merged sequentially, not duplicates." + "reason": "Related but not the same underlying change: #45333 focuses on per-layer config support, while #45332 extends that into model/runtime behavior. They are complementary pieces of a feature, not duplicate PRs that could be safely merged into one without losing scope separation." } ] }, @@ -18753,7 +18707,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18769,7 +18723,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "be330ad818b6d1b112338a4a6e7f89cfb0561602122c5a3c70e969cd626d8b36", + "prepared_review_unit_hash": "4aafcdcbd3673122ffdb763b4dc94f9ce85fb34f2c183204b43494bea4461a81", "component_budget": { "node_count": 2, "item_count": 2, @@ -18791,18 +18745,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests are about typing cleanup in Transformers, but they are not the same change: PR 45415 is a broad type-checking sweep across many files, while PR 45425 adds specific typing aliases/protocols for modeling-related code paths. They overlap in touched files, but the concrete fixes are different, so this is not a duplicate pair.", - "confidence": 0.96, + "summary": "Both PRs are about type annotations/type-checking, but they address different changes: one is a broad typing sweep across many files, the other adds specific typing protocols/aliases for modeling utilities and PEFT-related code. They are related in theme, not duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45425 is the more concrete, narrowly-scoped typing change; PR 45415 is a broader repository-wide type-checking sweep and not the same underlying modification.", + "canonical_pr_reason": "PR 45415 is the broader repository-wide typing sweep and is the better representative if this cluster is grouped by topic, but it is not the same concrete change as 45425.", "best_issue_reason": null, - "best_pr_reason": "PR 45425 is the best representative of the cluster because it targets a specific typing deficiency with a focused code change, whereas PR 45415 is a large unrelated type-checking pass.", + "best_pr_reason": "PR 45425 is the more focused, concrete typing-infrastructure change; it is a better standalone representative than the broad type-check cleanup in 45415, but still not a duplicate of it.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Shared typing-related files are not enough: 45415 is a broad type-checking enablement PR, while 45425 adds specific Protocol/TypeAlias support for modeling_utils and related integrations. They do not look like one mergeable fix for the same bug." + "reason": "Both are typing-related, but 45415 is a broad type-checking pass while 45425 adds specific protocols/aliases for different code paths; they do not fix the same underlying bug or change." } ] }, @@ -18816,7 +18770,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18837,58 +18791,58 @@ "pull_request:45817|pull_request:45843", "pull_request:45836|pull_request:45843" ], - "prepared_review_unit_hash": "fc9ccae092a3f0a75e783e0b86b0bd57c05fd013a3e9bdd4fbbc214bd317a0a9", + "prepared_review_unit_hash": "ba76cf9517786da69ba67cd2809dc680236c43e87c97904d7d3543c5798ae770", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 10390, - "estimated_input_tokens": 2598, - "estimated_eval_tokens": 5452 + "serialized_chars": 10388, + "estimated_input_tokens": 2597, + "estimated_eval_tokens": 5450 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 10390, - "estimated_input_tokens": 2598, - "estimated_eval_tokens": 5452 + "serialized_chars": 10388, + "estimated_input_tokens": 2597, + "estimated_eval_tokens": 5450 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are loosely related by model-parallel compatibility, but they fix different model families and one is only test skips; they should not be clustered as duplicates.", - "confidence": 0.84, + "summary": "These PRs are all loosely about model-parallel or model-related test issues, but they target different models and different failure modes. They should not be deduplicated into one change.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45817 is the best representative of the cluster because it contains a concrete code-path fix for model-parallel behavior plus tests, while the others are either different model-specific fixes or just test skips.", + "canonical_pr_reason": "PR 45843 is the best representative because it contains substantive source changes for a real model-parallel fix, while the others are either different model families or test-skips only.", "best_issue_reason": null, - "best_pr_reason": "PR 45817 is the strongest standalone fix in this set: it makes the model-parallel path device-safe in Gemma4 and adds coverage. The other PRs do not describe the same underlying bug.", + "best_pr_reason": "PR 45843 is the strongest standalone PR in the set: it fixes a concrete model-parallel implementation issue in DeIMv2 and is broader/more central than the test-only or model-specific fixes.", "soft_edge_verdicts": [ { "left": "pull_request:45487", "right": "pull_request:45843", "accept": false, - "reason": "Both are model-parallel fixes, but they target different architectures (AltCLIP/ChineseCLIP vs Deimv2) and different code paths, so they are not the same underlying change." + "reason": "Both mention model parallel, but they fix different model families and different code paths (AltCLIP/ChineseCLIP token type handling vs DeIMv2 backbone no-split propagation)." }, { "left": "pull_request:45487", "right": "pull_request:45836", "accept": false, - "reason": "45487 fixes actual model-parallel code in model files; 45836 only skips failing tests for MiniCPM-V/Qwen3.5 cache incompatibilities. Different problems." + "reason": "45487 is a source-code model-parallel fix; 45836 only skips failing tests for MiniCPM-V 4.6/Qwen3.5 cache compatibility. Not the same change." }, { "left": "pull_request:45817", "right": "pull_request:45843", "accept": false, - "reason": "Gemma4 model-parallel handling and Deimv2 model-parallel handling are separate bug fixes in different implementations, not one mergeable PR." + "reason": "Gemma4 and DeIMv2 are unrelated implementations; these are separate model-parallel bugs, not one shared underlying defect." }, { "left": "pull_request:45836", "right": "pull_request:45843", "accept": false, - "reason": "45836 is a test-skip PR for quantized cache failures, while 45843 changes Deimv2 model-parallel code. They do not fix the same bug." + "reason": "One is a test skip for invalid quant-cache coverage, the other is an implementation fix for model parallelism. Different underlying issues." } ] }, @@ -18902,7 +18856,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18918,7 +18872,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "b9795b21175364ec095637349af731c88eeecf88b6f218233a83ef6906ff1ccd", + "prepared_review_unit_hash": "3667855bcea49dca5032169456bd7a1909d91eabe8b34f76490abc2d0cb844af", "component_budget": { "node_count": 2, "item_count": 2, @@ -18940,18 +18894,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related to expert-parallel distributed loading, but they target different concrete failures: one fixes EP + DeepSpeed ZeRO-3 config/loading behavior, the other fixes EP + FSDP2 expert parameters being overwritten during load. Not a duplicate pair.", - "confidence": 0.92, + "summary": "Two open PRs both touch expert-parallel loading, but they fix different failure modes: one targets Accelerate/DeepSpeed ZeRO-3 config handling, the other targets FSDP2/rank-0 broadcast overwriting experts. They are related in subsystem, not the same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45662 is the better canonical representative because it addresses the more explicit concrete loading bug (expert parameters being overwritten on rank-0 broadcast) with a broader implementation across loading paths.", + "canonical_pr_reason": "PR #45662 is the stronger canonical PR because it addresses a concrete overwrite bug during EP model loading with broader code-path coverage and a clearer end-to-end fix.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the strongest standalone fix in this cluster: it has a clear symptom, touches the relevant loading code paths, and targets a specific distributed-model initialization failure. PR #45548 is narrower and concerns a different loading/configuration issue.", + "best_pr_reason": "PR #45662 is the best representative of the cluster since it fixes a specific runtime corruption issue in the loading path, whereas #45548 is a narrower config/launch integration fix.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both are about EP/distributed model loading, but they fix different bugs and different code paths: ZeRO-3/accelerate config handling versus FSDP2 expert params being overwritten. Too broad to merge as duplicates." + "reason": "Both concern expert parallelism, but #45548 fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch, while #45662 fixes EP + FSDP2 experts being overwritten by rank-0 broadcast. Different concrete bugs and different code paths." } ] }, @@ -18965,7 +18919,7 @@ { "label": "pull_request", "component_index": 24, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18981,12 +18935,12 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "85c02197e1a54820d2b597e2797ecf27180da6d09d66dfa71bf115c88f0ea6d9", + "prepared_review_unit_hash": "e4d0fad74a13f12cbf26e2b5211882bb31651573dbade4a13548f2e743e1598e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4590, + "serialized_chars": 4589, "estimated_input_tokens": 1148, "estimated_eval_tokens": 2552 }, @@ -18994,7 +18948,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4590, + "serialized_chars": 4589, "estimated_input_tokens": 1148, "estimated_eval_tokens": 2552 }, @@ -19003,18 +18957,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same captured_info/xdist logging problem; 45645 is the broader superset with the workflow and notification updates needed to consume the new per-worker log files.", - "confidence": 0.79, + "summary": "Both PRs address the same xdist/captured_info collision in testing debug logs; PR 45645 is the broader, more complete change and subsumes the narrower 45639 fix.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the more complete version of the xdist-safe captured_info fix, including the downstream CI log aggregation changes needed for the new file layout.", + "canonical_pr_reason": "PR 45645 is the better canonical PR because it includes the core xdist-safe logging fix in testing_utils plus the CI workflow adjustment needed to consume multiple captured_info files.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the best standalone candidate because it covers the root code-path fix and the supporting CI/notification plumbing, making it the most complete fix for the underlying artifact collision issue.", + "best_pr_reason": "PR 45645 most fully represents the underlying change: it fixes the collision, preserves debug log collection in CI, and covers the same affected paths as 45639.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Accept: 45645 is a superset of 45639\u2019s exact captured_info xdist-safety fix and appears to target the same underlying bug, with additional downstream handling that could plausibly live in the same PR." + "reason": "Same underlying bug: xdist collisions for captured_info/debug logs. 45645 directly incorporates the same fix and adds the workflow follow-up, so they could plausibly be merged into one PR." } ] }, @@ -19028,7 +18982,7 @@ { "label": "pull_request", "component_index": 25, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19091,7 +19045,7 @@ { "label": "pull_request", "component_index": 26, - "component_count": 26, + "component_count": 27, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19153,6 +19107,69 @@ "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45858-2", + "nodes": [ + "pull_request:45858", + "pull_request:45890" + ], + "soft_pairs": [ + "pull_request:45858|pull_request:45890" + ], + "prepared_review_unit_hash": "6e3fb1ec61a4376dc12129a1db8ff42f107e2641bbbb7194020f031c654f1640", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4140, + "estimated_input_tokens": 1035, + "estimated_eval_tokens": 2326 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4140, + "estimated_input_tokens": 1035, + "estimated_eval_tokens": 2326 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are XPU test-expansion changes, but they target different model test files and different concrete expectations, so they are related in theme but not duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No single PR is a clear duplicate representative here: each changes a different model test and adds distinct XPU expectations for a different code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45858", + "right": "pull_request:45890", + "accept": false, + "reason": "They both add XPU expectations, but for different models/tests (hy_v3 vs exaone4_5) and different expected outputs. This is a shared testing pattern, not the same underlying bug or change, and they would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null } ] } diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index a6a5e5fb1cdfe00518101f33327e634be34cbbd8..5be1ef8330fbeb510390abc3552a3230bce39e3b 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260511t060028z", + "analysis_id": "hybrid-model-20260511t120023z", "archived_artifacts": { - "hybrid": "snapshots/20260511T060028Z/analysis-runs/hybrid-model-20260511t060028z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260511T060028Z/analysis-runs/hybrid-model-20260511t060028z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-05-11T06:10:42Z", + "published_at": "2026-05-11T12:10:27Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260511T060028Z", + "snapshot_id": "20260511T120023Z", "variant": "hybrid" } diff --git a/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..74e7fd0732101fc56955e399da96a22210c55f0c --- /dev/null +++ b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json @@ -0,0 +1,1792 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260511T120023Z", + "generated_at": "2026-05-11T12:10:24Z", + "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-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.805, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.854, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 713, + 720 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 699, + 705 + ], + [ + 1301, + 1307 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-43824-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43824.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43824, + "canonical_pr_number": 43836, + "issue_numbers": [ + 43824 + ], + "pr_numbers": [ + 43836, + 43842 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43836, + "right_pr_number": 43842, + "code_similarity": 0.332, + "size_similarity": 0.088, + "file_overlap": 1.0, + "area_overlap": 0.033, + "patch_similarity": 0.017, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 359, + 429 + ], + [ + 584, + 590 + ], + [ + 1892, + 1910 + ], + [ + 1917, + 1923 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, + { + "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-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": 0.532, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.543, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3529, + 3536 + ], + [ + 3753, + 3787 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "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 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43824-3", + "canonical_pr_number": 43836, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43824-3 are treated as duplicates because they converge on issue #43824 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-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." + } + ], + "best_issue": { + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 12 discussion comments plus 8 inbound references.", + "score": 162.0 + }, + "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/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..331811966ca4443c0b6ab85aabacee0a44001d0a --- /dev/null +++ b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,19175 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260511T120023Z", + "generated_at": "2026-05-11T12:10:24Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "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": "7d84f1fbd8936e94fa9a48cec974049e4b6fa203f37cfeb08274c64d8e2ab352", + "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": "This is not a duplicate cluster: the three issues concern different problems (Tapas tokenization mismatch, private-model auth/error messaging, and inability to load bert-base-uncased). The soft similarities are too broad to justify merging.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the broadest and most user-visible model-loading failure and has the most discussion, so it is the best representative only if a single issue must be kept; however, it is not the same bug as the other two.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tokenization discrepancy in Tapas vs TensorFlow code is a different bug from a generic bert-base-uncased loading failure." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model login/error-message handling is an auth/download UX issue, not the same underlying model-loading defect." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45070", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45874" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43541|issue:43716", + "issue:43716|issue:45685", + "issue:43828|issue:44292", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:44292|issue:44918", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:43039|issue:44805", + "issue:41762|issue:43596", + "issue:43901|issue:44512", + "issue:39290|issue:41720", + "issue:44841|issue:45823", + "issue:43039|issue:44560", + "issue:43866|issue:44863", + "issue:43854|issue:43866", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:43404|issue:43866", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:36010|issue:39290" + ], + "prepared_review_unit_hash": "78fbaba6baf415ac7a86665ff31acdbace7bfa3ee928980fd9f9aaca8fd423c9", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 69, + "serialized_chars": 48321, + "estimated_input_tokens": 12081, + "estimated_eval_tokens": 24418 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: several unrelated issue families are mixed together by loose text similarity. The only clear near-duplicate I found is the ZeRO-3/Bert index-out-of-bounds pair; the rest are different bugs, docs cleanups, or model-specific failures.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43596 is the clearest representative of the ZeRO-3/Bert size-0 index bug and is nearly identical to 43638.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 has the cleanest, most specific reproduction for the only strong duplicate family in the cluster, so it is the best issue anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 problems: config attribute lookup vs streaming chat/completions failure." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different hardware/backend failures: AMD GPU loading vs MPS histogram not implemented." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different model families and failure modes: Mistral-3 dtype mismatch vs GPT-OSS-on-AMD." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same bug as GPT-OSS on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model/backend issues: AMD GPU runtime vs GraniteMoEHybrid invalid method call." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different model-specific runtime errors; no concrete shared code-path." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Both are Deepspeed/model-loading issues, but one is collective mismatch and the other is ONNX export; not the same bug." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Different model/export failure vs Qwen3Moe FP8 failure." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error and Qwen3 auto device-map cuda assert are unrelated failures." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanups, but they point to different removed docs items and are not the same concrete change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export bug is unrelated to GenerationMixin import relocation." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Generic mask-shape mismatch vs Gemma4 device mismatch; different symptoms and code paths." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both mention index/mask style errors, but one is ZeRO-3/Bert init and the other is a generic mask-length mismatch." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Same broad error class, but different concrete failure: Bert ZeRO-3 init vs token mask shape mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask shape mismatch is unrelated to GPT-OSS MoE TP-only failure." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Different tensor-shape failure vs Qwen3.5 embedding unpacking with TRL." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve Deepspeed, but collective mismatch and Bert zero-size index errors are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Near-identical ZeRO-3/Bert index-out-of-bounds reports with the same underlying init failure." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy dispatch and GPT-OSS TP-only failure are different issues." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Detokenization regression is unrelated to LoRA merge/save/load output mismatch." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA save/reload mismatch are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP recompute mismatch is not the same as the mask-shape index error." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed training weight-dimension error and collective mismatch are different failures." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Bert ZeRO-3 init failure is unrelated to GPT-OSS MoE tensor-parallel issues." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP checkpointing shape mismatch is unrelated to CustomTrainer.compute_loss signature breakage." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model/backend areas: Bert ZeRO-3 init vs GPT-OSS MoE tensor parallel." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3TextConfig sliding_window_pattern and Mistral tokenizer regex patch are different model/tokenizer bugs." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Gemma4 device-map mismatch during forward and expensive reverse embedding lookup are different code paths." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch-version incompatibility is unrelated to Qwen-3-8B-NVFP4 runtime failure." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same as GPT-OSS MoE TP-only failure." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "Qwen3.5 embedding unpacking with TRL and GPT-OSS TP-only failure are different bugs." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure is unrelated to Mistral-3 image-preprocessor dtype mismatch." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different model/backend failures: Mistral-3 dtype mismatch vs MPS histogram not implemented." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Autocast dtype mismatch and Qwen-3-8B-NVFP4 runtime error are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the MPS Int histogram limitation." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Mistral-3 dtype mismatch and Qwen-3-8B-NVFP4 runtime error are different issues." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen-related failures with different triggers and code paths." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross_entropy dispatch and Qwen3.5 embedding unpacking are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading failure and Bert ZeRO-3 loading failure share a backend, but not the same concrete bug." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs updates for v5 removals, but they cover different removed items and pages." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3-vl StopIteration." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Phi autocast dtype mismatch is unrelated to the MPS histogram backend limitation." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3.5 embedding/TRL issue is unrelated to MPS Int histogram support." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Liger Kernel cross_entropy dispatch is unrelated to Phi-tiny-MoE autocast dtype mismatch." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and Trainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP mismatch is unrelated to GenerationMixin import error." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross_entropy call and mask-shape index error are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 loading issues, but Gemma3 and Bert fail differently." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "TextClassificationPipeline return_all_scores docs and run-command docs are distinct documentation fixes." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3 sliding_window_pattern and Qwen3 auto device map assert are different bugs." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Voxtral processor failure is unrelated to Gemma4 device mismatch." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross_entropy call is unrelated to Qwen3-vl video StopIteration." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading problems: corrupted Ovis2 checkpoint vs NemotronH checkpoint compatibility." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash test loading failure and Ovis2 checkpoint corruption are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Gemma3 sliding_window_pattern and Qwen3-MoE sliding_window are different model config bugs." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Mask-index mismatch and Bert ZeRO-3 zero-size index error are different concrete failures." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility and pydantic PretrainedConfig field breakage are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3/Bert init failure is unrelated to torch/transformers version incompatibility." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper loading failure and pydantic PretrainedConfig breakage are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Qwen3.5 chat crash and streaming chat/completions failure are different bugs." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 lm_head tying bug and Ovis2 checkpoint corruption are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 sliding-window config and Gemma4 PLE device mismatch are different model issues." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming inference failure and MPS histogram backend limitation are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Torch Dynamo grouped_mm tracing failure is unrelated to MPS Int histogram support." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different bugs." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import error is unrelated to Gemma3 sliding_window_pattern." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43596 and issue:43638, is well supported by their near-identical ZeRO-3/Bert index-out-of-bounds reports. The other soft-edge rejections are phrased cautiously and do not overclaim duplicate status. No PR-pair mergeability concerns apply here because this cluster contains issues, not pull requests." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 2, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:39186", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45542", + "issue:45632", + "issue:45698", + "issue:45800", + "issue:45823" + ], + "soft_pairs": [ + "issue:44841|issue:45084", + "issue:45800|issue:45823", + "issue:41093|issue:41762", + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:45020|issue:45800", + "issue:43531|issue:45070", + "issue:43541|issue:44589", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:45542|issue:45800", + "issue:40990|issue:44960", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:43645|issue:45399", + "issue:44530|issue:45005", + "issue:44360|issue:44512", + "issue:42915|issue:45399", + "issue:43873|issue:45005", + "issue:44863|issue:45005", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:43950|issue:44534", + "issue:43644|issue:43950", + "issue:44292|issue:45507", + "issue:39692|issue:43054" + ], + "prepared_review_unit_hash": "cdd922facfff498c8bb0063da0fef50144165f456446b01473460b9066578b83", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37495, + "estimated_input_tokens": 9374, + "estimated_eval_tokens": 19004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issues were grouped by weak lexical similarity. The only clear duplicate cluster is the Transformers v5 non-persistent buffer corruption regression (`43644` / `43950` / `44534`); the rest should stay separate.", + "confidence": 0.84, + "canonical_issue_reason": "`43950` is the best canonical issue for the real duplicate set because it states the regression most explicitly (`from_pretrained` corrupts non-persistent buffers) and cleanly subsumes `43644` and `44534`.", + "canonical_pr_reason": null, + "best_issue_reason": "`43950` is the strongest representative of the cluster\u2019s actual shared bug: a v5 regression corrupting `register_buffer(persistent=False)` state during load.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Different problems: a model-specific processor/load failure versus a compiler error about non-template nodes." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Unrelated failures: version compatibility between torch/transformers versus a Gemma4 device-map forward mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are IndexErrors under ZeRO-3, but the reported symptoms and code paths differ too much to treat as the same bug." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "One is a vLLM loading failure for PaddleOCR-VL; the other is a request for GGUF support for gpt-oss." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "A missing config field after v5 update is not the same as a pydantic model-field regression." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/OOM behavior and tied-weights handling are different code paths and different failure modes." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Torch Dynamo grouped_mm tracing error is distinct from a tensor-parallel support issue in GPT-OSS MoE." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch and streaming chat/completions failure are not the same concrete bug." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument during `init_empty_weights` is unrelated to the pydantic `PretrainedConfig` field regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support and Qwen3.5 streaming inference are different feature areas and failures." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Remote-code breakage is a different issue than a torch/transformers version incompatibility." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "A sliding-window bug in Qwen3-MoE is unrelated to the v5 pydantic config-field problem." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "These are different runtime errors in different subsystems; only superficial similarity exists." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both mention index errors with ZeRO-3, but one is a mask-shape mismatch while the other is an empty-tensor load failure." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both mention SigLIP2, but one reports poor embedding quality while the other reports nonsensical AutoModel/pipeline outputs; not clearly the same bug." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "A missing TensorFlow backend check and a torch/transformers incompatibility are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on GPT-OSS and a GLM5 issue are not the same underlying problem." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image preprocessor/model dtype mismatch is unrelated to Qwen3.5 input embedding unpacking in TRL SFT trainer." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP `weight must be 2-D` and TRL embedding unpacking failures are different code paths and different bugs." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom-code loading, but one is a cache-path collision for local leaf directories and the other is wrong-module loading after save/load; too distinct to merge." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but they target different model families and don\u2019t clearly describe the same concrete regression." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback selection are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache crashing on `linear_attention` is unrelated to tied-weights handling." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA-indexer discussion and a docs typo are not the same underlying change or bug." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FineGrainedFP8Config failure and flash-attn fallback blocking are different quantization/attention issues." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization offloading behavior is not the same as tied-weights regressions." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is unrelated to tied-weights issues." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo and a removed run-command doc issue are just documentation-level differences." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRL SFT embedding unpacking and an invalid model method call are different bugs." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Incorrect detokenization and an NVFP4 runtime error do not point to the same fix." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Both describe the same v5 regression where non-persistent buffers get corrupted/junk during loading." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Clear duplicate: both report the same non-persistent buffer corruption regression in Transformers 5.x." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "A Qwen-3-8B-NVFP4 runtime error is not the same as GraniteMoEHybrid calling an invalid method." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 docs-example errors and poor SigLIP2 text embeddings are related only by model name, not by the same concrete bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusions are grounded in the packet. The canonical/best issue choice is supported by the titles and descriptions, and the only accepted soft-edge pairs are the clear non-persistent-buffer regression duplicates (`43644`, `43950`, `44534`). The remaining pair rejections are phrased conservatively and do not overclaim duplicate status." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 3, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:41553", + "issue:41628", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44589", + "issue:44661", + "issue:44910", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45237", + "issue:45356", + "issue:45399", + "issue:45520", + "issue:45685", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:43828|issue:44928", + "issue:36331|issue:41762", + "issue:41628|issue:45800", + "issue:43866|issue:44403", + "issue:43872|issue:44589", + "issue:43054|issue:45701", + "issue:41553|issue:43883", + "issue:38175|issue:43054", + "issue:44910|issue:45859", + "issue:41553|issue:43881", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:44403|issue:44451", + "issue:43643|issue:45020", + "issue:43975|issue:45520", + "issue:43975|issue:45237", + "issue:44403|issue:44488", + "issue:42947|issue:43856", + "issue:43881|issue:43883", + "issue:41762|issue:42915", + "issue:43782|issue:44960", + "issue:42491|issue:44368", + "issue:43975|issue:44451", + "issue:45685|issue:45823", + "issue:43824|issue:43872", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:44991|issue:45399", + "issue:43828|issue:45823", + "issue:45020|issue:45356", + "issue:45081|issue:45399", + "issue:43827|issue:44509", + "issue:43404|issue:43950", + "issue:43742|issue:45399" + ], + "prepared_review_unit_hash": "687b549a922000cb982914a9a7330d21a686f0bfbfda5dc64f387063d4daf7b8", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37385, + "estimated_input_tokens": 9347, + "estimated_eval_tokens": 18950 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous cluster of Transformers issues. The only clearly duplicate soft edge is the docs cleanup pair; the SigLIP2 pair is thematic but not strong enough to merge as a duplicate from the titles/evidence shown.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45020 is still the best umbrella representative: it captures the broad v5-era remote_code/loading regression theme that overlaps with several nearby reports, even though it is not a true duplicate of most items.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best global issue candidate because it is the broadest, most central report in the cluster and best represents the common version-related loading breakage theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different bugs and code paths: autocast dtype mismatch vs Flash Attention/position_ids handling." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Trainer API kwarg mismatch vs ZeRO-3 loading index error; unrelated failures." + }, + { + "left": "issue:41628", + "right": "issue:45800", + "accept": false, + "reason": "Missing import/export for AutoImageProcessor vs torch/transformers version incompatibility." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Corrupted checkpoint vs generic loading noise; not the same defect." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "bitsandbytes Int8Params constructor incompatibility vs missing Float8 storage object; different backends." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality regression vs tokenization changes; different underlying behavior." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "AutoTokenizer error for Voxtral vs missing all_tied_weights_keys on Molmo." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but zero probabilities and degraded embeddings are not clearly the same concrete bug." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Both involve Qwen3.5, but one is a Flash Attention crash from 3D position_ids and the other is CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading failures for different models and code paths." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Gradient checkpointing/LoRA issue vs int4 quantization memory regression; not the same bug." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Only a thematic SigLIP2 match; the titles point to a generic quality complaint vs a specific HF-vs-JAX implementation discrepancy, so duplicate status is not established." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise vs inability to load a specific model; not duplicates." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Missing fields in AutoConfig with trust_remote_code is a specific config-return bug, not the same concrete failure as the broader remote_code breakage report." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Detokenization regression vs flash_attn import KeyError; unrelated subsystems." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Detokenization regression vs AMD GPU support failure; different problems." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Loading noise vs model-specific loading failure; not duplicates." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "LoRA checkpointing behavior vs MoE training memory usage; related area, different bug." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Both loading-related, but one is a model load failure and the other is a missing attribute during weight tying." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 load crash vs Qwen3 MoE LoRA compatibility; different models and code paths." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight_only load error vs GLM5 issue; unrelated." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen3 MoE/LoRA, but one is a cross-version incompatibility and the other is a warning; not the same concrete bug." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Detokenization regression vs specific model loading failure; different defects." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram op limitation vs Gemma4 device mismatch during forward; separate failures." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Missing model class import vs bitsandbytes Int8Params constructor error; unrelated." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote_code loading regressions vs Gemma4 PLE device placement bug; different code paths." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer mapping helper failure vs remote_code loading breakage; not the same bug." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure for a specific model vs flash-attn fallback gating; unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi MoE vs Gemma4 device mismatch; different issues." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote_code loading regression vs Kimi-K2.5 tokenizer regression; different model and mechanism." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch crash vs flash-attn fallback gating; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Both are the same docs cleanup problem: outdated pipeline-task references removed in v5." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Tied-weight bug in Mistral3 vs non-persistent buffer corruption in from_pretrained; different loading semantics." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM load key error vs flash-attn fallback checks; different failures." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The overall cluster summary is reasonable and conservative, but one soft-edge verdict is too strong for the evidence available: `issue:43054` vs `issue:43493` is only a plausible thematic match (both SigLIP2-related), not clearly the same bug/PR from the titles and packet alone. The `issue:43827` vs `issue:44509` duplicate call looks grounded, but because at least one accepted pair overstates the evidence, I can\u2019t accept the full evaluation." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 4, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:43257", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43638", + "issue:43645", + "issue:43646", + "issue:43792", + "issue:43824", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43940", + "issue:43950", + "issue:43957", + "issue:44155", + "issue:44220", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45538", + "issue:45584", + "issue:45588", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834" + ], + "soft_pairs": [ + "issue:43792|issue:44220", + "issue:42915|issue:45020", + "issue:43257|issue:43866", + "issue:43638|issue:45823", + "issue:43872|issue:44291", + "issue:43940|issue:45313", + "issue:42915|issue:43872", + "issue:41553|issue:44991", + "issue:43854|issue:44863", + "issue:43257|issue:45305", + "issue:36331|issue:41093", + "issue:44928|issue:45588", + "issue:43872|issue:45538", + "issue:41093|issue:44805", + "issue:41553|issue:44488", + "issue:44530|issue:44863", + "issue:44479|issue:44560", + "issue:24643|issue:30064", + "issue:43366|issue:43828", + "issue:45663|issue:45823", + "issue:44155|issue:45584", + "issue:43366|issue:43531", + "issue:43950|issue:45834", + "issue:43550|issue:44910", + "issue:44560|issue:45685", + "issue:43824|issue:43957", + "issue:43645|issue:43646", + "issue:43366|issue:44589", + "issue:42915|issue:43957", + "issue:42617|issue:43366", + "issue:44560|issue:44918", + "issue:43950|issue:45325" + ], + "prepared_review_unit_hash": "48dbe281c8815820c0efc3885d2ccccb22d19f01ca4f7fa23c3599ebbd1a884d", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37381, + "estimated_input_tokens": 9346, + "estimated_eval_tokens": 18948 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy mix of unrelated Transformers model-loading/training regressions. Clear duplicates exist for the `_is_hf_initialized` init-empty-weights bug, the mask-shape IndexError, the custom-model initialization regression, and likely a couple of Qwen ZeRO-3/video regressions.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here: it is open, spans recent Transformers versions, and captures the recurring remote-code / model-loading regression theme better than the narrower model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the best representative because it is the most general and still-open issue among many specific loading failures, so it can anchor follow-up duplicates and related reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio-related, but one is a generic Whisper failure and the other is a specific fbank helper issue; not enough evidence they are the same bug." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "Both mention model failures, but one is a Qwen3MoE FP8/config problem and the other is a broad remote_code regression; different code paths." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Different models and failure modes: Qwen3 MoE weight conversion with Accelerate/DeepSpeed vs a corrupted Ovis2 checkpoint." + }, + { + "left": "issue:43638", + "right": "issue:45823", + "accept": false, + "reason": "Both involve loading/training, but one is a ZeRO-3/BERT index error and the other is a Gemma4 device-mismatch forward bug." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying `_is_hf_initialized` / `init_empty_weights` constructor regression; the second report is the concrete TypeError variant of the first." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": true, + "reason": "Both describe the same DeepSpeed ZeRO-3 weight-loading failure on Qwen3-family models, producing missing parameters during load." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "FP8/Qwen3MoE issues are unrelated to the bitsandbytes `_is_hf_initialized` constructor bug." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/loading complaints, but they concern different models and different problems (bad error message vs actual tokenizer load failure)." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and different load failures; no shared concrete bug is evident." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Both involve training/loading with distributed setups, but one is weight conversion under DeepSpeed while the other is gradient averaging with GAS." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different regressions: a custom compute_loss signature change vs a tensor/mask length mismatch IndexError." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Qwen3.5 SDPA/position_id issues are unrelated to the flash_attention `s_aux=None` crash." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "The bitsandbytes constructor regression is unrelated to CLIPTokenizer's `model_max_length` behavior." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same generic mask-shape IndexError; the later issue is the same failure class with a different reproduction." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they target different tokenizers/models and do not share a concrete code path." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache's `linear_attention` crash is unrelated to NemotronH checkpoint loading." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": true, + "reason": "Both describe the same Qwen VL video-input regression in Transformers 5.3.0; the second is a specific Qwen3-VL manifestation." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training weight-shape error and void segmentation-map processing are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss architecture support and a Phi MoE autocast dtype mismatch are different problems." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4-related, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo token leakage between tracks and Whisper empty-transcription generation failure are different audio bugs." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window behavior are unrelated." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to Kosmos2.5 OCR index errors." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different model families and different failure modes: SDPA torch.compile crash vs flash-attention illegal memory access from 3D position_ids." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen video regression and an MPS histogram dtype limitation are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing a missing class and meta-device loading regressions are different issues." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same custom model initialization regression; the notebook reproduction is just a context detail." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF architecture support is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3MoE FP8/config failure and meta-device loading regression do not share the same bug." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "A 3d_parallel.py runtime failure and gpt-oss GGUF support are not the same issue." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Qwen3-VL video StopIteration and TRL input-embedding unpacking are different code paths." + }, + { + "left": "issue:43950", + "right": "issue:45325", + "accept": false, + "reason": "Buffer corruption in from_pretrained is unrelated to the Qwen2.5-VL rope-index scaling bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the cluster is mostly heterogeneous, and the rejected pairs are appropriately cautious. The accepted edges are limited to closely matching failure classes and stay within 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": 2, + "component_count": 17, + "review_unit_index": 5, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43720", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44075", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44560", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45030", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45812", + "issue:45823", + "issue:45834" + ], + "soft_pairs": [ + "issue:44661|issue:45823", + "issue:36010|issue:42915", + "issue:38175|issue:43994", + "issue:43632|issue:45834", + "issue:44560|issue:44805", + "issue:44877|issue:45030", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:44960|issue:44977", + "issue:45488|issue:45812", + "issue:45313|issue:45357", + "issue:44451|issue:44488", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:43708|issue:44075", + "issue:43854|issue:43931", + "issue:42491|issue:45276", + "issue:42915|issue:44756", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:42491|issue:43931", + "issue:43856|issue:43873", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:45362|issue:45406", + "issue:43856|issue:44387", + "issue:44910|issue:45588", + "issue:42491|issue:45292", + "issue:36010|issue:39404", + "issue:45276|issue:45335" + ], + "prepared_review_unit_hash": "00a67d79b91120d25edaad19053fd17c3fc2ee2061eca997f4b69171c5f4bf24", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37782, + "estimated_input_tokens": 9446, + "estimated_eval_tokens": 19148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous Transformers issue set with a few true duplicate pairs around the same regressions (non-persistent buffers junk, SigLIP2 bad outputs, and resize_token_embeddings propagation). Most other soft pairs are only thematically similar, not the same underlying bug.", + "confidence": 0.59, + "canonical_issue_reason": "Best duplicate anchor in the set: a concrete Transformers v5 regression with a clear symptom and a close near-duplicate (44534), making it the cleanest issue to represent a subcluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Best overall issue to represent the cluster's broad regression theme: it is the most general, actionable Transformers v5 bug here and has stronger centrality than the model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: tokenizer-mapping logic vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Qwen3MoE FP8 loading." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 base model/output regression; both report nonsensical/zero results with AutoModel/pipeline." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized/deepspeed-v5 issue is unrelated to Kosmos2.5 long-input index errors." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different code paths and symptoms: video StopIteration vs mask/tensor shape mismatch." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both involve config validation, but they are different model-specific loading failures." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Zero3 Bert initialization error is unrelated to add-new-model-like tokenizer mapping failure." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Non-pretrained Bert Zero3 index error is a different bug from TOKENIZER_MAPPING_NAMES handling." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer-mapping bug and mask-shape IndexError are not the same underlying problem." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Dequantized offload load failure is different from mmap-related OOM mitigation." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Same non-persistent buffer junk regression in Transformers v5, just reported with slightly different wording." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but doc-example errors and runtime zero-probabilities are different bugs." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention crashes are different model-specific issues." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer pre-tokenizer override regression is unrelated to Granite token-id corruption." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "ZeRO-3 weight loading and save_pretrained visual-key regression are distinct." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are BERT loading failures, but the models and failure details are too different to treat as one bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3 MoE hf4/hf5 compatibility is unrelated to BitNet packed-weight loading." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Auto device mapping CUDA assert and FineGrainedFP8Config failure are different issues." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Checkpoint resume max_steps bug is unrelated to SGD args being ignored." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading and Qwen3-VL weight-shape mismatch are distinct model-loading bugs." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Qwen3 MoE loading compatibility is unrelated to resize_token_embeddings propagation." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3MoE FP8 failure is unrelated to mmap OOM on Strix Halo." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7-Flash test loading." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "SigLIP2 doc-example issues are not the same as the zero/nonsensical output regression." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility and Qwen3-VL shape mismatch are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Training memory inefficiency and quantization offload behavior are different problems." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but for different models and likely different root causes." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Non-persistent buffer corruption and Jupyter custom-model initialization are separate regressions." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Flash-attention packed-sequence bug and s_aux=None AttributeError are different code paths." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Qwen3 MoE loading issue is unrelated to resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "GenerationMixin import failure is unrelated to Whisper return_language pipeline behavior." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings propagation bug affecting model-specific embedding/output modules." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The three accepted soft pairs are plausible duplicate/subduplicate matches based on shared symptoms and model/feature context, and the rejected pairs do not overclaim sameness. The overall cluster description as heterogeneous with a few true duplicate regressions is well supported." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 6, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42947", + "issue:43257", + "issue:43335", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43828", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44683", + "issue:44743", + "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:45701", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:44164|issue:44291", + "issue:43404|issue:43454", + "issue:41720|issue:44155", + "issue:44402|issue:44488", + "issue:43425|issue:43957", + "issue:39692|issue:44960", + "issue:42947|issue:45704", + "issue:43873|issue:44387", + "issue:43335|issue:45663", + "issue:45663|issue:45759", + "issue:45663|issue:45859", + "issue:44291|issue:44756", + "issue:44155|issue:45823", + "issue:44387|issue:44756", + "issue:43854|issue:45216", + "issue:44292|issue:44912", + "issue:42947|issue:43381", + "issue:43957|issue:44756", + "issue:44743|issue:44960", + "issue:45276|issue:45292", + "issue:45684|issue:45698", + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:43646|issue:43708", + "issue:44479|issue:44877", + "issue:45588|issue:45663", + "issue:45292|issue:45335", + "issue:43645|issue:44164", + "issue:43257|issue:44863", + "issue:44402|issue:44451", + "issue:43828|issue:43957", + "issue:43425|issue:43828", + "issue:39404|issue:45701", + "issue:44589|issue:44756", + "issue:44683|issue:45823" + ], + "prepared_review_unit_hash": "605e66edac7bb0768e9fd35c7bd084f550d000fad4ef100fed29d96c13797efd", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38197, + "estimated_input_tokens": 9550, + "estimated_eval_tokens": 19356 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Only one duplicate pair is well supported in this packet: the Gemma4 `resize_token_embeddings` regression (`45276` \u2194 `45292`). The nearby Transformers 5.0.0 custom-model initialization issues are not supported here as a duplicate pair and should not be claimed as such.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 45276 is the best canonical issue because it gives the broader Gemma4 `resize_token_embeddings` failure, explicitly covering both `embed_tokens_per_layer` and `output_embeddings`, while 45292 is the narrower variant.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45276 is the strongest overall issue candidate for the supported duplicate pair because it describes the broader symptom set and subsumes 45292.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both are save/load failures, but they describe different bugs: `extra_state` handling vs unexpected `_is_hf_initialized` in `init_empty_weights`." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both mention `lm_head` tying, but they affect different model families and different code paths." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Unrelated failures: Qwen3 auto device-mapping CUDA assert vs AudioFlamingo3 batched inference token leak." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Similar loading complaints, but different checkpoints with no clear shared root cause." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility and meta-device loading regression are different issues." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Different areas entirely: SigLIP2 documentation errors vs a GLM5 model issue." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Different bugs: PEFT LoRA gradient checkpointing behavior vs a T5 RMSNorm memory leak." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both are quantization/memory themed, but one is offloading behavior and the other is an int4 reserved-memory/OOM regression." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Different models and bugs: sparse layer creation vs Gemma4 FSDP2 shared-kv-state rebuilding." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Different composite-model failures: per-layer shared-kv-state bug vs missing `text_config` unwrap." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3.5-MoE problems: attention state rebuild vs CPU offload incompatibility in rotary embedding." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Both involve loading/runtime failures, but the concrete bugs differ: unexpected init argument vs mmap OOM on Strix Halo." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "Different device-related bugs: embedding/token leak between tracks vs Gemma4 PLE device mismatch under auto device mapping." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Different regressions: int4 memory growth/OOM vs mmap-related OOM on Strix Halo." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Both are model-loading/saving issues, but they concern different models and failure modes." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both are quantized model loading issues, but they target different models/quantization paths." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Same broad gradient-checkpointing area, but one is ineffective under PEFT LoRA and the other is an eval-mode restriction." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regression and Strix Halo mmap OOM are different underlying bugs." + }, + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Different models and symptoms: Qwen3_5 recurrent-state reset vs generic GLM5 issue." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same `resize_token_embeddings` regression; 45276 is the broader Gemma4 report and 45292 is the narrower duplicate focused on `output_embeddings`." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Related save/load area, but different problems: read-only permission propagation vs wrong custom module loaded after `save_pretrained`." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Unrelated: eval-mode gradient checkpointing restriction vs RLHF NaN / SDPA fallback instability." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Different save/load and memory issues with no shared root cause." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Both mention training/custom behavior, but they are different bugs: custom-model init in notebooks vs `resume_from_checkpoint` max-steps calculation." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Different model families and failures: video input regression vs strict config loading of `granite_speech`." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Different attention-path bugs: `s_aux=None` flash-attention crash vs Gemma4 shared-kv-state rebuild." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": false, + "reason": "Both are `resize_token_embeddings` complaints, but they affect different model components and do not clearly share one bug." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook/custom-model initialization regression and `extra_state` save/load handling are different bugs." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Different loading failures for different checkpoints and execution paths." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading failures, but the titles indicate different checkpoints and no shared root cause." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Different regressions: autocast dtype mismatch on Phi-tiny-MoE vs meta-device loading breakage." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch 2.10 compatibility is unrelated to the specific autocast dtype mismatch bug." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` regression is unrelated to tokenization changes." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Different low-level failures: missing Float8 storage object vs mmap/OOM behavior." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Different issues: compiled flex_attention on torch>=2.9 vs Gemma4 device mismatch during forward." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The Gemma4 resize_token_embeddings pair looks grounded, but the summary overreaches on the Transformers 5.0.0 custom-model initialization claim: the packet does not provide a supported duplicate pair for that bug, and the listed soft pair for nearby issues was rejected. Please keep the summary limited to what the packet actually supports and avoid asserting an additional true duplicate without evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 7, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42947", + "issue:43299", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:43931|issue:44863", + "issue:43906|issue:44530", + "issue:43611|issue:43645", + "issue:43425|issue:45823", + "issue:44530|issue:45084", + "issue:43425|issue:43716", + "issue:39404|issue:44912", + "issue:43901|issue:44509", + "issue:44265|issue:44898", + "issue:43299|issue:43931", + "issue:44479|issue:44805", + "issue:43994|issue:44079", + "issue:39404|issue:44671", + "issue:44589|issue:45084", + "issue:40990|issue:42947", + "issue:43425|issue:44918", + "issue:44671|issue:44991", + "issue:44164|issue:45759", + "issue:36032|issue:43452", + "issue:43596|issue:44683", + "issue:43856|issue:44756", + "issue:32090|issue:39404", + "issue:44991|issue:45081", + "issue:44737|issue:45859", + "issue:44164|issue:44479", + "issue:43582|issue:43632", + "issue:44488|issue:45081", + "issue:44661|issue:44683", + "issue:40990|issue:43421", + "issue:44488|issue:44987", + "issue:43611|issue:43644", + "issue:43716|issue:45084", + "issue:44206|issue:44220" + ], + "prepared_review_unit_hash": "3dec5884f03e9d20aa7f1914a837c8fd580a25685b8377d87b926622b61eedf0", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37642, + "estimated_input_tokens": 9411, + "estimated_eval_tokens": 19078 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of Transformers v5 regression issues, mostly around model/tokenizer loading and initialization, but the soft pairs generally point to different concrete failures rather than the same underlying bug.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43611 is the broadest and most representative loading-regression report in the set, centered on v5 breakage in model loading with a clear generic symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is a strong anchor for the cluster because it describes a reusable v5 model-loading regression and is closer to the recurring theme than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:44863", + "accept": false, + "reason": "Both are loading failures, but they concern different models and different failure modes (shape mismatch vs checkpoint incompatibility)." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of an earlier issue is unrelated to the Qwen3.5 PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both involve v5 model initialization/loading regressions, but one is base_model_prefix loading and the other is notebook custom-class initialization; not the same bug." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch version compatibility and Gemma4 device-map forward mismatch are different problems." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different subsystems and symptoms: cache group-type handling vs template compilation error." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "A framework version incompatibility is not the same as a multimodal dtype mismatch." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline return_language regression and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs issues, but they cover different pipeline/task deprecations and are not the same change." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export with torch_compilable_check is unrelated to Perceiver interpolation/resolution failures." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a dev0 regression on MoE models and the other is a specific weight-shape mismatch; not clearly the same underlying bug." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video input regression for Qwen VL models is unrelated to a generic mask-shape IndexError." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "A model producing bad outputs is different from ModelOutput key assignment behavior." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline regression and CamemBERT masked-LM prediction changes are different code paths and symptoms." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure and template compilation error are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "Perplexity quality on a model is not the same as gradient checkpointing being ineffective with LoRA." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch compatibility and Qwen3.5 embedding unpacking with TRL are different bugs." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT prediction regression and tokenizer loading failure for EMBEDDIA/est-roberta do not match." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "save/from_pretrained extra_state handling and AutoModelForCausalLM config unwrapping are separate serialization/config issues." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5Tokenizer method conflict and gguf_file loading failures are different tokenizer/model-loading bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed zero3/BertModel indexing and compiled flex_attention on newer torch versions are unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Both mention memory, but one is Qwen3 MoE training inefficiency and the other is mmap OOM avoidance on Strix Halo." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and Whisper return_language regression are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer loading for est-roberta and Mistral regex patching fail in different tokenizer code paths." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "CPU device placement bug in XLNet and CPU-offload incompatibility in Qwen3_5Moe are different model-specific issues." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Serialization extra_state handling and Qwen VL video-input regression are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and _is_hf_initialized flag breakage are separate regressions." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading cjvt/sleng-bert and Mistral regex patch crashes are unrelated tokenizer issues." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES add-new-model-like failure and compiled flex_attention failure are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "High perplexity on a model is not the same as a runtime special-token/post-processor update bug." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "A tokenizer loading failure for one model is not the same as a multimodal config-loading failure for another." + }, + { + "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 initialization; not the same underlying bug." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Multimodal dtype mismatch and template compilation error do not share a concrete code path." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Unsupported center argument in feature extraction and _torch_extract_fbank_features issues are related in area but not evidently the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 8, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:40444", + "issue:42915", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44062", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45507", + "issue:45584" + ], + "soft_pairs": [ + "issue:43883|issue:44488", + "issue:43643|issue:44530", + "issue:34567|issue:40444", + "issue:36032|issue:42915", + "issue:44811|issue:45584", + "issue:43644|issue:43646", + "issue:43541|issue:45084", + "issue:44451|issue:44987", + "issue:43452|issue:43482", + "issue:42915|issue:44530", + "issue:43317|issue:43482", + "issue:44849|issue:45325", + "issue:36032|issue:43645", + "issue:43883|issue:44991", + "issue:44987|issue:44991", + "issue:45084|issue:45507", + "issue:43404|issue:44466", + "issue:43262|issue:43377", + "issue:43646|issue:44534", + "issue:43994|issue:44610", + "issue:43645|issue:44534", + "issue:44403|issue:45081", + "issue:43322|issue:43540", + "issue:43377|issue:43819", + "issue:42915|issue:45084", + "issue:44451|issue:45081", + "issue:44265|issue:44610", + "issue:44220|issue:44683", + "issue:44987|issue:45081", + "issue:45216|issue:45357", + "issue:43632|issue:44756", + "issue:43404|issue:45127", + "issue:42915|issue:44291", + "issue:44062|issue:44987", + "issue:36032|issue:43742" + ], + "prepared_review_unit_hash": "7f02c409a302593a65e150ced39e40f3c9dfb3ab9353453310fdc1d4018d75b6", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 35, + "serialized_chars": 37696, + "estimated_input_tokens": 9424, + "estimated_eval_tokens": 19104 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly distinct regression/compatibility reports across different model families and APIs. A few are thematically related (especially around Transformers v5 model-loading and initialization), but none of the soft pairs look like the same underlying bug/change.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 43646 is the broadest representative of the cluster\u2019s recurring v5 initialization regressions, so it is the best single anchor if one issue must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 most clearly captures a central, reusable regression theme (custom model initialization breaking in v5) and is more general than the narrower model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading failures, but they involve different models and different symptoms (missing all_tied_weights_keys vs failing to load a specific BERT checkpoint)." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "These are unrelated failures: missing fields from trust_remote_code config loading versus a PagedAttentionCache/group-type crash in Qwen3.5." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Trainer token counting not updating and Qwen2.5-VL IterableDataset multi-image failure are different code paths and bug classes." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer class construction error and Qwen3Moe FP8 config failure are unrelated issues despite both being loading problems." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription generation after align_special_tokens are distinct processor bugs." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Related v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; not the same bug." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo tracing on Mixtral grouped_mm and template-node compilation errors are separate compilation failures." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading failures, but they affect different checkpoints and likely different root causes." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "gguf_file API breakage and Qwen2.5-GGUF loading failure overlap in area but are not the same concrete failure." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache linear_attention crash are different model/runtime bugs." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "gpu+cpu offload loading of dequantized models and Qwen2.5-GGUF v5 loading failure are different loading paths with different symptoms." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 output_hidden_states regression and Qwen2.5-VL rope-index temporal scaling bug are unrelated model logic issues." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method conflict and notebook custom model initialization breakage are separate regressions." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Both concern loading/serialization in v5, but one is an attribute error on a model class and the other is a tokenizer load failure for a specific checkpoint." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both mention v5 loading failures, but they are different artifacts and likely different breakpoints (model vs tokenizer, different checkpoints)." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation failure and GraniteMoEHybrid invalid method call are not the same bug or code path." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Weight-tying bug in Mistral3ForConditionalGeneration and inconsistent lm_head serialization are related only by weights, not the same failure." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate default and MIMI batched-vs-single padding-mask mismatch are unrelated processor/encoder bugs." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization breakage and non-persistent buffers filled with junk are both v5 regressions but distinct issues." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor/model size mismatch are different vision-model problems." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Notebook model initialization breakage and junk-filled non-persistent buffers are separate v5 regressions, not one bug." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise and Mistral regex patch crash are unrelated; one is UX/noise, the other is an AttributeError in tokenizer patching." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segmentation fault and Qwen3OmniMoe video-input ValueError are different multimodal loading/processing failures." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI padding-mask batching bug and DAC STE mismatch are both model correctness issues, but on different models and mechanisms." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and template-node compile failure are unrelated errors in different subsystems." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Loading a specific BERT model and crashing in fix_mistral_regex are different tokenizer-loading failures with different roots." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export export failure with torch_compilable_check and OmDet-Turbo image-size mismatch are unrelated bugs." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "_torch_extract_fbank_features and compiled flex_attention on torch>=2.9 are different low-level failures." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Model loading failure for a checkpoint and Mistral tokenizer regex patch crash are separate issues." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 save_pretrained regressions, but one concerns checkpoint correctness in general and the other incorrect visual-encoder keys specifically." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "_is_hf_initialized flag breakage and disabling mmap on Strix Halo to avoid OOM are unrelated bugs." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Weight-tying in Mistral3ForConditionalGeneration and LoRA merge collapse with extended vocab are different tied-weights problems with different triggers." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and init_empty_weights unexpected _is_hf_initialized argument are separate v5 compatibility issues." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken special keyword conflict and model loading failure for physical-intelligence/fast are different tokenizer/model load errors." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer method conflict and MobileLLM-125M key error are unrelated loading failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 9, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:44079", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812" + ], + "soft_pairs": [ + "issue:43296|issue:43531", + "issue:44479|issue:45325", + "issue:42915|issue:44589", + "issue:43611|issue:44534", + "issue:43531|issue:44877", + "issue:40444|issue:45325", + "issue:36032|issue:41720", + "issue:40990|issue:44811", + "issue:43425|issue:44403", + "issue:41553|issue:43582", + "issue:43632|issue:44534", + "issue:44079|issue:45663", + "issue:42175|issue:45542", + "issue:44462|issue:45812", + "issue:43421|issue:44568", + "issue:43792|issue:44589", + "issue:43818|issue:45357", + "issue:43317|issue:43856", + "issue:44403|issue:44991", + "issue:44683|issue:45020", + "issue:33666|issue:40444", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:35707|issue:40444", + "issue:44291|issue:44589", + "issue:44079|issue:44534", + "issue:43482|issue:43525", + "issue:43377|issue:45491", + "issue:43749|issue:44079", + "issue:43531|issue:44291", + "issue:41720|issue:45081", + "issue:43644|issue:44079" + ], + "prepared_review_unit_hash": "6ff7c7438dde47f96e0241cb4a3b1759b5f7ccc643ce45d2e56b21e49ba70db5", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37344, + "estimated_input_tokens": 9336, + "estimated_eval_tokens": 18928 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose pile of unrelated Transformers regressions that all happen to mention v5/v5.3 loading or model behavior. Only one soft pair looks like a true duplicate: the non-persistent-buffer corruption regression.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the clearest duplicate anchor: it states the exact non-persistent-buffer corruption bug, and issue:44534 is the same regression with a slightly reworded title.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best representative issue here because it matches the exact duplicate pair and is more specific than the other broad v5-regression reports in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different problems: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Related vision/VL area, but different concrete bugs: broad video-input regression vs still-image position_id scaling." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Different failure modes: FineGrainedFP8Config issue vs float8 storage lookup TypeError." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but they describe different bugs and code paths: base_model_prefix loading vs non-persistent buffer corruption." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Unrelated models and failures: sliding_window on Qwen3-MoE vs strict config blocking granite_speech." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both involve VL models, but one is multi-image IterableDataset training and the other is temporal position_id scaling." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer AttributeError vs device-mapping cudaErrorAssert; no shared underlying bug." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Different functionality: perplexity evaluation issue vs Whisper batch_decode skip_special_tokens bug." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Different issue types: torch/backend compatibility vs logging noise when loading a transformer." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "AutoTokenizer error messaging vs Apple Silicon warmup TypeError are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but they affect different internals: _is_hf_initialized flag vs non-persistent buffers." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "Different model families and bugs: ModelOutput key assignment vs Gemma-4 FSDP2 shared_kv_states reconstruction." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Packaging/backend installation issue vs tf backend undefined error; not the same bug." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Both tokenizer-related, but one is repo tokenizer.json resolution and the other is Granite token ID regression." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer runtime behaviors: post-processor update vs add_special_tokens not adding BOS/EOS." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure vs float8 storage object error; unrelated." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different Video-LLaVA/Qwen3.5 issues: missing temporal attention/shared weights vs wrong visual encoder keys on save_pretrained." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different loading/offload and training-memory problems; same broad area only." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Loading noise vs tokenizer loading failure for a specific model; not duplicates." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Torch >=2.9 compiled flex_attention failure is unrelated to remote_code model breakage." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both mention multi-image training, but they are different model-specific failures with different setups." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure vs CLIPTokenizer model_max_length bug; different issues." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both are training/classification-related, but one is flash_attention_3 degeneration and the other is single-label zero-loss behavior." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Different generation/training contexts and different bugs." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError vs float8 storage lookup error; different code paths." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment bug is unrelated to non-persistent buffer corruption." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Different models and failures: Qwen2.5-GGUF loading vs Llama4Config missing pad_token_id." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both involve batching, but one is MIMI padding-mask mismatch and the other is Gemma3 NaN embeddings with sliding-window attention." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP_CPU_RAM_EFFICIENT_LOADING breakage vs ModelOutput key assignment; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue vs init_empty_weights argument mismatch are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "cudaErrorAssert on auto device mapping vs Mistral regex patch crash; unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer corruption and ModelOutput key assignment 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": 2, + "component_count": 17, + "review_unit_index": 10, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43792", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43957", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44610", + "issue:44617", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45325", + "issue:45381", + "issue:45584", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45834" + ], + "soft_pairs": [ + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43632|issue:45072", + "issue:42175|issue:43296", + "issue:43611|issue:43646", + "issue:45325|issue:45381", + "issue:43994|issue:44898", + "issue:43454|issue:44466", + "issue:45325|issue:45834", + "issue:43792|issue:43957", + "issue:42915|issue:43482", + "issue:44387|issue:44617", + "issue:43296|issue:45800", + "issue:43881|issue:44488", + "issue:43701|issue:45325", + "issue:44479|issue:45823", + "issue:44291|issue:44877", + "issue:40990|issue:45584", + "issue:43262|issue:43746", + "issue:43606|issue:44265", + "issue:43873|issue:45701", + "issue:34634|issue:35707", + "issue:44186|issue:44265", + "issue:43317|issue:43873", + "issue:35707|issue:41720", + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:41720|issue:45823", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:43377|issue:44898", + "issue:39692|issue:40444", + "issue:43572|issue:45759" + ], + "prepared_review_unit_hash": "c67d1b240a71fbc1ae5845d05ae4eac54d8e0028d6cd1bb3982d85e41344111f", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37524, + "estimated_input_tokens": 9381, + "estimated_eval_tokens": 19018 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated bug reports across models and framework regressions; I only found one likely duplicate pair in the Qwen2.5-VL position-id/rope-index area.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45325 is the clearest root report for the Qwen2.5-VL position-id/rope-index regression; issue:45381 looks like the same underlying bug surfaced through video input.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45325 is the best anchor in this set because it names the concrete function/regression most directly and matches the accepted duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL-related, but one is a finetuning/IterableDataset multi-image failure and the other is a video position-id bug; different concrete problems." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export with torch_compilable_check is unrelated to bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized init behavior is a different regression from dtype mismatches during inference." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Missing TensorFlow in the torch extra and PaddleOCR-VL loading are unrelated install/load issues." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both mention Transformers v5 model loading/init breakage, but they describe different failures and code paths." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Both report the same Qwen2.5-VL position-id/rope-index bug, one via still-image temporal ids and the other via video vision_position_ids." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failures: SigLIP2 nonsensical outputs vs Perceiver resolution/interpolate_pos_encoding behavior." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "One is missing/tied lm_head weights in AyaVision; the other is inconsistent serialization of lm_head.weight by device." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL position-id handling and Kosmos2.5 long-OCR index errors are different model-specific bugs." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure is not the same as meta-device loading regressions in other models." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe and GGUF loading failure in Qwen2.5 are different issues." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Generic OOM under int4 quantization is not the same as SAM3Video running out of memory." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "PaddleOCR-VL/vLLM loading and torch/Transformers version incompatibility are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading failures with no shared concrete bug." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL rope-index calculation." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Video regression in Qwen models and Gemma4 PLE device mismatch are different code-path problems." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument during init and strict config rejection are separate regressions." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "High perplexity on gpt-oss and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch for Bark and torch.export failures are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Offloading/quantization problems and tokenization changes are too broad to be the same concrete bug." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer padding/truncation crashes are unrelated to torch.export/torch_compilable_check failures." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both mention offloading, but one is dequantized model loading with device_map=auto and the other is a broad quantization/offloading complaint." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation cache handling and Qwen3 auto device-map CUDA asserts are different failures." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Bark CPU-offload mismatch and DAC.from_latents STE mismatch do not share the same code-path problem." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "AyaVision lm_head tying and Bamba torch.compile/SDPA issues are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both involve device placement, but the model families and failure modes are different." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask divergence and OmDet-Turbo processor/model image-size mismatch are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "lm_head weight tying/serialization and bf16 dtype mismatches are different issues." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Different models and problems: padding-mask behavior vs Perceiver resolution handling." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 docs/example errors and Qwen2.5-VL multi-image finetuning failures are not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and text_config unwrapping for composite Qwen configs are different regression classes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pair rejections are conservative and the only accepted edge, issue:45325 \u2194 issue:45381, is plausibly the same Qwen2.5-VL positional-index regression by title and scope. I don't see an overstatement that would justify rejecting this review." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 11, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43749", + "issue:43792", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44008", + "issue:44186", + "issue:44387", + "issue:44423", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44918", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45812", + "issue:45823" + ], + "soft_pairs": [ + "issue:43526|issue:43883", + "issue:43749|issue:44534", + "issue:43881|issue:44991", + "issue:44466|issue:45357", + "issue:43296|issue:44877", + "issue:43454|issue:43720", + "issue:43792|issue:44683", + "issue:41720|issue:43322", + "issue:41720|issue:44977", + "issue:44008|issue:45823", + "issue:40990|issue:45701", + "issue:45632|issue:45684", + "issue:43404|issue:45357", + "issue:42617|issue:42915", + "issue:44186|issue:45072", + "issue:43577|issue:43708", + "issue:44387|issue:45704", + "issue:43873|issue:44756", + "issue:44423|issue:45406", + "issue:44877|issue:44991", + "issue:36331|issue:42617", + "issue:36331|issue:42915", + "issue:43450|issue:45701", + "issue:43550|issue:43720", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:44479|issue:44918", + "issue:42915|issue:43317", + "issue:41720|issue:42617", + "issue:43550|issue:43606", + "issue:45081|issue:45812", + "issue:43295|issue:45701" + ], + "prepared_review_unit_hash": "a8a75d2178fdbd3592ae5c485596274578ef9341d9bc696e1ef17638088fb7ba", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37484, + "estimated_input_tokens": 9371, + "estimated_eval_tokens": 18998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: trainer API regressions, tokenizer issues, model-loading/offload bugs, and unrelated model-specific failures. The soft pairs are mostly thematic overlaps rather than true duplicates, so no pair should be merged as the same underlying issue.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items cover distinct bugs across training, tokenization, loading/offload, and model-specific runtime paths.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the broadest open report with multiple references, but it still is not a safe canonical for the rest of this heterogeneous set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Different bugs: BeitImageProcessorFast label reduction vs Molmo missing all_tied_weights_keys." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Different mechanisms: broken FSDP CPU-efficient loading vs non-persistent buffer initialization junk." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different symptoms and code paths: GLM-4V load failure vs tokenizer loading regression for est-roberta." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve tied/serialized lm_head weights, but one is generic v5 serialization/device behavior and the other is a Qwen3.5 visual encoder save regression." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Model loading failures, but one is PaddleOCR-VL in vLLM and the other is strict config parsing for granite_speech." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Both mention weight handling, but AyaVision lm_head tying and BitNet packed-weight unpacking are separate bugs." + }, + { + "left": "issue:43792", + "right": "issue:44683", + "accept": false, + "reason": "Whisper model loading failure is unrelated to compiled flex_attention failing on torch >= 2.9." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Both are CUDA crashes, but one is Qwen3 auto device mapping and the other is Llava Next loading segfaults." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Different Qwen3-family failures: cudaErrorAssert during device mapping vs flash-attention generation bug." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma-family, but one is a variable-name collision in forward() and the other is a device mismatch in PLE during forward." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Both discuss tokenizer/tokenization, but one is perplexity quality on a specific model and the other is a broad tokenization-version change report." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Both concern save/load workflows, but cache-path collisions and read-only permission propagation are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Both touch weight tying in multimodal models, but Mistral3 lm_head tying and Qwen3.5 visual encoder save keys are distinct." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Both are runtime failures, but 3d_parallel.py execution and Qwen3Moe FineGrainedFP8Config errors are separate problems." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype/processing-related, but tokenizer crashes and dtype mismatches in CI are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "BLIP2 dtype propagation and Trainer resume/max_steps calculation are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Different memory issues: int4 quantization OOM in Transformers v5 vs apex FusedRMSNorm memory leak in T5." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both relate to offloading/memory, but quantization offload behavior and Strix Halo mmap OOM are not the same failure." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Different serve-path crashes: multimodal continuous batching vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config loading and tokenizer loading regressions are separate issues." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss API breakage is unrelated to the 3d_parallel.py execution failure." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Unexpected num_items_in_batch kwarg in Trainer is unrelated to Qwen3Moe FP8 loading/runtime failure." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug and broad tokenization-change report are not the same code-path." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Both involve model loading/compile paths, but torch.compile SDPA failure for Bamba and BitNet packed-weight unpacking are separate." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Processor output size mismatch for OmDet-Turbo is unrelated to Perceiver non-default-resolution failure." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "NER/batched-padding tokenizer crash and OmDet processor resolution mismatch are different processor bugs." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Video-input regression in v5.3.0 and unpacking Qwen3.5 embeddings with TRL SFT trainer are different failures." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8 failure and device_map=auto dequantized offload failure are separate loading paths." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "CUDA assert on Qwen3 auto mapping is unrelated to not being able to run 3d_parallel.py." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both mention offload/compile-ish behavior, but Bamba SDPA torch.compile failure and Bark CPU-offload device mismatch are not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer patch crash on Mistral and Granite token ID regression are separate tokenizer bugs." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor.tokenizer regression is a custom model API break, not the same as a general tokenization-change report." + } + ] + }, + "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": 17, + "review_unit_index": 12, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:43927", + "issue:44186", + "issue:44361", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44625", + "issue:44671", + "issue:44743", + "issue:44779", + "issue:44913", + "issue:45059", + "issue:45161", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45859" + ], + "soft_pairs": [ + "issue:43316|issue:44913", + "issue:43526|issue:43881", + "issue:44671|issue:44743", + "issue:41720|issue:45161", + "issue:43531|issue:43572", + "issue:43421|issue:45812", + "issue:36010|issue:42617", + "issue:43540|issue:43782", + "issue:44186|issue:44361", + "issue:42915|issue:43322", + "issue:43388|issue:45059", + "issue:45406|issue:45464", + "issue:43421|issue:45701", + "issue:30064|issue:34634", + "issue:33357|issue:35707", + "issue:44492|issue:44509", + "issue:44779|issue:45812", + "issue:41720|issue:43746", + "issue:43606|issue:43720", + "issue:43873|issue:45704", + "issue:43550|issue:43653", + "issue:44514|issue:44545", + "issue:44625|issue:45479", + "issue:43531|issue:45759", + "issue:43606|issue:43653", + "issue:34634|issue:41720", + "issue:43421|issue:43927", + "issue:34567|issue:35707", + "issue:45759|issue:45859" + ], + "prepared_review_unit_hash": "e228b4a92d1421e73aec5cbfaf7bc6e457f861209adf96852e8bbb918b667c8e", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35682, + "estimated_input_tokens": 8921, + "estimated_eval_tokens": 18098 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: config, tokenizer, generation, serving, and model-loading bugs across different model families. Only one clear duplicate pair stands out: the two Qwen2_5_VLProcessor apply_chat_template crash reports.", + "confidence": 0.95, + "canonical_issue_reason": "issue:44514 is the earlier and clearer report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input/padding=False crash as issue:44545.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44514 is the best canonical issue candidate because it cleanly represents the only confirmed duplicate pair and has the more stable, complete report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Different bugs: Gemma3TextConfig API mismatch vs GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated issues affecting different models/components: BeitImageProcessorFast label reduction vs glm-4v-9b load failure." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Different model-path failures: CamemBERT masked LM regression vs Qwen3.5 recurrent state reset in cache logic." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both involve Qwen/MoE, but the concrete failures differ: device-mapping cudaErrorAssert vs tensor-parallel not working." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Distinct config/runtime bugs in different models: Qwen3-MoE sliding_window vs StableLmConfig pad_token_idx." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer-related, but one is runtime post-processor syncing and the other is wrong Granite token IDs; not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "Completely different problems: GenerationMixin import error vs inability to run 3d_parallel.py." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different multimodal model failures: Qwen3OmniMoe video input ValueError vs Qwen3VL weight_only/from_pretrained error." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both tokenizer issues, but they affect different tokenizers and failure modes; not the same underlying bug." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Unrelated: Qwen3Moe FP8 config failure vs Llava Next segmentation fault on load." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different subsystems and symptoms: gather_for_metrics label truncation vs SAM3 PCS text/bbox behavior." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different serve/runtime failures: Gemma4Processor missing _tokenizer vs Qwen3.5-0.8B streaming chat/completions issue." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "The latter is a broad version-dependent tokenization regression, not the same specific special-token post-processor bug." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Both mention processors, but the affected processors and bugs are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "No concrete overlap: MacOS bus error with CLIP vs progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "Typo in cache strategies is unrelated to docs mentioning removed pipeline tasks." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different model families and different failure mechanisms." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different issues: Qwen3 auto device mapping crash vs GraniteSpeech PEFT local checkpoint loading failure." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offload paths, but the concrete bugs differ: CPU offload device mismatch vs packed BitNet weights not unpacked." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Completely different areas: torch.compile/SDPA failure vs tokenizer special-token registration." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False; these are duplicates." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Related to num_labels, but one is config propagation in Qwen3.5 and the other is a generic zero-loss training bug; not the same issue." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen3 variants, but one is sliding_window behavior and the other is config unwrapping in from_config." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "No shared underlying bug: CPU offload device mismatch vs BigBirdTokenizer mask-token registration." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "Different model families and failure modes; nothing indicates the same bug." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "One is runtime tokenizer post-processing; the other is config save/load losing custom token IDs." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "Training-state accounting bug vs generation behavior bug; unrelated." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is composite config handling and the other is CPU-offload incompatibility in rotary embedding; distinct fixes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the packet supports a mostly heterogeneous cluster with only one clearly confirmed duplicate pair (issue:44514 / issue:44545). The soft-edge rejections are conservative and stay within the evidence from titles/bodies. Minor caution: phrases like \"more stable, complete report\" are not directly evidenced in the packet, but they do not materially affect the clustering judgment." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 13, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42886", + "issue:42915", + "issue:43381", + "issue:43421", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43746", + "issue:43782", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45588", + "issue:45704", + "issue:45759", + "issue:45859" + ], + "soft_pairs": [ + "issue:44451|issue:44991", + "issue:44368|issue:45357", + "issue:44898|issue:45072", + "issue:43782|issue:45381", + "issue:43531|issue:44589", + "issue:44077|issue:44625", + "issue:44987|issue:45042", + "issue:44062|issue:44451", + "issue:43421|issue:43653", + "issue:44918|issue:44936", + "issue:45588|issue:45759", + "issue:39692|issue:44610", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:43421|issue:44186", + "issue:43994|issue:44671", + "issue:43525|issue:43531", + "issue:45704|issue:45859", + "issue:45081|issue:45356", + "issue:41720|issue:45704", + "issue:43596|issue:44220", + "issue:39692|issue:43873", + "issue:43746|issue:43873", + "issue:42915|issue:45759", + "issue:40444|issue:43746", + "issue:43746|issue:45381", + "issue:44220|issue:44661", + "issue:43540|issue:44008", + "issue:30064|issue:36010" + ], + "prepared_review_unit_hash": "d307e2304c3faaef1c71c358b8dcb7b6bc21a303703bc1dc293f01510f8e4c94", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36092, + "estimated_input_tokens": 9023, + "estimated_eval_tokens": 18302 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated bugs across different models, tokenizers, quantization/offload paths, and docs/examples. None of the soft pairs look like the same underlying fix, so I would not merge any of them as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No clear canonical issue: the items span many distinct code paths and model families, so there is no single issue that cleanly represents the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a strong global representative; even the closest-looking pairs differ by model, failure mode, or subsystem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model loading problems, but they affect different models and failure modes; not the same bug." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a tie_word_embeddings warning during LoRA fine-tuning; the other is incorrect visual-encoder keys being saved. Different issues." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Both involve vision/inference, but one is a positional-encoding image-size bug and the other is a bfloat16 dtype mismatch." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen variants and different failures: weight_only loading error vs wrong video vision_position_ids." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "One is about optional post_init in patchtsmixer; the other is num_labels propagation in Qwen3.5 configs." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Different subsystems: loading a specific model on 5.1.0 vs PIL image processors incorrectly requiring torchvision." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken keyword collision is unrelated to the model-loading failure in EMBEDDIA/est-roberta." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both involve special tokens, but one is runtime post-processor updates and the other is tokenizer registration/empty decode behavior." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "TRL embedding unpacking after SFT is unrelated to trainer.evaluate() failing after train()." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Different code paths: flash-attention sink handling vs from_config not unwrapping text_config for composite Qwen models." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both are vision-related, but they concern different models and different mismatches (docs/quantization vs processor input size)." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing in eval mode and ZeRO-3 gradient explosion are separate training bugs." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both touch offline hub behavior, but one is cache loading and the other is an unconditional model_info() call in mistral regex patching." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Special-token post-processor updates are not the same as LayoutLMv2 tokenizer crashes on NER padding/truncation." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and symptoms: SigLIP2 bad outputs vs CamemBERT masked-LM prediction regression." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "T5 apex memory leak and Qwen3_5Moe CPU offload incompatibility are different problems." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both concern tokenizer patching/regression, but one is a crash in _patch_mistral_regex and the other is Kimi-K2.5 codec handling." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 auto device mapping cuda asserts and T5 apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed ZeRO-3/BertModel indexing error is unrelated to the fbank feature extraction issue." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 docs/quantization issues are not the same as quantized offloading not working as expected." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT local checkpoint loading for GraniteSpeech is a different bug from quantization/offloading behavior." + }, + { + "left": "issue:42915", + "right": "issue:45759", + "accept": false, + "reason": "Both are Qwen config-related, but one is FP8Config failure and the other is from_config handling of composite multimodal configs." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Multiple-image Qwen2.5-VL finetuning failure and GraniteSpeech adapter loading are unrelated." + }, + { + "left": "issue:43746", + "right": "issue:45381", + "accept": false, + "reason": "Different models and paths: PEFT adapter loading vs Qwen2.5-VL video position IDs." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank feature extraction is unrelated to add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Qwen3OmniMoe video-input processing and Gemma3n forward variable collision are different bugs." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Void segmentation map processing and GenerationMixin import failure are completely 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": 17, + "review_unit_index": 14, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:39692", + "issue:43295", + "issue:43322", + "issue:43335", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43825", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44448", + "issue:44485", + "issue:44488", + "issue:44509", + "issue:44561", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45685", + "issue:45834" + ], + "soft_pairs": [ + "issue:43720|issue:43819", + "issue:43618|issue:43697", + "issue:43335|issue:43756", + "issue:45183|issue:45479", + "issue:43653|issue:43720", + "issue:43716|issue:44008", + "issue:43295|issue:45834", + "issue:44062|issue:44488", + "issue:43550|issue:45072", + "issue:43550|issue:43819", + "issue:43322|issue:45685", + "issue:44186|issue:44898", + "issue:44843|issue:45081", + "issue:43531|issue:43643", + "issue:43577|issue:43818", + "issue:43643|issue:44877", + "issue:43421|issue:44811", + "issue:43653|issue:43819", + "issue:44448|issue:44671", + "issue:43540|issue:45381", + "issue:43782|issue:44977", + "issue:45584|issue:45834", + "issue:43644|issue:43749", + "issue:44485|issue:44509", + "issue:44821|issue:45084", + "issue:43825|issue:43994", + "issue:44561|issue:45020", + "issue:39692|issue:45834" + ], + "prepared_review_unit_hash": "88def731470adc5aff4713ca9bd0170e64b8d24b6f48bd0cb74ec0b8505e4673", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36342, + "estimated_input_tokens": 9086, + "estimated_eval_tokens": 18428 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is broad and noisy: most items are distinct Transformers regressions across unrelated models, tokenizers, multimodal loaders, and docs. I do not see strong duplicate evidence for most soft edges; the closest ones share a subsystem name or helper function, not the same concrete bug.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45020 is the broadest umbrella here and best matches the recurring theme of recent-version regressions breaking model loading/usage, especially around remote-code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most representative issue for the cluster\u2019s general regression theme; it is broader and more central than the highly specific model- or tokenizer-only reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different models and failures: packed-weight loading vs DAC latent reconstruction/STE mismatch." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both are v5 behavior regressions, but they affect different models and code paths." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Unrelated architecture bugs: sparse-layer config creation vs RoPE layer dropping." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Different validation/loss bugs in separate code paths; not the same underlying issue." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug vs packed-weight loading bug; unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Different multimodal model failures with different root causes." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal input issues, but one is processor/tokenizer regression and the other is OCR index handling." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer problems: AddedToken construction vs loading a specific model/tokenizer." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs in different models: torch.compile/SDPA vs dtype mismatch under bfloat16." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "No shared underlying code path; compile/SDPA failure is unrelated to DAC latents." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Both are runtime errors, but in unrelated models and hardware backends." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer/padding crash vs image classification resolution/interpolation failure; unrelated." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper name, but different failures: offline hub call vs missing backend_tokenizer attribute." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window MoE bug is unrelated to remote-code config field loss." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype propagation bug vs Video-LLaVA weight/temporal-attention issue." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both involve loading configs/models, but the concrete failures differ." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; not the same issue." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent decoding are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v5 output regressions, but on different model families and causes." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both mention video/Qwen, but they target different model variants and different input bugs." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different Qwen issues: from_pretrained weight-only loading vs flash-attention generation behavior." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation bug is unrelated to Kosmos2.5 long-OCR indexing." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer initialization bug vs FSDP CPU RAM-efficient loading regression." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "RoPE implementation discussion vs outdated docs; no shared bug." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "ImageProcessor URL loading and template-node compilation are unrelated." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Error-message/docs issue vs SigLIP2 wrong results from AutoModel/pipeline mismatch; not the same bug." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Related broad area (remote-code regressions), but these are distinct breakages with different triggers." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "Different model-specific multimodal/documentation problems with no shared failure mode." + } + ] + }, + "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": 17, + "review_unit_index": 15, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:43262", + "issue:43329", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43650", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:44060", + "issue:44186", + "issue:44291", + "issue:44368", + "issue:44442", + "issue:44479", + "issue:44554", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44936", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45406", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45701", + "issue:45803", + "issue:45834" + ], + "soft_pairs": [ + "issue:30064|issue:36331", + "issue:41720|issue:44991", + "issue:43825|issue:43827", + "issue:44479|issue:45834", + "issue:44186|issue:44442", + "issue:43577|issue:43582", + "issue:44805|issue:44936", + "issue:44671|issue:45701", + "issue:43606|issue:45072", + "issue:34634|issue:36010", + "issue:44610|issue:45072", + "issue:43782|issue:45325", + "issue:42175|issue:45081", + "issue:43720|issue:45072", + "issue:45081|issue:45406", + "issue:43749|issue:44291", + "issue:44060|issue:44368", + "issue:43582|issue:44857", + "issue:43650|issue:43906", + "issue:43262|issue:43819", + "issue:43526|issue:43697", + "issue:43526|issue:44554", + "issue:45084|issue:45538", + "issue:45459|issue:45803", + "issue:34634|issue:36331", + "issue:43329|issue:45600", + "issue:43742|issue:45020" + ], + "prepared_review_unit_hash": "12bef7c90d365f5e98ee5cfd51b697f82a88dbe9d609ba49bdbcb2a5137aa5c5", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35510, + "estimated_input_tokens": 8878, + "estimated_eval_tokens": 18012 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft pairs share broad themes like model loading, tokenizers, or v5 regressions, but not the same concrete bug or change. I would not merge any of the listed soft edges.", + "confidence": 0.89, + "canonical_issue_reason": "No single issue is a true canonical duplicate target for this mixed set; if one must be chosen, issue 45020 is the broadest umbrella regression report, but it is not actually the same bug as most others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the most general, user-facing open regression and the closest thing to an umbrella report in this set, but it should still not absorb the unrelated issues as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Different subsystems and failures: image segmentation processing vs trainer compute_loss signature mismatch." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated problems: CUDA assert during device mapping vs tokenizer load failure for a specific model." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both concern v5 pipeline removal, but one is an error message bug and the other is stale docs; not the same underlying defect." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Different models and code paths: Qwen video regression vs Kosmos2.5 OCR long-input index error." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer bugs affecting different models and inputs." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Model dtype loading bug vs Apple Silicon allocator TypeError; unrelated." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask shape IndexError in one workflow vs evaluate-after-train failure; not the same bug." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT prediction regression is not the same as a broad tokenization-change report." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Different inference issues: CPU offload device mismatch vs bf16 dtype mismatch." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "Bark voice preset behavior and GenerationMixin import error are unrelated." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch vs dtype mismatch; different code paths." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Both are VL-related, but one is weight_only loading and the other is rope/position-id scaling; distinct bugs." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Packaging/backend dependency issue vs tokenizer regex patch crash; unrelated." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight loading failure is a different issue from generic dtype mismatches." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Similar symptom class (missing attribute), but different processors and different failure sites." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both involve loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights arg compatibility." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Different warnings and model families; not the same tied-weights problem." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon warmup TypeError vs CUDA AMP loss crash; unrelated." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "One is a meaningless report and the other is a reproduction of a specific upstream issue; no duplicate relation." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio sampling-rate default bug and DAC latent consistency bug are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Beit reduce_labels bug and RTDetrV2 output regression are different model-specific defects." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Completely different topics: label reduction vs MPS attention correctness." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation error vs tokenizer max_length bug; unrelated." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they are different components and different swallowed-error paths." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor preset issue and trainer compute_loss signature issue are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Multimodal token counting bug vs stale auto_mappings references; different files and problems." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A key error for one model is not the same concrete bug as a broad remote_code compatibility regression." + } + ] + }, + "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": 17, + "review_unit_index": 16, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43454", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43638", + "issue:43643", + "issue:43645", + "issue:43653", + "issue:43696", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:43931", + "issue:44479", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45042", + "issue:45356", + "issue:45488", + "issue:45701", + "issue:45759", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:43653|issue:44912", + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:44589|issue:45042", + "issue:43746|issue:44811", + "issue:43854|issue:44977", + "issue:43257|issue:43696", + "issue:38175|issue:44610", + "issue:36010|issue:43645", + "issue:39692|issue:44912", + "issue:43819|issue:43927", + "issue:43454|issue:43577", + "issue:43720|issue:44610", + "issue:43550|issue:44912", + "issue:44568|issue:45488", + "issue:43531|issue:45042", + "issue:43262|issue:44811", + "issue:44568|issue:45356", + "issue:43329|issue:44479", + "issue:36010|issue:36331", + "issue:43749|issue:44756", + "issue:43643|issue:45759", + "issue:44849|issue:45701", + "issue:43257|issue:43931", + "issue:43329|issue:43450", + "issue:34567|issue:40990", + "issue:43606|issue:45859", + "issue:43257|issue:43756", + "issue:45823|issue:45859" + ], + "prepared_review_unit_hash": "cd1b29853f7b678ac1af1421b6785f381e8db2826db35f41db53272d0af337a3", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36389, + "estimated_input_tokens": 9098, + "estimated_eval_tokens": 18452 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated closed issues spanning tokenizer regressions, loading/offload bugs, multimodal processor problems, and trainer behavior changes. The only plausible duplicate-style match is the Qwen VLM video-input regression pair.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45701 is the broadest open, user-facing umbrella here: it directly describes version-dependent tokenization changes and has inbound references, making it the best representative for the tokenizer-regression slice of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the best single issue to keep as a representative anchor because it is open, broadly worded, and closest to the recurring tokenizer-regression theme seen among several items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Special-token decode failure vs MXFP4 quantization/loading fallback; different subsystems and symptoms." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3/Bert indexing failure vs post-train evaluate failure; not the same code path." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Generic gguf/from_pretrained breakage vs a specific tokenizer regression for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading issue vs multimodal video token counting bug; different defects." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup error vs image processor torchvision dependency regression." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading bug vs Whisper batch_decode skip_special_tokens bug." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash loading test failure vs Qwen3.5 flash-attention generation bug." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Loading/conversion bug vs out-of-memory report; unrelated failure modes." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 probability issue vs OmDet processor size mismatch; different model behavior." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "GenerationMixin import error vs Jupyter custom-model initialization regression." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Docs example errors for SigLIP2 vs GPT-oss quantization loading; no shared bug." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC forward/from_latents mismatch vs config save/load custom token ID loss." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Weight-tying/text generation bug vs dtype propagation bug during Blip2 load." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Accelerate packed-weight unpacking bug vs processor image-shape mismatch." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA failure vs quantization fallback; different concrete problems." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both tokenizer-related, but one is add_special_tokens/BOS-EOS handling and the other is pre-tokenizer override; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Sliding-window MoE bug vs PIL backend torchvision dependency issue." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode special-token skipping." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regressions with different root causes and model families." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both point to the same video-input multimodal path; 43329 names the broken video token-counting branch, and 44479 reports a broad video regression on the same VLM families." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "Import-time GenerationMixin breakage vs compute_loss signature change; unrelated API regressions." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU-RAM-efficient loading bug vs mmap/OOM issue; different loading paths and causes." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Missing fields from trust_remote_code AutoConfig vs AutoModelForCausalLM.from_config text_config unwrap issue." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs broad tokenization-change report; not the same underlying defect." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE loading under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Video token counting undefined-symbol bug vs batched video processor shape bug." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Trainer token-count tracking bug vs GPT-oss perplexity being too high; different layers of the stack." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both involve CPU offload/device placement, but they affect different models and code paths." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "MoE loading/conversion bug vs Smollm3 RoPE-layer discrepancy; unrelated model behavior." + }, + { + "left": "issue:45823", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 device mismatch during forward vs Qwen3_5MoE rotary embedding CPU-offload incompatibility." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is broadly grounded in the packet, and the soft-edge decisions are conservative overall. The lone accepted pair (issue:43329 / issue:44479) is reasonably mergeable as a single video-input regression fix. No clear overstatement of evidence stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 17, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43317", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43824", + "issue:43906", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44737", + "issue:44821", + "issue:44898", + "issue:44912", + "issue:44938", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45137", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45538", + "issue:45588", + "issue:45823" + ], + "soft_pairs": [ + "issue:43697|issue:45357", + "issue:44987|issue:45538", + "issue:44291|issue:45042", + "issue:43452|issue:43582", + "issue:42915|issue:43742", + "issue:42915|issue:44821", + "issue:33357|issue:34567", + "issue:42915|issue:45538", + "issue:41720|issue:43540", + "issue:44479|issue:45381", + "issue:43317|issue:45538", + "issue:43927|issue:44912", + "issue:42175|issue:43824", + "issue:43577|issue:45072", + "issue:43540|issue:44560", + "issue:43540|issue:44479", + "issue:44062|issue:44991", + "issue:42175|issue:44938", + "issue:45137|issue:45405", + "issue:42175|issue:44488", + "issue:38175|issue:44898", + "issue:44737|issue:45588", + "issue:28282|issue:34689", + "issue:43742|issue:44291", + "issue:43906|issue:45381", + "issue:42915|issue:43749", + "issue:43756|issue:45325", + "issue:43526|issue:45479", + "issue:28282|issue:29127", + "issue:43540|issue:45823", + "issue:43582|issue:43782", + "issue:43493|issue:43994", + "issue:44291|issue:44821" + ], + "prepared_review_unit_hash": "5dc0fadce238ac3a62d405c3f7c7c1eb6a5b261e13a5d3d08c8b2c8debd0aabe", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37719, + "estimated_input_tokens": 9430, + "estimated_eval_tokens": 19116 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous, with a few small duplicate-like mini-clusters around Qwen2.5-VL video/position-id regressions, tokenizer loading, and SigLIP2 behavior.", + "confidence": 0.71, + "canonical_issue_reason": "issue:44479 is the closest thing to a canonical anchor because it matches the Qwen2.5-VL video/position-id regression cluster; the rest of the items are too diverse to form one true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the strongest local hub among the issues, but it is still only suitable as a partial anchor rather than a global representative of the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different bugs: RTDetrV2 output drift vs Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different failure modes: tokenizer load regression vs CLIPTokenizer max_length misuse." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "One is init_empty_weights argument handling; the other is PIL backend unexpectedly requiring torchvision." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Unrelated issues: gguf_file loading breakage vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different model-specific load bugs: Qwen3Moe FP8 config vs MobileLLM key error." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Different bugs and code paths; no shared concrete failure is evident." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error on CLIP model is unrelated to TrainerState token counting." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Qwen3Moe FP8 config issue is unrelated to CLIPTokenizer model_max_length." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Both involve Qwen models, but the concrete failures differ: auto device mapping assert vs video-input ValueError." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both describe the same Qwen2.5-VL video/position-id regression in 5.3.0; one is a broader report and the other a narrower repro." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map offload failure is unrelated to CLIPTokenizer max_length." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Different topics: DiaConfig save/load token IDs vs MXFP4 quantization fallback." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend availability issue is unrelated to missing Qwen2_5_VL import." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but they affect different model classes and code paths." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input bugs, but they are different model families and different exceptions." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Different Qwen model families and different concrete failures; not the same bug." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": true, + "reason": "Likely the same tokenizer bug: the AddedToken/special-argument TypeError appears to be the concrete cause of the EMBEDDIA/est-roberta tokenizer load failure." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Backend packaging issue vs Python 3.14 load failure are unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT minimum-version bump." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Missing tensorflow backend in installation is unrelated to a model-loading failure." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different models and symptoms: SigLIP2 zero probabilities vs Perceiver non-default-resolution failure." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU arange bug is unrelated to flash-attention s_aux=None handling." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Generic PyTorch import error vs Llama 3.2 Vision model-loading regression are different issues." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM key error and init_empty_weights unexpected-arg failure are separate loading problems." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "A reproduction of another issue is not the same as the Qwen2.5-VL video-position-id bug." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Smollm3 RoPE-layer behavior and Qwen2.5-VL temporal position-id scaling are different model-specific bugs." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT fast processor label reduction bug is unrelated to zero-loss classification behavior." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch import availability and LayoutLMv3 box-clarity messaging are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45823", + "accept": false, + "reason": "Qwen3OmniMoe video input failure is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and Qwen3VL weight_only load error are different bugs." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Same underlying SigLIP2 implementation problem: one compares HF to JAX, the other reports the same model giving nonsensical outputs in HF usage." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights argument handling and AutoImageProcessor URL loading are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment stays grounded in the packet and the soft-edge calls are mostly conservative. The accepted duplicate-like pairs are reasonably supported by the titles, and the rejected pairs do not overclaim similarity. The summary also appropriately treats 44479 as a local anchor rather than a global canonical issue." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 18, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:24643", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43819", + "issue:44062", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44488", + "issue:44514", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44792", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:45005", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45290", + "issue:45356", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45600", + "issue:45823" + ], + "soft_pairs": [ + "issue:34567|issue:36331", + "issue:44062|issue:45081", + "issue:43643|issue:45375", + "issue:43452|issue:44488", + "issue:44792|issue:44964", + "issue:45042|issue:45081", + "issue:43742|issue:44589", + "issue:43531|issue:44821", + "issue:45405|issue:45600", + "issue:43688|issue:43749", + "issue:43697|issue:45216", + "issue:44484|issue:44704", + "issue:44514|issue:45381", + "issue:43720|issue:44898", + "issue:43643|issue:44561", + "issue:41762|issue:44936", + "issue:43452|issue:43577", + "issue:44625|issue:44877", + "issue:24643|issue:29127", + "issue:43452|issue:44843", + "issue:43295|issue:45356", + "issue:44466|issue:45005", + "issue:43819|issue:44811", + "issue:44514|issue:45290", + "issue:43653|issue:44568", + "issue:44936|issue:45823", + "issue:44464|issue:44743" + ], + "prepared_review_unit_hash": "0ae36d680b47f599b6150580f72028d554fda88aa5602b06bebc624ebbdacdc0", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36118, + "estimated_input_tokens": 9030, + "estimated_eval_tokens": 18316 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated transformer bugs. The only soft edge that looks like a true duplicate is the pair about v5 tied-weights/save_pretrained serialization behavior.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44466 is the better canonical issue: it states the underlying tied-weights serialization regression more generally and describes the device-dependent behavior directly; issue:45005 reads like a model-specific manifestation of the same v5 tied-weights problem.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44466 is the best representative of the cluster's only plausible duplicate thread because it captures the core regression rather than a downstream translation-model symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Different trainer problems: token counting state vs a custom compute_loss signature mismatch." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Both tokenizer-related, but one is an AddedToken kwarg error and the other is a Mistral regex/backend_tokenizer crash." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config fields, but the first is missing fields from trust_remote_code AutoConfig output and the second is a specific Qwen3.5MoeVisionConfig field dropped by strict parsing." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading failures, but they involve different APIs and likely different root causes." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both are multimodal model loading/test failures, but they concern different models and symptoms." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated subsystems: image processor torchvision dependency vs tokenizer regex crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Different load-time failures with different low-level causes; one is a model key error, the other is a missing float8 storage object." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version pinning and auto_mappings referencing removed configs are separate CI/release issues." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization and FSDP CPU RAM efficient loading are different bugs in different code paths." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "One is RTDetrV2 output drift in v5; the other is Qwen3.5 checkpoint saving. Same general version area, different bug." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "Save shard size question is unrelated to AutoProcessor kwargs forwarding." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL issues, but one is chat-template batching crash and the other is incorrect vision_position_ids for video input." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver image-classification resolution handling are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Different regressions: missing config fields under trust_remote_code vs removal of is_torch_fx_available breaking remote-code models." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 loading crash and trainer.evaluate-after-train failure do not share a concrete code path." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "GGUF loading regression and BLIP2 dtype propagation are distinct model-loading issues." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config handling, but one is num_labels propagation and the other is strict config blocking granite_speech loading." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "Training DeepSpeed weight shape error is unrelated to a LayoutLMv3 box-validation message issue." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "GGUF loading breakage and mistral_regex offline/model_info behavior are different tokenizer/loading problems." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions around processor/tokenizer behavior, but one is processor.tokenizer/image handling and the other is Kimi-K2.5 codec/fix_mistral_regex warnings." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both point to the same v5 tied-weights serialization regression: inconsistent lm_head.weight handling for tied-weight models, with translation models as a specific affected case." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both affect apply_chat_template in Qwen2.5-VL, but the failure modes differ: batched padding=False vs assistant tool-call messages without content." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer special-token regressions for different models and different APIs." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer evaluation failure and Gemma4 PLE device mismatch are not the same bug." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both concern generation/runtime correctness, but one is chunked generation with compiled forward and the other is recurrent state reset in a specific Qwen modular file." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the single accepted soft edge (44466/45005) is a conservative, well-supported duplicate call based on the issue titles. The other pair verdicts are appropriately cautious given the limited evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 19, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28282", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:43122", + "issue:43278", + "issue:43295", + "issue:43329", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44451", + "issue:44464", + "issue:44479", + "issue:44554", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44849", + "issue:44977", + "issue:44991", + "issue:45325", + "issue:45356", + "issue:45357", + "issue:45440", + "issue:45588", + "issue:45701", + "issue:45704" + ], + "soft_pairs": [ + "issue:43688|issue:45440", + "issue:30064|issue:33666", + "issue:38175|issue:43295", + "issue:44361|issue:44442", + "issue:43653|issue:44361", + "issue:43688|issue:45704", + "issue:28282|issue:32090", + "issue:43577|issue:43697", + "issue:28282|issue:33357", + "issue:41720|issue:43906", + "issue:28282|issue:35141", + "issue:43646|issue:43749", + "issue:43688|issue:45701", + "issue:44843|issue:44991", + "issue:44977|issue:45356", + "issue:43531|issue:43742", + "issue:43653|issue:43927", + "issue:44464|issue:44977", + "issue:43329|issue:45325", + "issue:43577|issue:45588", + "issue:43550|issue:43927", + "issue:43577|issue:44554", + "issue:44479|issue:44625", + "issue:43122|issue:43906", + "issue:43550|issue:44361", + "issue:32090|issue:34567", + "issue:41720|issue:44464", + "issue:43504|issue:43550", + "issue:43504|issue:43720", + "issue:32090|issue:33666", + "issue:43504|issue:44610", + "issue:43452|issue:44451", + "issue:43742|issue:44821", + "issue:44849|issue:45357", + "issue:43278|issue:43688" + ], + "prepared_review_unit_hash": "630e90614d011610badb22f0bc629ce280caeb875a8d45dd4c5a844bdbdfe2ce", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 35, + "serialized_chars": 37686, + "estimated_input_tokens": 9422, + "estimated_eval_tokens": 19100 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly connect unrelated bugs/regressions across tokenization, multimodal processing, training, loading, and dtype/device-specific behavior. I would keep all of the proposed soft edges separated.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items are only superficially similar and do not share one underlying bug or change.", + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE-related, but one is auxiliary-loss normalization and the other is a broader implementation divergence; not the same bug." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Image-processor segmentation bug vs multi-GPU training topic; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different failure modes: zero probabilities in SigLIP2 vs processor/tokenizer API regression." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer classes and errors; no shared code-path bug." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token decode issue vs MLukeTokenizer AttributeError; unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Both model-internal issues, but different components and failure mechanisms; not mergeable as one fix." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "ImportError for missing PyTorch vs trainer broadcast NoneType error; unrelated." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Different model families and different dtype/output issues." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Import-time dependency error vs MacOS bus error; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "No evidence they reproduce the same concrete crash path; only loose similarity around model loading/runtime failures." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "PyTorch missing dependency vs embedding reinitialization bug; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization regression vs FSDP CPU RAM efficient loading breakage; different code paths." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Generic tokenization/version-change report vs MoE auxiliary-loss normalization; not the same defect." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline hub metadata call bug vs tokenizer load failure for a specific model; different causes." + }, + { + "left": "issue:44977", + "right": "issue:45356", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug vs Kimi tokenizer/codec regression; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug vs MobileLLM loading key error; different issues." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration bug vs config token-ID persistence bug; not the same underlying problem." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency vs flash-attention generation bug; distinct failures." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Different VLM video/rope bugs; both multimodal, but not the same code-path problem." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Different model/dtype bug vs flash-attention crash on s_aux=None; unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "torch.compile SDPA failure in Bamba vs tokenizer AttributeError; unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype loading issue vs MPS attention correctness issue; different subsystems." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression in Qwen VL models vs num_labels propagation bug; unrelated." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Tokenizer version-difference report vs isolated reproduction of another issue; not enough evidence of the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Model compile/SDPA crash vs tokenizer task AttributeError; unrelated." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast NoneType error vs TrainerState counter not updating; different trainer behaviors." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Device-mapping cuda assert vs compiled chunked-generation inconsistency; different code paths." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit pretrained-load legacy field bug vs Bamba compile/SDPA failure; unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy field preset loading issue vs packed-weight accelerate loading issue; different models and fixes." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer broadcast error vs Qwen2-VL multi-GPU training request; unrelated." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Different processor/model mismatch problems; not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "GGUF loading breakage vs inability to load a specific tokenizer; related only at a high level, not the same defect." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM key error vs AutoImageProcessor-from-URL failure; unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5 output_hidden_states bug vs save_pretrained visual-key regression; distinct issues." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift in eval vs MoE auxiliary-loss normalization; 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": 17, + "review_unit_index": 20, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43720", + "issue:43756", + "issue:43824", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44466", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44734", + "issue:44843", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45759", + "issue:45812", + "issue:45823" + ], + "soft_pairs": [ + "issue:44568|issue:45812", + "issue:33666|issue:36010", + "issue:43493|issue:43825", + "issue:43756|issue:44913", + "issue:43262|issue:43295", + "issue:43525|issue:43572", + "issue:43606|issue:43927", + "issue:43720|issue:43927", + "issue:28218|issue:33357", + "issue:33666|issue:36331", + "issue:39692|issue:43493", + "issue:43720|issue:44361", + "issue:43493|issue:43844", + "issue:43278|issue:43844", + "issue:44964|issue:45405", + "issue:43540|issue:44977", + "issue:43334|issue:43572", + "issue:43257|issue:43299", + "issue:44625|issue:45584", + "issue:43756|issue:45083", + "issue:44488|issue:44843", + "issue:43334|issue:45759", + "issue:45200|issue:45229", + "issue:43646|issue:44466", + "issue:43824|issue:44589", + "issue:44589|issue:45520", + "issue:39692|issue:43295", + "issue:45229|issue:45823", + "issue:33357|issue:33666", + "issue:43572|issue:44291", + "issue:43504|issue:43606", + "issue:43927|issue:44361", + "issue:44423|issue:44734", + "issue:43575|issue:45405" + ], + "prepared_review_unit_hash": "f6a5d4ae2a8b3b8bd5e9059c973e6013f160924cb383e52e6648af87a771b309", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38205, + "estimated_input_tokens": 9552, + "estimated_eval_tokens": 19360 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous, and none of the soft pairs look like the same underlying bug or change. Most are either different model families, different failure modes, or one is documentation while the other is a runtime bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43824 is the most central issue in the set by discussion/inbound references and is a clear user-facing regression.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 has the strongest signal of impact and cross-linking, so it is the best representative issue among these unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer-related, but one is missing BOS/EOS behavior for mdeberta while the other is wrong token IDs for Granite models; different bugs and model families." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and a GenerationMixin import failure are unrelated code paths." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation mismatch vs an incorrect pipeline translation-task message; not the same issue." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Both mention rotary/rope settings, but one is a Smollm3 layer-selection bug and the other is GPTNeoX config reload regression." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio sampling-rate defaulting and processor.tokenizer/image handling are different regressions." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing pad-token fields, but for different configs and different attribute names; not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch in bark-small is unrelated to DiaConfig token-ID loss on save/load." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a different failure than DiaConfig losing custom token IDs." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing bug vs MacOS bus error loading a CLIP model; completely different problems." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and a custom trainer loss-signature TypeError are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "One is documentation example errors, the other is an implementation discrepancy for SigLIP2; not the same concrete bug." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 parity discrepancy vs a ZeRO-3 gradient inflation bug; different code paths and symptoms." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype drift in eval is not the same as abnormal gradient growth with DeepSpeed ZeRO-3." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal loading failure vs an unmet PEFT version requirement; different root causes." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError and Qwen3.5 flash-attention generation issues are different bugs." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-VL pad_token_id missing and StableLmConfig pad_token_idx missing are separate model/config regressions." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed differs from Qwen3VL MoE loading breakage in dev; not mergeable as one fix." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:45083", + "accept": false, + "reason": "Smollm3 RoPE layer dropping and a qwen3_omni_moe helper length bug are not the same defect." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model loading failure for sleng-bert and an offline-mode model_info call bug in AutoTokenizer are different issues." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both touch Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is incorrect unwrapping of text_config in from_config." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only finetuning defaults vs multi-GPU inference OOM are different Gemma4 problems." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage and device-dependent lm_head.weight serialization are distinct v5 regressions." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Qwen2_5_VL import failure and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "A Float8 storage TypeError and a flash_attn import-utils KeyError on Python 3.13 are different failures." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Documentation example mistakes are not the same as the processor.tokenizer regression in v4.57.5." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 OOM during multi-GPU inference and a device mismatch with device_map='auto' are separate runtime problems." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error on CLIP loading and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and an unexpected _is_hf_initialized arg under init_empty_weights are different compatibility issues." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy-field model loading in Beit is unrelated to bark-small CPU offload device mismatch." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig save/load token-ID loss and MLukeTokenizer task AttributeError are separate bugs." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are transformers serve crashes, but on different endpoints and different failure mechanisms: str.to on multimodal batching vs KV-cache tensor indexing." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Qwen2-57B-A14B-Instruct OOM with tensor parallelism and a PEFT version bump 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": 17, + "review_unit_index": 21, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29942", + "issue:30333", + "issue:34567", + "issue:37428", + "issue:38175", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43493", + "issue:43504", + "issue:43540", + "issue:43582", + "issue:43606", + "issue:43653", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44112", + "issue:44315", + "issue:44464", + "issue:44466", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44568", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45356", + "issue:45405", + "issue:45446", + "issue:45520", + "issue:45584", + "issue:45701", + "issue:45823" + ], + "soft_pairs": [ + "issue:44464|issue:45584", + "issue:43504|issue:43653", + "issue:43994|issue:44493", + "issue:43278|issue:43701", + "issue:43606|issue:44112", + "issue:44792|issue:45405", + "issue:29942|issue:30333", + "issue:44521|issue:45290", + "issue:28218|issue:34567", + "issue:45084|issue:45520", + "issue:43844|issue:44964", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:43232|issue:44849", + "issue:43582|issue:44977", + "issue:44568|issue:45701", + "issue:43540|issue:45200", + "issue:37428|issue:43957", + "issue:45072|issue:45198", + "issue:45200|issue:45823", + "issue:44315|issue:44849", + "issue:44554|issue:45356", + "issue:43493|issue:43701", + "issue:43873|issue:44464", + "issue:43295|issue:44610", + "issue:44466|issue:45127" + ], + "prepared_review_unit_hash": "e74357a0621218bf8748faddbced3295a90b489831a38446172933b0bde42115", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35013, + "estimated_input_tokens": 8754, + "estimated_eval_tokens": 17764 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft links are mostly false positives: the issues cover different models, subsystems, and failure modes, so none should be merged as duplicates. If a representative issue is needed, the most active and referenced one is the multimodal chat-template bug.", + "confidence": 0.56, + "canonical_issue_reason": "No clear canonical duplicate: the candidate pairs do not point to one shared underlying bug, so the cluster is better treated as heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the best representative by activity and inbound references, and it sits on an active multimodal API regression with broader user impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Both involve generation, but they are different bugs in different code paths/models: compiled forward chunked generation vs Whisper empty-transcription behavior." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Unrelated failures: BEiT semantic-segmentation loading with a legacy field vs BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Different symptoms and likely different fixes: SigLIP2 output discrepancy vs a broad position-id key warning across many models." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "One is a dtype drift in evaluation; the other is a resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Both mention device-related CI behavior, but they affect different models and different mechanisms (CPU offload mismatch vs stale override test)." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure and an unreleased PEFT version bump are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLflow reporting after training failure are unrelated subsystems." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both are chat-template related, but one is all-zero assistant masks for multimodal inputs while the other is a crash on tool-call assistant messages; not the same bug." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token-count tracking." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Non-template-node compile errors and a flash_attn import KeyError on Python 3.13 are separate issues." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient blow-up and a model-loading regression under latest transformers are different failure modes." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability outputs." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness with differing head dims is not the same as a flex_attention import/version check bug." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs sync_gpus behavior and Qwen3.5 hidden-states output are different code paths." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator TypeError and Qwen3.5 flash-attention generation problems do not share a concrete bug." + }, + { + "left": "issue:44568", + "right": "issue:45701", + "accept": false, + "reason": "A specific tokenizer regression for BOS/EOS addition is broader than a general report that transformers version changes tokenization." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video processing failure in Qwen3OmniMoe and Gemma 4 text-only mm_token_type_ids behavior are separate multimodal issues." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Flash-support import failure and meta-device model-loading regressions are unrelated." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "Bfloat16 dtype mismatches in CI and Wav2Vec2 save/tokenization failures are not the same underlying bug." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma 4 related, but one is about defaulting mm_token_type_ids for text-only finetuning while the other is a device-map mismatch during forward." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application via model_init and Qwen3.5 hidden-state output are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "An MPS attention correctness issue and a Kimi-K2.5 tokenizer codec regression are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 HF-vs-JAX discrepancy and checkpoint resume key mismatch do not describe the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44464", + "accept": false, + "reason": "Quantized offloading behavior and compiled-forward chunked generation inconsistencies are different problems." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer API regression and OmDet-Turbo input-size mismatch are separate processor/model integration bugs." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both involve tied weights/serialization themes, but one is device-dependent serialization of lm_head.weight while the other is LoRA merge collapse with extended vocab." + } + ] + }, + "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": 17, + "review_unit_index": 22, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43575", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43901", + "issue:43957", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44442", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44792", + "issue:44857", + "issue:44898", + "issue:44936", + "issue:44987", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45245", + "issue:45325", + "issue:45412", + "issue:45563", + "issue:45584" + ], + "soft_pairs": [ + "issue:44898|issue:45412", + "issue:37428|issue:43824", + "issue:43572|issue:44987", + "issue:44792|issue:45563", + "issue:43572|issue:43957", + "issue:45200|issue:45325", + "issue:43334|issue:43525", + "issue:43844|issue:44936", + "issue:43295|issue:43450", + "issue:43824|issue:43975", + "issue:43295|issue:43873", + "issue:43388|issue:43526", + "issue:42915|issue:43824", + "issue:38175|issue:45412", + "issue:28218|issue:35141", + "issue:38175|issue:43493", + "issue:44442|issue:45198", + "issue:43278|issue:43856", + "issue:44514|issue:44521", + "issue:43827|issue:43901", + "issue:43653|issue:44112", + "issue:43450|issue:43575", + "issue:43450|issue:43493", + "issue:44206|issue:45584", + "issue:45127|issue:45245", + "issue:44360|issue:44561", + "issue:43262|issue:43450", + "issue:43819|issue:45083", + "issue:44610|issue:44857", + "issue:43819|issue:44112" + ], + "prepared_review_unit_hash": "bb9f7b824f4c51fad782bbea4d688ce3891cf0b84271fd8da6259724d161dc06", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36859, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a mostly heterogeneous cluster with one clearly plausible duplicate pair (the two SigLIP2 issues) and many adjacent-but-distinct processor/model regressions. For the representative issue, I\u2019d anchor on the most activity-heavy broad multimodal/import case rather than overclaiming that issue:44521 is strongest overall.", + "confidence": 0.8, + "canonical_issue_reason": "issue:43824 is a broad, high-activity multimodal import regression with the strongest discussion/inbound signal among the cluster\u2019s more generic candidates, so it works better as the representative issue than issue:44521.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the best global issue representative because it is active, broadly phrased, and sits near the center of the cluster\u2019s multimodal/import failures without relying on an unsupported claim of being the most active overall.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different model families and failure modes: Perceiver resolution handling vs RT-DETR memory release." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different symbols, models, and code paths." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Different loading regressions on different models; similarity is too broad." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure and a stale generate warning are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Both concern loading, but the underlying problems are distinct." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different Gemma/Qwen multimodal bugs affecting different preprocessing logic." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Same general missing-field pattern, but different configs and no clear shared defect." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "ZeRO-3 training instability and post-train evaluate failure are different issues." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor.tokenizer regression and batched video-shape handling hit different paths." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Different model-specific import/detokenization failures." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Custom processor API breakage and quantization offloading are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both mention labels, but one is metric gathering and the other is image-processor label reduction." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Different model families and different failure classes." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 output mismatch and RT-DETR memory release are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization on resize." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": true, + "reason": "Both are SigLIP2 implementation discrepancies, and the zero-probability symptom plausibly stems from the same HF-vs-reference mismatch." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer-related, but different model families and different bugs." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift in eval is not the same as MoE memory inefficiency." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Same processor area, but batched-input crashing and all-zero assistant masks are different failures." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they reference different APIs and regressions." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device-override behavior are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video preprocessing and tensor-parallel OOM are different problems." + }, + { + "left": "issue:43450", + "right": "issue:43493", + "accept": false, + "reason": "Batched video shape bug and SigLIP2 implementation mismatch are different paths." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported feature-extractor arg and empty-transcription generation are distinct audio bugs." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/extended-vocab collapse and category-count limit errors are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer activation choice and a removed helper affecting trust_remote_code are different subsystems." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio apply_chat_template sampling-rate default and video batching are unrelated processor bugs." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent/forward mismatch is unrelated to qwen3_omni_moe feature-length handling." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Preprocessing size mismatch and AMP crash are different issues." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent mismatch is unrelated to a GraniteSpeech CI test failure." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level summary is mostly conservative, but one key claim is not grounded in the packet: issue:44521 is not clearly the strongest by discussion/inbound activity (e.g. issue:43824 has higher discussion and inbound counts). Because the representative-issue choice is justified with an unsupported superlative, I can\u2019t accept this as fully precise. The soft-edge verdicts themselves look mostly conservative, including the single SigLIP2 acceptance, but the top-line rationale needs to be toned down or corrected." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 23, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:32090", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43931", + "issue:44112", + "issue:44206", + "issue:44292", + "issue:44351", + "issue:44448", + "issue:44462", + "issue:44493", + "issue:44561", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45478", + "issue:45520", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:37428|issue:43425", + "issue:43278|issue:43825", + "issue:43335|issue:43550", + "issue:44857|issue:45593", + "issue:43606|issue:45083", + "issue:44448|issue:44743", + "issue:37428|issue:44351", + "issue:45200|issue:45588", + "issue:29942|issue:32090", + "issue:45290|issue:45381", + "issue:43611|issue:43931", + "issue:39692|issue:43526", + "issue:44206|issue:45823", + "issue:44462|issue:44561", + "issue:44561|issue:44779", + "issue:43335|issue:43653", + "issue:45310|issue:45478", + "issue:43819|issue:44493", + "issue:45588|issue:45859", + "issue:29127|issue:32090", + "issue:43550|issue:44112", + "issue:44292|issue:45520", + "issue:44855|issue:45083", + "issue:43335|issue:43504", + "issue:39692|issue:44493", + "issue:43335|issue:45083", + "issue:44448|issue:44779", + "issue:38175|issue:44493", + "issue:43335|issue:45072", + "issue:43335|issue:45593", + "issue:43720|issue:44112", + "issue:44977|issue:45229", + "issue:34689|issue:35141", + "issue:44448|issue:44977" + ], + "prepared_review_unit_hash": "c8dc2187a0f187fa7f0b87cc35eb4edf4079ec7fea45342beefdf18d770cd956", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37360, + "estimated_input_tokens": 9340, + "estimated_eval_tokens": 18936 + }, + "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 identical Qwen3.5 MoE from_pretrained regression. The rest are related only at a very broad Transformers/modeling level, not the same underlying bug.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45310 is the cleanest canonical issue because it exactly matches issue:45478 in title and failure mode, and it was filed first.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best overall representative of the only true duplicate pair in this set; other issues are unrelated one-off bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Different failures: a missing flash-attention utility import vs a general torch-version incompatibility." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "One is an embedding dtype regression; the other is a pipeline error-message issue." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Different code paths: SwitchTransformers sparse-layer config vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "AMP/CUDA loss crash and missing auxiliary losses are separate bugs." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to qwen3_omni_moe feature-length behavior." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Different models and root causes: Pegasus output drift vs Qwen recurrent-state reset." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and different broken APIs." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 token-type-id defaulting and flash_attention aux-state crash are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and a trainer broadcast NoneType error do not share the same bug." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call handling and qwen2.5-vl video position IDs are different issues." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Both are loading errors, but they point to different models and different shape/attribute problems." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Documentation example mistakes are not the same as BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different regressions: unsupported center argument vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Tokenizer.json precedence and removal of is_torch_fx_available are unrelated breakages." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "A removed utility breaking trust_remote_code is unrelated to Deepseek tokenizer output changes." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are different subsystems." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Exact same Qwen3.5 MoE from_pretrained error, same title and same failure mode." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC latent/forward mismatch and unexpected position-id keys are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both mention offload/flash-attention-adjacent code, but the concrete failures differ." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMV3 error-message clarity and GPU broadcast NoneType are not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA compile failure and GraniteSpeech stale device override are different code paths." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime failure and Python 3.13 flash_attn import KeyError are distinct." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python parser/import failure versus qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different BEiT/SwitchTransformers loading bugs with no shared root cause." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs issues and position-id key warnings are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "Sparse-layer creation and dtype mismatch in bfloat16 inference are different bugs." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output differences and Deepseek tokenizer regression are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as position-id key warnings." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Different models and failures: Pegasus v4/v5 output drift vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weight unpacking in BitNet and GraniteSpeech device override are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issue and Gemma4 multi-GPU OOM are different problems." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision model loading and post_init reinitializing embeddings are separate bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: the only accepted pair, issue:45310 and issue:45478, is an exact title/failure match and the earlier filing is correctly identified as canonical. The remaining soft-edge verdicts stay conservative and are consistent with the issue titles and the packet\u2019s lack of stronger link evidence. No overstatement detected." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 24, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28218", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:35141", + "issue:37428", + "issue:40444", + "issue:41720", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43701", + "issue:43723", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:43994", + "issue:44448", + "issue:44462", + "issue:44466", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45712", + "issue:45803" + ], + "soft_pairs": [ + "issue:43723|issue:44554", + "issue:40444|issue:43526", + "issue:43606|issue:44857", + "issue:44857|issue:45072", + "issue:44568|issue:45491", + "issue:28218|issue:32090", + "issue:45198|issue:45803", + "issue:43927|issue:45698", + "issue:45459|issue:45712", + "issue:44977|issue:45563", + "issue:29942|issue:35141", + "issue:37428|issue:44987", + "issue:43701|issue:45083", + "issue:43873|issue:44493", + "issue:45372|issue:45588", + "issue:43550|issue:45083", + "issue:32090|issue:33357", + "issue:43412|issue:45588", + "issue:43278|issue:43329", + "issue:43873|issue:45229", + "issue:45245|issue:45563", + "issue:43976|issue:45440", + "issue:41720|issue:45229", + "issue:28218|issue:29127", + "issue:45372|issue:45406", + "issue:45372|issue:45459", + "issue:43066|issue:44462", + "issue:43065|issue:43994", + "issue:44857|issue:45588", + "issue:43278|issue:43381", + "issue:44448|issue:44466" + ], + "prepared_review_unit_hash": "662ae0457c8e3485879e77b74eea02b7ab301bdd14bb22df5265639a2afd891e", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37205, + "estimated_input_tokens": 9302, + "estimated_eval_tokens": 18860 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No duplicate cluster emerges here: every soft pair spans different concrete bugs or unrelated subsystems, so none should be merged.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue fits this set; the soft pairs are heterogeneous and appear to be false positives rather than duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good representative duplicate target because there is no accepted overlap across the candidates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading in v5 and an MPS attention correctness issue are unrelated code paths." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Multimodal iterable-dataset failures and BEiT label reduction are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch and float16 AMP loss crashes are different failure modes." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both involve dtype issues, but they affect different models and concrete code paths." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token handling and Gemma3 NaN attention outputs are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to Trainer broadcast NoneType errors." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failures and a bare-except tensor conversion bug are not the same issue." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Config token-ID persistence and wrong custom-module loading after save/load are different bugs." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking and leftover dummy classes are unrelated maintenance issues." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation failure and a stale warning for num_return_sequences are different symptoms." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures and embedding reinitialization on resize are unrelated." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Missing flash-attention import symbol and a specific model-load failure are not the same bug." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to qwen3_omni_moe feature-length calculation." + }, + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Quantization offload issues and unexpected position-id keys are different problems." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor import failure and a flash-attention s_aux=None crash are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure for Bamba is unrelated to qwen3_omni_moe output-length logic." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer GPU broadcast errors and a MacOS bus error on CLIP are different code paths." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/erfinv support and flash-attention None handling are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in eval and a multimodal token-count video-branch bug are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization offloading and multi-GPU CUDA OOM are different resource-management issues." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count overflow and a stale generation warning are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility and DeepSeekV3MoE implementation divergence are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 device-mapping assertion and Gemma4 OOM are different models and failures." + }, + { + "left": "issue:28218", + "right": "issue:29127", + "accept": false, + "reason": "Tokenizer added-space behavior and LayoutLMv3 box validation messaging are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma4 processor import failure and Gemma4Processor missing _tokenizer are distinct bugs." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 processor import failure and protobuf error masking are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type and ignoring tokenizer.json are different tokenizer-loading issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d cleanup and Siglip2 nonsensical outputs are unrelated." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP/CUDA loss crash and flash-attention s_aux=None crash are different bugs." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval dtype drift and gradient-checkpointing-in-eval behavior are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus output differences and lm_head serialization inconsistency are different concrete issues." + } + ] + }, + "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": 17, + "review_unit_index": 25, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:36010", + "issue:39692", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43257", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44451", + "issue:44479", + "issue:44610", + "issue:44743", + "issue:44871", + "issue:44977", + "issue:45103", + "issue:45198", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45491", + "issue:45520", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45834", + "issue:45859" + ], + "soft_pairs": [ + "issue:25251|issue:32090", + "issue:43065|issue:45412", + "issue:45588|issue:45823", + "issue:43653|issue:44743", + "issue:44451|issue:45520", + "issue:44186|issue:45198", + "issue:44206|issue:44479", + "issue:43994|issue:45412", + "issue:44361|issue:45198", + "issue:44871|issue:44977", + "issue:42915|issue:42994", + "issue:43421|issue:45491", + "issue:44265|issue:45198", + "issue:29942|issue:30064", + "issue:43257|issue:43278", + "issue:43756|issue:45440", + "issue:33357|issue:39692", + "issue:29127|issue:34567", + "issue:32090|issue:33290", + "issue:44610|issue:45412", + "issue:45103|issue:45372", + "issue:43824|issue:44279", + "issue:45440|issue:45593", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:43761|issue:45859", + "issue:43526|issue:45834", + "issue:29942|issue:36010", + "issue:29127|issue:33357", + "issue:43526|issue:44186" + ], + "prepared_review_unit_hash": "e5d6edfd8bf45ca3a88fd2fb92b8d4db3d53bd5d4a7a8367ab6d1068470fb84e", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36670, + "estimated_input_tokens": 9168, + "estimated_eval_tokens": 18592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items do not appear to form a real duplicate cluster; they span many unrelated bugs, regressions, and feature requests. I would reject all soft edges.", + "confidence": 0.92, + "canonical_issue_reason": "issue:45588 is the most concrete, localized bug report in the set (`integrations/flash_attention.py` crash on `s_aux=None`), but even so this is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45588 is the strongest standalone triage target because it has a specific code path, clear failure mode, and a plausible one-PR fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline `top_k` nesting bug and trainer `_gpu_broadcast_one` NoneType error are unrelated code paths." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy `Conv2d` in Sam3PixelDecoder is unrelated to RT-DETR memory being retained after deletion." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and Gemma4 device-map mismatch are different failures." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBird mask-token registration/empty decode is unrelated to Qwen3.5 recurrent-state reset behavior." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "Model loading failure for ScandiBERT and Python 3.13 `flash_attn` import-key handling are different issues." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 NER/padding crash and Wav2Vec2 save/tokenization failure do not share the same bug." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported `center` arg in `LasrFeatureExtractor` is unrelated to the Qwen video-input regression." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 wrong outputs with AutoModel/pipeline and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLuke tokenizer task AttributeError is not the same as Wav2Vec2 save_pretrained/tokenization failure." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos token config mismatch and Qwen3.5 flash-attention generation issue are different underlying bugs." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and quantized-model saving failure are separate quantization problems." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor update bug is unrelated to Gemma3 NaN embeddings from sliding-window attention." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "`torch.export`/`torch_compilable_check` failure is unrelated to Wav2Vec2 save/tokenization behavior." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems and failures." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to embedding dtype changing at eval time." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer omission and DeepSeekV3MoE implementation divergence are not the same bug." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS CLIP bus error and SigLIP2 docs example errors are different issues despite both involving vision models." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count tracking are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "`_gpu_broadcast_one` NoneType crash and DeepSpeed Adafactor OOM are distinct failure modes." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "OmDet-Turbo processor size mismatch and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation crash and mistral_common `ReasoningEffort` import error do not share a code path." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing `Qwen2_5_VLForConditionalGeneration` export and a generic dependency issue are not the same problem." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeekV3MoE divergence and D-FINE auxiliary-loss behavior are unrelated model bugs." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode and checkpoint key mismatch are different checkpointing problems." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBird special-token registration bug is unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden-state regression and Qwen3_5Moe CPU-offload incompatibility are separate issues." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast `reduce_labels` bug is unrelated to Kosmos2.5 OCR index error." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention 2 test failures and `GenerationMixin` import error are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error clarity and CLIP MacOS bus error are not the same failure." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast label handling and LayoutLMv2 tokenizer crashes are different components and bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 26, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:29127", + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:39692", + "issue:42907", + "issue:42915", + "issue:43232", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43388", + "issue:43441", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43611", + "issue:43646", + "issue:43701", + "issue:43723", + "issue:43761", + "issue:43931", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45072", + "issue:45245", + "issue:45290", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45698", + "issue:45803" + ], + "soft_pairs": [ + "issue:43334|issue:43531", + "issue:29127|issue:35141", + "issue:44464|issue:45491", + "issue:44977|issue:45245", + "issue:42915|issue:43931", + "issue:44442|issue:45072", + "issue:43262|issue:44206", + "issue:43526|issue:44857", + "issue:43232|issue:43388", + "issue:43575|issue:43701", + "issue:43441|issue:45588", + "issue:44743|issue:44811", + "issue:43575|issue:44387", + "issue:43646|issue:44077", + "issue:45003|issue:45440", + "issue:43761|issue:44079", + "issue:43723|issue:45479", + "issue:44442|issue:45803", + "issue:43761|issue:44188", + "issue:43316|issue:43525", + "issue:33290|issue:34567", + "issue:44479|issue:45290", + "issue:44442|issue:44898", + "issue:42907|issue:45698", + "issue:34689|issue:43611", + "issue:39692|issue:45491", + "issue:33290|issue:34689", + "issue:33290|issue:35141" + ], + "prepared_review_unit_hash": "5dd5b343c5c64766c489f2cbe907fa395aafe98290de50da1f7a7ec15ef7479f", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35643, + "estimated_input_tokens": 8911, + "estimated_eval_tokens": 18078 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the issues span unrelated model families, trainer/runtime regressions, tokenizer/processor bugs, and memory/loading failures. None of the proposed soft pairs look like the same underlying bug or change, so I would not merge any of them as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is a grab bag of unrelated bugs, and the soft pairs are only superficial title-level similarities.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best representative only because it is open, has inbound references, and concerns a broad save/load custom-module regression with ongoing impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Different failures: missing pad_token_id on Qwen3-VL text config vs a sliding_window/recurrent-state bug in Qwen3-MoE." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "Unrelated bugs: unclear LayoutLMv3 box error messaging vs token-embedding resizing reinitializing the output embedding." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Both involve generation, but one is compiled forward chunking inconsistency and the other is a Gemma3 sliding-window NaN issue with mixed-length batches." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Completely different problems: FlashAttention generation regression vs a categories-cardinality RuntimeError." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Different root causes: FineGrainedFP8Config failure on Qwen3Moe vs weight-shape mismatch when loading Qwen3-VL." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer-loading failure and bfloat16 dtype mismatch are unrelated despite both being model-adjacent." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Both touch audio processing, but one is a chat-template sampling-rate default bug and the other is an unsupported `center` argument crash." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Different subsystems: BEiT label reduction returns the wrong shape, while LwDetrImageLoss crashes under float16 AMP on CUDA." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs syncing and metric-gathering label truncation are separate code paths with different failure modes." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "OOM during tensor-parallel model load is not the same bug as a resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention FlashAttention, but they hit different models and different failures: Ministral-3 regression vs `s_aux=None` AttributeError." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Unrelated: Qwen3.5 recurrent-state reset bug vs Whisper batch_decode ignoring skip_special_tokens." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both are OOMs, but one is TP model loading and the other is increased reserved memory under int4 quantization; not the same issue." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "General custom-model initialization breakage in v5 is broader and distinct from the specific patchtsmixer post_init constraint." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "sys.modules access safety bug and DeepSeekV3MoE implementation divergence are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "The former is a CLIPVisionModel forward regression; the latter is a general ModelOutput key-assignment bug. Similar symptoms, different bugs." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer auto-loading and degenerate zero-loss classification are completely different problems." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure is unrelated to Fuyu's bare-except tensor conversion bug." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Both are regressions, but one is hidden_states output handling and the other is attention-kernel divergence under torch.compile." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both involve config APIs, but one is a broad Gemma3TextConfig discrepancy and the other is a specific missing pad_token_id on Llama4Config." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Optimizer/deepspeed OOM and TrainerState token-count bookkeeping are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression across Qwen VL variants is unrelated to chat-template crashes on tool-call assistant messages." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading and Perceiver non-default-resolution classification failure are not the same code path." + }, + { + "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." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both affect model loading, but they target different mechanisms: a specific Llama 3.2 vision regression vs base_model_prefix handling in v5." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "Documentation-example errors and Gemma3 NaN embeddings are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "OOM in deepspeed/adafactor and model-loading failure for Llama 3.2 vision are different issues." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Completely different: training-memory OOM vs embedding reinitialization during post_init after resizing tokens." + } + ] + }, + "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": 17, + "review_unit_index": 27, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:41628", + "issue:42175", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43493", + "issue:43540", + "issue:43550", + "issue:43643", + "issue:43673", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43825", + "issue:43827", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44617", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45137", + "issue:45216", + "issue:45372", + "issue:45759" + ], + "soft_pairs": [ + "issue:43381|issue:43873", + "issue:43931|issue:45216", + "issue:43493|issue:43827", + "issue:43550|issue:44743", + "issue:44077|issue:44479", + "issue:41628|issue:45372", + "issue:43316|issue:45759", + "issue:43720|issue:44442", + "issue:43643|issue:44913", + "issue:43749|issue:43931", + "issue:43873|issue:44871", + "issue:33290|issue:36331", + "issue:34689|issue:36010", + "issue:43381|issue:43856", + "issue:44190|issue:44496", + "issue:44855|issue:45137", + "issue:43381|issue:43742", + "issue:43316|issue:44877", + "issue:42175|issue:43316", + "issue:43540|issue:44514", + "issue:25251|issue:28282", + "issue:43673|issue:44743", + "issue:43065|issue:44617", + "issue:43404|issue:43441", + "issue:44162|issue:44792", + "issue:43388|issue:43688", + "issue:43295|issue:43825" + ], + "prepared_review_unit_hash": "96e4d0039f1a1eed1367b69a33c25ed61b1c174dbb3820739befde62c59dd65a", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35282, + "estimated_input_tokens": 8821, + "estimated_eval_tokens": 17898 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: it mixes unrelated import, config, training, loading, and multimodal processing issues. None of the soft-similarity pairs look like the same underlying bug, so all soft edges are rejected.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43825 is the clearest, most self-contained user-facing regression in the set, with a specific problem statement and broad pipeline-v5 relevance.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43825 is the safest anchor if one issue must represent the cluster: it is actionable, well-scoped, and easier to generalize than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Different failure modes: gradient checkpointing vs quantization/offloading." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen models, but one is a shape mismatch on load and the other is a save_pretrained regression." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is unrelated to docs still mentioning pipeline()." + }, + { + "left": "issue:43550", + "right": "issue:44743", + "accept": false, + "reason": "Different model families and bugs: torch.compile/SDPA vs recurrent state reset with cache." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Unrelated model-specific validation issue versus a video-input regression." + }, + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "Both are import errors, but caused by different packages and different processor paths." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Related to config APIs, but these are different composite-model bugs with different triggers." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Packed-weight loading bug is unrelated to tokenizer loading failure." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are config reload problems, but one drops missing fields while the other resets a specific parameter." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Different subsystems: FSDP CPU RAM loading versus Qwen3-VL shape mismatch." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to Gemma eos_token_id inconsistency." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "Optimizer OOM in DeepSpeed is unrelated to a trainer method signature change." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Both are model/loading errors, but one is a specific model regression and the other is a missing import path." + }, + { + "left": "issue:43381", + "right": "issue:43856", + "accept": false, + "reason": "Eval-mode gradient checkpointing and Qwen3 MoE memory usage are different problems." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Dataset loading in a training script is unrelated to missing model_type in config.json." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python parser error in torch.jit scripting is unrelated to DeepSpeed deque underflow." + }, + { + "left": "issue:43381", + "right": "issue:43742", + "accept": false, + "reason": "Gradient checkpointing eval-mode issue is unrelated to a MobileLLM key error." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config handling, but one is an API discrepancy and the other is a strict loader rejection." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Packaging/backend dependency issue is unrelated to Gemma3TextConfig API mismatch." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are multimodal input bugs, but they affect different processors, inputs, and code paths." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k nesting bug is unrelated to a missing PyTorch dependency error." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is GenerationMixin cache availability and the other is recurrent state reset logic." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d implementation in Sam3PixelDecoder is unrelated to a Sam3Video OOM report." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Different Mistral-family issues: weight tying versus FlashAttention failure." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "Completely unrelated: ESM2 breakage versus a janus image-generation test failure." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric label dropping and auxiliary-loss normalization are different training/eval bugs." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both are regression reports, but one is processor/tokenizer behavior and the other is pipeline v5 messaging." + } + ] + }, + "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": 17, + "review_unit_index": 28, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:32090", + "issue:35141", + "issue:41950", + "issue:42617", + "issue:43065", + "issue:43295", + "issue:43452", + "issue:43475", + "issue:43504", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43701", + "issue:43742", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45290", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45698" + ], + "soft_pairs": [ + "issue:16998|issue:22355", + "issue:41950|issue:43475", + "issue:44077|issue:44534", + "issue:44871|issue:45584", + "issue:45127|issue:45200", + "issue:43550|issue:43701", + "issue:43761|issue:44479", + "issue:43606|issue:43701", + "issue:43295|issue:44871", + "issue:43575|issue:43742", + "issue:42617|issue:43575", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:44734|issue:45406", + "issue:43504|issue:45698", + "issue:44617|issue:45584", + "issue:43575|issue:43873", + "issue:43761|issue:44242", + "issue:43931|issue:44060", + "issue:32090|issue:35141", + "issue:45563|issue:45584", + "issue:44464|issue:44871", + "issue:44792|issue:44977", + "issue:44479|issue:44734", + "issue:43065|issue:44610", + "issue:43827|issue:43994", + "issue:44534|issue:44913", + "issue:44521|issue:45381", + "issue:44933|issue:45362", + "issue:44464|issue:45563", + "issue:43906|issue:45290" + ], + "prepared_review_unit_hash": "e62f28d9f211f864c2b85fabe8c2daedf37fadee412a754206fd5fe0789812df", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36836, + "estimated_input_tokens": 9209, + "estimated_eval_tokens": 18674 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab bag of mostly unrelated Transformers issues across loading, generation, multimodal processors, quantization/offload, and model-specific regressions. None of the soft pairs look like true duplicates or the same fixable code-path bug, so all soft edges should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "issue:44521 is the best anchor: it is open, has the most external attention in this set, and describes a concrete reproducible multimodal bug rather than a vague question or isolated model-specific failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone issue to keep: specific reproduction, active discussion, and likely to benefit from continued triage/fix work.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Different problems: a model_max_length question for DeBERTa-V3 vs a missing `transformers.onnx` module import error." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Both involve video/SAM-like models, but one is a pipeline processor selection bug and the other is a missing attribute on a specific SAM3 output object." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Different code paths: one is about `post_init` being optional for patchtsmixer, the other about non-persistent buffers being filled with junk in v5." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Different bugs: Gemma-3 eos token configuration mismatch vs Whisper failing on empty transcription after `align_special_tokens`." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "Both relate to embeddings/tokens, but one is LoRA merge collapse with tied embeddings and the other is Gemma 4 multimodal token type IDs defaulting to zeros." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "Unrelated failures: a torch.compile/SDPA issue for Bamba vs a resume-from-checkpoint key mismatch." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "Different models and regressions: CLIPVision hidden_states handling vs Qwen video-input regression in v5.3.0." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "Different issues: CPU offload device mismatch for bark-small vs checkpoint key mismatch during resume." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Processor/tokenizer regression vs Gemma-3 eos_token configuration mismatch; not the same underlying bug." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "Both are load-time failures, but one is TP/OOM for Qwen2-57B and the other is a key error loading MobileLLM-125M." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "A runtime script failure (`3d_parallel.py`) and a model-loading OOM are different root causes." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Different loading problems: `gguf_file` breaking tokenizer/model from_pretrained vs an unrecognized model type in config.json." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model initialization breakage in v5 is unrelated to GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Different subsystems: KV-cache continuation indexing in serve vs Gemma4Processor missing `_tokenizer`." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy-field preset loading for BEiT is unrelated to wrong custom module resolution after save_pretrained." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is not the same bug as Whisper failing on empty transcription." + }, + { + "left": "issue:43575", + "right": "issue:43873", + "accept": false, + "reason": "Both mention large-model memory behavior, but one is TP loading OOM and the other is quantization/offloading not working as expected." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "Different model behaviors: CLIPVision hidden_states regression vs MoE load-balancing loss omission." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch on load is unrelated to the Qwen3-Next tied-weights warning bug." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer `_gpu_broadcast_one` NoneType error and embedding reinitialization after resize are different failure modes." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale warning in paged generate is not the same as Whisper's empty-transcription crash." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is unrelated to eos_token_id config mismatch." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation failure are separate model-specific regressions." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both are generation-related, but one is a video-input regression for specific Qwen VL models and the other is a serve KV-cache indexing crash." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo processor size mismatch are different vision-model issues." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing removed `pipeline()` APIs is unrelated to SigLIP2 producing nonsensical outputs." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization junk and GPTNeoX rotary_pct not reloading are unrelated configuration/runtime bugs." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/video-adjacent, but one is assistant mask construction in chat templating and the other is wrong video vision position IDs." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import from `image_utils` is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Both touch generation, but one is a correctness regression in compiled chunked generation and the other is only a stale warning." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "An isolated reproduction of a prior issue and a chat-template crash on tool-call messages 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": 2, + "component_count": 17, + "review_unit_index": 29, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:41628", + "issue:41720", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43388", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43582", + "issue:43644", + "issue:43650", + "issue:43976", + "issue:44016", + "issue:44077", + "issue:44079", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44279", + "issue:44466", + "issue:44496", + "issue:44683", + "issue:44734", + "issue:44737", + "issue:44792", + "issue:44855", + "issue:44913", + "issue:44938", + "issue:44964", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45357", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45803" + ], + "soft_pairs": [ + "issue:44190|issue:44964", + "issue:41720|issue:45245", + "issue:44792|issue:45600", + "issue:44855|issue:45588", + "issue:44737|issue:45803", + "issue:45127|issue:45357", + "issue:43388|issue:43582", + "issue:44938|issue:45542", + "issue:43335|issue:44855", + "issue:43644|issue:44913", + "issue:44077|issue:44079", + "issue:43976|issue:44855", + "issue:43550|issue:44190", + "issue:44496|issue:45803", + "issue:45276|issue:45584", + "issue:44112|issue:45600", + "issue:43540|issue:44734", + "issue:43650|issue:44220", + "issue:43335|issue:44190", + "issue:43295|issue:45276", + "issue:43116|issue:44190", + "issue:45245|issue:45584", + "issue:41950|issue:44016", + "issue:36246|issue:41628", + "issue:43479|issue:45375", + "issue:36246|issue:38617", + "issue:42175|issue:44279", + "issue:44016|issue:44220", + "issue:38175|issue:43065", + "issue:44186|issue:44734", + "issue:44466|issue:44913", + "issue:44683|issue:45446" + ], + "prepared_review_unit_hash": "0a90c049b74b336552fa37f4498ae8821ed6b21ab2c3be03b103c757bbb18726", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37510, + "estimated_input_tokens": 9378, + "estimated_eval_tokens": 19012 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: most pairs are only superficially similar (often just the same subsystem or API area) and do not describe the same underlying bug. I would not merge any of the soft edges; the strongest recurring theme is import/API regressions, but even those are distinct failures.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 41628 is the broadest import/API-regression report among the set and is the closest match to the other import-related items, but the cluster overall is too mixed to treat it as a true duplicate center.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41628 is the best single representative if one must be chosen because it reflects a common user-facing import failure pattern, but it should not be used as a duplicate canonical for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Both are model/script loading problems, but one is a local dataset issue in an example script and the other is a multimodal model load failure; different code paths." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert is unrelated to the 2^24 category-limit runtime error." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus image-generation test failure and Sam3LiteText auto_mappings CI breakage are different failures in different areas." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 import parsing error in DebertaV2Model is unrelated to the flash_attention s_aux=None crash." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet CPU tensor creation bug and FuyuBatchFeature bare-except issue are distinct code-path problems." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen serialization, but one is LoRA/tied-weights collapse after merge while the other is incorrect visual encoder keys in save_pretrained." + }, + { + "left": "issue:43388", + "right": "issue:43582", + "accept": false, + "reason": "gather_for_metrics label truncation and AppleSilicon warmup TypeError are unrelated." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Both are packaging/backend-related, but one is Python 3.14 import failure and the other is missing TensorFlow backend wiring for tensorboard." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and DebertaV2 Python 3.13 import parsing error are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer serialization junk and GPTNeoX rotary_pct reload default are different state/serialization bugs." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "Optional post_init in patchtsmixer and ModelOutput key assignment from None are separate model-output/config issues." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "A Python-version compatibility complaint and a Python 3.13 indentation/import parsing error are not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile SDPA failure is unrelated to the local dataset loading issue." + }, + { + "left": "issue:44496", + "right": "issue:45803", + "accept": false, + "reason": "Unrecognized model/config.json failure and Fuyu bare-except bug are unrelated." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resize propagation and Whisper empty transcription handling are different model behaviors." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both involve CI failures, but one is a stale GraniteSpeech test and the other is auto_mappings referencing removed configs." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video input validation and serve KV-cache continuation indexing are different code paths." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "The vague 'ADD THE DATA' issue and the fbank extraction bug are not clearly the same underlying problem." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers config construction bug and local dataset loading error are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer regression and Gemma4 resize_token_embeddings propagation bug are different API failures." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Multi-label classification empty results and local dataset loading are unrelated example-script issues." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category cardinality limit error and Whisper empty transcription are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and notebook syntax error are not the same bug." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import failures, but they concern different symbols and different surfaces: a missing Qwen2.5 VL image processor submodule export versus top-level AutoImageProcessor import." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Phi4MultimodalConfig default initialization and Qwen3.5MoEVisionConfig missing deepstack_visual_indexes are distinct config-schema bugs." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import errors, but one is a missing model class export and the other is a missing utility from configuration_utils; different regressions." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "TensorFlow backend packaging issue and generic dependency issue are too broad and not the same defect." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and audio feature extraction bug are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 and dummy Conv2d in Sam3PixelDecoder are unrelated model-specific bugs." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER padding/truncation and serve KV-cache continuation indexing are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "lm_head.weight serialization inconsistency and GPTNeoX rotary_pct reload default are different serialization/config issues." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and incorrect PyTorch version check for AuxRequest import are unrelated compatibility bugs." + } + ] + }, + "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": 17, + "review_unit_index": 30, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:41950", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43404", + "issue:43441", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44112", + "issue:44242", + "issue:44246", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44561", + "issue:44734", + "issue:44933", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45020", + "issue:45042", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45325", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45406", + "issue:45584", + "issue:45593", + "issue:45600", + "issue:45632", + "issue:45759" + ], + "soft_pairs": [ + "issue:44734|issue:45632", + "issue:44561|issue:45632", + "issue:43404|issue:45216", + "issue:44246|issue:45003", + "issue:43643|issue:44496", + "issue:45406|issue:45632", + "issue:44964|issue:45600", + "issue:44112|issue:44964", + "issue:43295|issue:45335", + "issue:43854|issue:44964", + "issue:42907|issue:43441", + "issue:44016|issue:44062", + "issue:43976|issue:44987", + "issue:43673|issue:44336", + "issue:43531|issue:45600", + "issue:44496|issue:45042", + "issue:43824|issue:44933", + "issue:43526|issue:43901", + "issue:45335|issue:45584", + "issue:43299|issue:45216", + "issue:44075|issue:44368", + "issue:43976|issue:44279", + "issue:43901|issue:44496", + "issue:43381|issue:45593", + "issue:43299|issue:45357", + "issue:44496|issue:45759", + "issue:44351|issue:45278", + "issue:44521|issue:45325", + "issue:41950|issue:45020", + "issue:44242|issue:45200", + "issue:43976|issue:45405" + ], + "prepared_review_unit_hash": "f1e8b44bf8fe077aff81c724efe46eae2e196dff1028673c5ea7086f8001d020", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36956, + "estimated_input_tokens": 9239, + "estimated_eval_tokens": 18734 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is dominated by Transformers regressions around model loading, remote_code, multimodal processors, and version upgrades, but the soft pairs mostly describe different concrete failure modes. I would not merge any of the proposed soft edges.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45020 is the broadest representative because it captures the shared theme of recent-version regressions breaking remote_code/model-loading flows across multiple models, while most other issues are narrower symptoms or model-specific variants.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best cluster anchor for duplicate triage because it is the most general, user-facing description of the underlying regression class.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Serve KV-cache indexing failure and trust_remote_code cache-path collision are unrelated code paths." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Both mention trust_remote_code, but one is a missing import/helper regression and the other is a cache-directory collision." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different models and different failure modes: tied weights vs incorrect save_pretrained checkpoint contents." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import slowness and unsafe sys.modules access are related only at a very broad level, not the same bug." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "One is missing fields from AutoConfig with trust_remote_code; the other is an unrecognized model/config validation error." + }, + { + "left": "issue:45406", + "right": "issue:45632", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer in serve is not the same as a trust_remote_code cache collision." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "A model-load regression and stale auto_mappings references are different defects." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI device-override flakiness is not the same as a user-facing model-loading failure." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor/tokenizer regression and decoder embedding resize behavior are separate bugs." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "They concern different models with different loading failures." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized models and FlashAttention loading errors are distinct issues." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken argument conflict have no shared code path." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility and a specific model-loading regression are not duplicates." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Cache missing in generation and ANSI-code emission in loading_report are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and removed auto_mappings configs are different problems." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Model-type recognition failure and PIL processor torchvision dependency are separate loading issues." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "These are both import errors, but for different symbols and different root causes." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "A processor bug and stale docs about return_all_scores are not the same underlying change." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize and Whisper empty-transcription handling are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Loading Qwen3VL MoE models and saving an incorrect Qwen3.5 checkpoint are different regressions." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD args and a tie_word_embeddings warning are not the same defect." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python support mismatch and a vague dependency issue are too different to merge." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Documentation mismatch and config/model recognition failure do not share a concrete code path." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior and D-FINE auxiliary loss logic are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen models, but one is loading and the other is save_pretrained key serialization." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Unrecognized model config and AutoModel text_config unwrapping are different configuration-path bugs." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing export and a broad collection of import errors are not the same concrete issue." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Assistant-mask generation and temporal position-id scaling are different multimodal bugs." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video pipeline processor lookup and remote_code breakage are only loosely related subsystem issues." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss and Gemma4 token-type defaults are separate training bugs." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python compatibility and an unreleased minimum dependency bump are unrelated release issues." + } + ] + }, + "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": 17, + "review_unit_index": 31, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:33290", + "issue:38175", + "issue:39692", + "issue:41950", + "issue:42994", + "issue:43065", + "issue:43295", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43479", + "issue:43525", + "issue:43577", + "issue:43650", + "issue:43653", + "issue:43723", + "issue:43854", + "issue:43901", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44462", + "issue:44466", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:45081", + "issue:45216", + "issue:45276", + "issue:45290", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45823", + "issue:45834" + ], + "soft_pairs": [ + "issue:43479|issue:43525", + "issue:41950|issue:45823", + "issue:38175|issue:44617", + "issue:44683|issue:45542", + "issue:44279|issue:45405", + "issue:43577|issue:45216", + "issue:39692|issue:43901", + "issue:43299|issue:43854", + "issue:43295|issue:44779", + "issue:44462|issue:45081", + "issue:41950|issue:43650", + "issue:44297|issue:44462", + "issue:33290|issue:45656", + "issue:44462|issue:44779", + "issue:44016|issue:44291", + "issue:45600|issue:45712", + "issue:43450|issue:44279", + "issue:43441|issue:45834", + "issue:44466|issue:45276", + "issue:42994|issue:43317", + "issue:44261|issue:44263", + "issue:43065|issue:44898", + "issue:43441|issue:44683", + "issue:43650|issue:45290", + "issue:43653|issue:45593", + "issue:44623|issue:45397", + "issue:44623|issue:44749", + "issue:44671|issue:44779", + "issue:43723|issue:44162" + ], + "prepared_review_unit_hash": "20967b1afdafc5984258ae8d4f7afb7a1e33cef20ea51d74bfa2c12f9a66cc3a", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36329, + "estimated_input_tokens": 9083, + "estimated_eval_tokens": 18422 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly unrelated model-specific bugs, regressions, and docs/config problems. None of the soft pairs look like true duplicates, so I would reject all soft edges.", + "confidence": 0.83, + "canonical_issue_reason": "If forced to pick one representative issue, 44261 is the strongest standalone report: it is open, detailed, and describes a concrete regression around a missing config value rather than a vague symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "44261 is the best anchor for the set because it is active, well-scoped, and technically actionable; however, the cluster overall does not appear to be a duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different bugs: multimodal config defaults vs a missing pad_token_id attribute on a different config class." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Different code paths: video pipeline processor lookup vs Gemma4 forward device placement mismatch." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Unrelated symptoms: zero probabilities in SigLIP2 vs CUDA OOM in Sam3Video." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Different subsystems: torch flex_attention compilation vs TensorFlow backend installation detection." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A vague dependency issue is not the same as a specific PEFT minimum-version bump." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: dtype propagation for BLIP2 vs incorrect checkpoint saving for Qwen3.5." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both mention docs, but they are different documentation problems for different pipelines and behaviors." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models and failures: Qwen3VL MoE loading vs GLM-4.7-Flash test loading." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions: processor.tokenizer/image handling vs Deepseek tokenization output correctness." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Loading tokenizer.json from repo is unrelated to the Mistral regex patch AttributeError." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "A video-classification pipeline bug is unrelated to the placeholder 'ADD THE DATA' issue." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Both touch tokenizer serialization/loading, but one is save_pretrained metadata mismatch and the other is AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Different deepspeed optimizer problems: OOM with Adafactor vs optimizer.step being called twice." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions in different components: repository tokenizer loading vs Deepseek tokenizer output changes." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "A notebook syntax error is unrelated to the init_empty_weights TypeError regression." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are CI hygiene issues, but they target different files and failures: stale mappings vs leftover dummy classes." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Incorrect batched video processor shapes are unrelated to a generic dependency issue." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Ministral FlashAttention failure and Kosmos2.5 OCR index error are different model-specific bugs." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Different serialization and resizing bugs affecting different models and code paths." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized model saving failure is not the same as dequantized model loading with gpu/cpu offload." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Completely different bugs: MLA layernorm epsilon precision issue vs torch.split return values in a MoE indexer." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to Perceiver non-default-resolution image classification failure." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Different attention-stack bugs: FlashAttention failure in a model vs compiled flex_attention failure on newer torch." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "The placeholder issue has no concrete overlap with chat template crashes on tool-call assistant messages." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration is unrelated to D-FINE auxiliary loss handling." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Missing processor.save_pretrained files and gemma-4 zero3 loading are different bugs." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing save_pretrained files and a performance slowdown after upgrading are not the same underlying issue." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Different model regressions: CamemBERT masked LM predictions vs Deepseek tokenizer output correctness." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Generic tokenizer loading regression is not the same as the broader ESM2 breakage report." + } + ] + }, + "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": 17, + "review_unit_index": 32, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42491", + "issue:42757", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43299", + "issue:43352", + "issue:43525", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44361", + "issue:44373", + "issue:44496", + "issue:44521", + "issue:44589", + "issue:44617", + "issue:44987", + "issue:45250", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:45823" + ], + "soft_pairs": [ + "issue:43784|issue:45542", + "issue:43262|issue:44521", + "issue:44016|issue:45823", + "issue:43673|issue:43761", + "issue:44361|issue:45593", + "issue:43901|issue:43994", + "issue:42175|issue:43976", + "issue:44987|issue:45542", + "issue:43122|issue:43650", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:44297|issue:45397", + "issue:43525|issue:43643", + "issue:44188|issue:44336", + "issue:43295|issue:44617", + "issue:44589|issue:45362", + "issue:42491|issue:43299", + "issue:43901|issue:44246", + "issue:41628|issue:43976", + "issue:45561|issue:45563", + "issue:44336|issue:44373", + "issue:44242|issue:45593", + "issue:43352|issue:45250", + "issue:41950|issue:44220", + "issue:44295|issue:45397", + "issue:44246|issue:44496", + "issue:43531|issue:45362", + "issue:36246|issue:42757", + "issue:36246|issue:44351" + ], + "prepared_review_unit_hash": "56ad95dd2a41d945ffc8306506b6c5deefc943d380ff17b5aef42bcab27c49c8", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35506, + "estimated_input_tokens": 8877, + "estimated_eval_tokens": 18010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose thematic bucket of Transformers regressions, import/config errors, and unrelated model-specific bugs rather than a true duplicate set. None of the soft pairs share the same concrete failure path, so they should stay separate.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the cluster, 43299 is the closest fit because it is a concrete v5 model-loading regression and is thematically central to the set, but it is still not a duplicate anchor for the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different code paths: sentence-transformers import NameError vs missing TensorFlow backend from a torch-only install." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Audio chat-template sampling-rate default bug vs multimodal assistant_mask generation bug." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error is unrelated to the Gemma4 device-map forward mismatch." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache/chunked prefill regression vs CLIPVisionModel hidden_states being dropped; different subsystems." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; no shared failure mechanism." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores vs bad model outputs for siglip2." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Packaging/backend dependency problem vs Python version support regression." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "Model-loading failure for a specific checkpoint vs tf backend undefined due to missing TensorFlow." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer behavior regression vs an unrelated placeholder/data issue." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Specific Qwen2.5-VL import export failure vs Python 3.9+/3.10+ compatibility problem." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config vs glm-4v loading failure; different model/config bugs." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer metadata mismatch on save_pretrained vs a zero3/from_pretrained loading issue." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing config attribute vs trust_remote_code stripping fields; related broadly to configs, but not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile attention-kernel divergence vs ANSI codes leaking from loading_report." + }, + { + "left": "issue:43295", + "right": "issue:44617", + "accept": false, + "reason": "Processor/tokenizer regression vs CUDA OOM; completely different symptoms and fixes." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage import/type error vs Qwen3.5 chat crash; no shared code path." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3-MoE/Qwen3VL loading, but one is cross-version LoRA compatibility and the other is a main-branch loading regression; not the same concrete bug." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Documentation issue vs import-time slowness; unrelated." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor export issue vs Python version compatibility regression." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "pytest-xdist race in testing utilities vs stale warning for num_return_sequences." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI escape codes in loading_report vs incorrect docstring for position_ids." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Missing load-balancing loss vs missing auxiliary losses; both are loss-related but for different models and logic." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Unsupported Flash Attention 2 warning for one model vs a generic Flash Attention topic; not the same fix." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline using image processors vs a torch fbank extraction issue." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Buffer/read error for position_ids vs Gemma4 loading failure; different layers." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowness vs unrecognized model config; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug vs Qwen3.5 chat crash; different model behaviors." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Qwen2.5-VL image processor import error vs missing is_offline_mode export from huggingface_hub." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Qwen2.5-VL image processor import error vs HybridCache import error; different missing symbols." + } + ] + }, + "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": 17, + "review_unit_index": 33, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36246", + "issue:42886", + "issue:43065", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43408", + "issue:43475", + "issue:43526", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43673", + "issue:43901", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44368", + "issue:44466", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45459", + "issue:45561", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45701", + "issue:45759" + ], + "soft_pairs": [ + "issue:43526|issue:45003", + "issue:44336|issue:44485", + "issue:45003|issue:45561", + "issue:43673|issue:44561", + "issue:44246|issue:44871", + "issue:44242|issue:44857", + "issue:36246|issue:43596", + "issue:44162|issue:45200", + "issue:44779|issue:45292", + "issue:44295|issue:44749", + "issue:44877|issue:45375", + "issue:45292|issue:45584", + "issue:45292|issue:45701", + "issue:43901|issue:45003", + "issue:42886|issue:45229", + "issue:44987|issue:45459", + "issue:45375|issue:45759", + "issue:44496|issue:45003", + "issue:43232|issue:44336", + "issue:43257|issue:45137", + "issue:43065|issue:43408", + "issue:43262|issue:44246", + "issue:44995|issue:45200", + "issue:44060|issue:44077", + "issue:43475|issue:43540", + "issue:44466|issue:45335", + "issue:44995|issue:45588", + "issue:43653|issue:44162", + "issue:44368|issue:44977", + "issue:44368|issue:45563" + ], + "prepared_review_unit_hash": "bd564781241a6646892dac37a2f53ac9608df3867fe9f42aa9da9a4bebbd334e", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36908, + "estimated_input_tokens": 9227, + "estimated_eval_tokens": 18710 + }, + "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. The only soft duplicate-looking pair is the `resize_token_embeddings` regression across model variants; the rest are different code paths or different model families.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 45292 is the broadest formulation of the `resize_token_embeddings` bug, with 45335 looking like a model-specific instance of the same failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "45292 is the best representative because it states the generic API break, while 45335 narrows it to t5gemma\u2019s decoder embeddings.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "Different subsystems: image processor label reduction vs unsafe `sys.modules` access in modeling utils." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "Unrelated bugs: terminal ANSI output vs GLM-5 RoPE implementation." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Generic import/cache behavior and xdist file race are different code paths and failure modes." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "One is generation cache handling; the other is removal of a helper breaking `trust_remote_code` models." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import-time slowdown is unrelated to Gemma eos token config mismatch." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "MoE load-balancing loss and AMP/CUDA image-loss crash are different model/components." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5 VL import error is unrelated to DeepSpeed ZeRO3/BertModel index failure." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 token-type defaulting are different model bugs." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer behavior regression is not the same as resizing token embeddings not updating outputs." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position ID buffer bug and v5 filtering slowdown do not share the same underlying cause." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-related, but one is strict config loading and the other is a missing field in Qwen vision config." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "`resize_token_embeddings` and Whisper empty-transcription failure after token alignment are different regressions." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize bug is unrelated to tokenizer behavior changing across versions." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for pipeline output is unrelated to unsafe module lookup." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache loading failure is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Physical-intelligence model load failure differs from tokenizer decode-error masking." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing config field in Qwen vision config is unrelated to `from_config` text_config unwrapping." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model config and unsafe `sys.modules` access are different issues." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation kwargs update bug is unrelated to ANSI rendering in loading_report." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Qwen3 MOE weight conversion and DeepSpeed deque underflow are different problems." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as sam3_video/sam3_tracker config mismatch." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio processor sampling-rate defaulting is unrelated to import-time slowdown." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Stale indexer cache crash in GlmMoeDsa is unrelated to Gemma 4 mm token type defaults." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning and patchtsmixer post_init policy are different concerns." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM 3 Video missing attribute and Qwen3OmniMoe video input validation are different bugs." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both mention tied/resized embeddings, but 44466 is a serialization inconsistency while 45335 is a resize propagation bug." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Second-forward stale cache crash and flash-attention `s_aux=None` crash are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage do not share a code path." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A warning about `tie_word_embeddings` is not the same as a flash-attention generation failure." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Different generation warnings: tied-embedding config vs stale `num_return_sequences` warning." + } + ] + }, + "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": 17, + "review_unit_index": 34, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43704", + "issue:43873", + "issue:43931", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44222", + "issue:44297", + "issue:44351", + "issue:44361", + "issue:44368", + "issue:44623", + "issue:44661", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45459", + "issue:45478", + "issue:45600", + "issue:45656", + "issue:45753" + ], + "soft_pairs": [ + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:43673|issue:44734", + "issue:43931|issue:44368", + "issue:44933|issue:45459", + "issue:44869|issue:45356", + "issue:43232|issue:43673", + "issue:44297|issue:44623", + "issue:43673|issue:45278", + "issue:43646|issue:44162", + "issue:44361|issue:44749", + "issue:44623|issue:45656", + "issue:42994|issue:43482", + "issue:28282|issue:30990", + "issue:44749|issue:45310", + "issue:43475|issue:43643", + "issue:44749|issue:45656", + "issue:42371|issue:45092", + "issue:43475|issue:45381", + "issue:41628|issue:44933", + "issue:45230|issue:45397", + "issue:44038|issue:44623", + "issue:45081|issue:45103", + "issue:42994|issue:43525", + "issue:45375|issue:45478", + "issue:43475|issue:44661", + "issue:42673|issue:43704", + "issue:45310|issue:45375", + "issue:43873|issue:44945", + "issue:43232|issue:44188", + "issue:41628|issue:44351", + "issue:42831|issue:44222", + "issue:44945|issue:45753" + ], + "prepared_review_unit_hash": "c8c7d5772c351d22eb551242e8f4f0efdd78bc4932154692a6b179267afcd3fa", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37327, + "estimated_input_tokens": 9332, + "estimated_eval_tokens": 18920 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous, but there are two clear duplicate-style pockets: the repeated Qwen3ForCausalLM VRAM leak reports, and the repeated Qwen3.5 MoE from_pretrained error reports. The rest span unrelated import, generation, tokenizer, save/load, and model-specific regressions.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella item: it captures the general Transformers v5 import-regression theme better than the more model-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the most reusable representative for this cluster because it reflects the broad upgrade/import breakage pattern rather than a narrow model-specific failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Different bugs: Zero3 from_pretrained failure vs deepspeed optimizer-step-count issue." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error and auto_mappings CI breakage are unrelated failure modes." + }, + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both involve generation internals, but the code paths differ: chunked_prefill cache handling vs /v1/responses tensor indexing." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch is not the same as a Qwen3.5 LoRA warning about tie_word_embeddings." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Missing image_utils import and tokenizer error masking without protobuf are separate problems." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash and Kimi tokenizer codec regression are different model/tokenizer bugs." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both touch generation state, but one is sync_gpus kwargs update and the other is chunked_prefill cache missing." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save metadata mismatch and processor.save_pretrained missing files are distinct serialization bugs." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "Generation cache handling is unrelated to the broad import-error regression report." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage is not the same as the ESM2 workflow regression." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError and performance slowdown after upgrade are different issues." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save_pretrained file omission does not match a deepspeed double-step optimizer bug." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving and GGUF loading failure are opposite directions and different bugs." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Missing PyTorch import error is not the same as a loading hang in Sentence Transformers." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Upgrade slowdown and Qwen3.5 Moe from_pretrained failure are not the same code-path problem." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM3 output attribute missing and AutoConfig trust_remote_code missing fields are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API-warning/deprecation and remote-code meta-init incompatibility are different topics." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are vision/model regressions, but the specific missing attribute and wrong position_ids bugs are distinct." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different causes." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report provides no evidence of the same underlying failure as the Gemma-4 Zero3 issue." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3.5-Moe transformers-v5 incompatibility and processor save_pretrained file loss are separate regressions." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex crash and auto_docstring annotation crash are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized save failure is unrelated to Llama4Config missing pad_token_id." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Related Qwen3.5-Moe family, but one is a missing config field dropped by strict and the other is a from_pretrained failure; not the same concrete bug." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM3 vision output attribute missing and add-new-model-like tokenizer mapping failure are different problems." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Same title and same VRAM-leak symptom for Qwen3ForCausalLM in multiple dataloader threads; this looks like the same underlying bug." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3.5-Moe-related, but one is a from_pretrained error and the other is a missing vision config field; too different to merge." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior and pipeline-parallel output corruption are different bugs." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs syncing and divergent attention kernels under torch.compile are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor import failure and HybridCache import failure are separate export issues." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FP8 accuracy regression and FP8 save_pretrained failure are distinct problems." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output mismatch and image-specific infinite loop are not the same bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Mostly conservative, but the summary overstates heterogeneity by saying the only clear duplicate-style pair is Qwen3ForCausalLM VRAM leak. The packet also contains another obvious duplicate-style pair with the same title, issue:45310 and issue:45478, both about the Qwen3.5 MoE from_pretrained error. That omission makes the summary a bit too strong. The pairwise rejections themselves look generally conservative, and the 42673/43704 acceptance is plausible." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 35, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43479", + "issue:43504", + "issue:43526", + "issue:43673", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44079", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44373", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44964", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45561", + "issue:45712", + "issue:45812" + ], + "soft_pairs": [ + "issue:42222|issue:45200", + "issue:44222|issue:45310", + "issue:42994|issue:43824", + "issue:36010|issue:41628", + "issue:43232|issue:44792", + "issue:42175|issue:45712", + "issue:43377|issue:44242", + "issue:43673|issue:44188", + "issue:44038|issue:45310", + "issue:43526|issue:45561", + "issue:44496|issue:45561", + "issue:44964|issue:45137", + "issue:43504|issue:45561", + "issue:42898|issue:44568", + "issue:42175|issue:44933", + "issue:43479|issue:44869", + "issue:30990|issue:35141", + "issue:43976|issue:44933", + "issue:43479|issue:44079", + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:44188|issue:44373", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:44734|issue:45103", + "issue:30333|issue:33290", + "issue:43723|issue:44568", + "issue:42898|issue:45356", + "issue:44060|issue:45125", + "issue:42898|issue:44779", + "issue:43874|issue:45092" + ], + "prepared_review_unit_hash": "c60c4989e89e544eb8fd3c3a564fe251961f1637a593ab92a32bc0dbf718e6b9", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "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": "The cluster is highly heterogeneous: a few items are tokenizer/model-loading regressions, but most are unrelated bugs across training, multimodal, generation, packaging, and test infrastructure. I would not collapse any soft pair as a duplicate, and there is no safe single canonical issue for the whole cluster.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different failures: vitpose model breakage vs Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both mention MoE/FP8, but one is save_pretrained and the other is from_pretrained loading; different code paths." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized model saving and a missing Qwen2.5-VL import are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different missing symbols from different modules: GenerationMixin vs AutoImageProcessor." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation cache/sync_gpus bug vs a janus image-generation test failure; not the same defect." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Packaging/backend dependency issue vs leftover dummy classes in repo checks." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask batching bug vs router-logits load-balancing loss omission." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Both involve generation behavior, but one is cache/chunked_prefill and the other is attention kernel divergence under torch.compile." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Different Qwen-related model compatibility errors; not the same concrete bug." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs pytest-xdist file race." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrecognized model/config loading vs test utility file contention." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Legacy BEiT preset loading bug vs pytest-xdist race; no common code path." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Tokenization cleanup behavior change vs mdeberta BOS/EOS token insertion; related area but different bug." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "TensorFlow backend packaging vs missing image_utils import are distinct issues." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Default multimodal config initialization vs Whisper word timestamp decode crash." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence-Transformers loading hang vs embedding reinit on resize; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility issue vs missing import from image_utils." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config-default bug vs ModelOutput key assignment bug; different objects and code paths." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions with different models and symptoms; not one underlying bug." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import NameError vs Qwen2.5-VL import error; different missing imports." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Attention-kernel divergence vs docstring typo; not a duplicate." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in evaluation vs DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "processor.save_pretrained file omission vs Qwen3.5 Moe from_pretrained failure; different fixes." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation crash in serve vs auto_docstring annotations crash; separate subsystems." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow job status reporting vs DeepSpeed/Adafactor OOM; unrelated training issues." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading/import problem vs BOS/EOS token behavior change; different failure modes." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is cleanup-space behavior and the other is Kimi-K2.5 codec/regex handling." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Incorrect tied-weights warning vs missing _tp_plan for tensor parallelism; different model-internal bugs." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions: cleanup-space behavior vs DeepSeek token IDs." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image patch-count method vs old InternVL2 meta-init incompatibility; 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": 2, + "component_count": 17, + "review_unit_index": 36, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43232", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43824", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44448", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44568", + "issue:44623", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44977", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45310", + "issue:45357", + "issue:45405", + "issue:45406", + "issue:45656", + "issue:45712", + "issue:45834" + ], + "soft_pairs": [ + "issue:45405|issue:45712", + "issue:42898|issue:44448", + "issue:44933|issue:45712", + "issue:44295|issue:45656", + "issue:44222|issue:44297", + "issue:44297|issue:45656", + "issue:43479|issue:43644", + "issue:44206|issue:45103", + "issue:44222|issue:44623", + "issue:42831|issue:44623", + "issue:30333|issue:36331", + "issue:22355|issue:28282", + "issue:41628|issue:42757", + "issue:30333|issue:34689", + "issue:43232|issue:45834", + "issue:43723|issue:44977", + "issue:28282|issue:30333", + "issue:30333|issue:35141", + "issue:44222|issue:44484", + "issue:44297|issue:45310", + "issue:44263|issue:45092", + "issue:44263|issue:44360", + "issue:44483|issue:45406", + "issue:44514|issue:45103", + "issue:44373|issue:44485", + "issue:44295|issue:44623", + "issue:43723|issue:44871", + "issue:44261|issue:45092", + "issue:43723|issue:44991", + "issue:45092|issue:45357", + "issue:30990|issue:38175", + "issue:43723|issue:44442", + "issue:41628|issue:43824", + "issue:39401|issue:44568", + "issue:43479|issue:44913", + "issue:42831|issue:45310" + ], + "prepared_review_unit_hash": "8bed832a48df8e6be74bb8719648cf2095381fb441721627d94b3c0992eedb61", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38260, + "estimated_input_tokens": 9565, + "estimated_eval_tokens": 19386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: it mixes unrelated import errors, tokenizer/model regressions, serialization bugs, training issues, and docs-style complaints. No PRs are present, and none of the soft pairs look like the same concrete underlying bug/change.", + "confidence": 0.73, + "canonical_issue_reason": "issue:44297 is the most self-contained, clearly reproducible bug in the set, with a narrow serialization mismatch (`tokenizer_class` in `tokenizer_config.json`) and a straightforward fix path.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44297 is the strongest representative issue because it is concrete, isolated, and easy to validate compared with the broader regression or discussion-style reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Unrelated problems: a version pin to unreleased PEFT versus leftover dummy classes leaking into `dir(transformers)`." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both mention v5-era behavior changes, but one is token cleanup semantics and the other is full model output drift; not the same concrete bug." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Image-utils import/export issue versus dummy PT object cleanup; different code paths and failure modes." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Registering/reading `position_ids` as a buffer is unrelated to a deepspeed double optimizer step." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 `save_pretrained` behavior for MoE models is a different serialization bug than tokenizer config mismatch." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save/load metadata mismatch is unrelated to optimizer stepping under deepspeed." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Different configuration bugs: default vision/audio config initialization vs non-persistent buffer junk values." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Unsupported `center` arg in `LasrFeatureExtractor` is unrelated to the auto-docstring annotations crash." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "MoE FP8 save bug and processor save_pretrained missing files are both save-related, but not the same artifact or code path." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy regression and processor save_pretrained file omission are distinct issues." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting on failed jobs is unrelated to `CustomTrainer.compute_loss` signature breakage." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Import errors, but one is missing `transformers.onnx` and the other is missing PyTorch; different root causes." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Missing `AutoImageProcessor` export is unrelated to `huggingface_hub.is_offline_mode` import failure." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Training-job MLflow reporting and Llama 3.2 model loading regression do not share the same underlying bug." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs update bug and Kosmos2.5 OCR index error are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading in v5 and Qwen3.5 flash-attention generation problems are different code paths and symptoms." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "PyTorch-not-found import error versus MLflow job state bug; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLflow reporting issue and embedding resize/post_init reinitialization bug are different subsystems." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained bug is unrelated to the `max_shard_size` default question." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer config mismatch is not the same as Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "`torch.split()` return-value handling in GlmMoeDsaIndexer is unrelated to InternVL2 meta-initialization compatibility." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same DSA/indexer area, but one reports split return values and the other a missing ReLU; not the same bug." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat completion endpoint rejection and Gemma4Processor missing `_tokenizer` are different failures." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Batched `apply_chat_template` crash and auto-docstring annotation crash are unrelated." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring mismatch for `position_ids` is not the same as GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer registration/read error and processor save_pretrained missing files are separate serialization/runtime issues." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading failure in v5 is unrelated to Gemma-3 `eos_token_id` inconsistency." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "RMS norm epsilon precision discussion and InternVL2 meta-init incompatibility are different model bugs." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but one is generic v5 loading and the other is a specific model; no evidence they share the same concrete cause." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Old InternVL2 meta-init incompatibility is distinct from Qwen3.5 visual encoder key saving regression." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading stuck and SigLIP2 zero probabilities are different model/runtime problems." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "Generic tokenizer loading regression is unrelated to FastSpeech2ConformerTokenizer AutoTokenizer failure." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Different import/export problems: missing `AutoImageProcessor` versus missing `Qwen2_5_VLForConditionalGeneration`." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Qwen3 offset mapping bug and mDeBERTa special-token behavior regression are unrelated tokenizer issues." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Config initialization for multimodal defaults is unrelated to GPTNeoX `rotary_pct` persistence on reload." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy regression and Qwen3.5 MoE from_pretrained failure are different problem types." + } + ] + }, + "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": 17, + "review_unit_index": 37, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:22355", + "issue:28218", + "issue:29127", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43576", + "issue:43756", + "issue:43784", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44829", + "issue:44945", + "issue:45200", + "issue:45278", + "issue:45290", + "issue:45593", + "issue:45600", + "issue:45753" + ], + "soft_pairs": [ + "issue:42886|issue:42994", + "issue:42757|issue:45278", + "issue:44263|issue:44373", + "issue:43756|issue:44060", + "issue:42831|issue:44038", + "issue:43232|issue:45593", + "issue:43576|issue:43784", + "issue:42831|issue:45753", + "issue:42886|issue:43066", + "issue:22355|issue:28218", + "issue:42371|issue:44263", + "issue:43122|issue:44521", + "issue:44117|issue:44704", + "issue:30990|issue:33357", + "issue:43232|issue:44743", + "issue:42907|issue:44945", + "issue:41628|issue:45278", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:39401|issue:44488", + "issue:44295|issue:44297", + "issue:43784|issue:45600", + "issue:43122|issue:44514", + "issue:42907|issue:43937", + "issue:43408|issue:43475", + "issue:44360|issue:44485", + "issue:44261|issue:44373", + "issue:39401|issue:44521", + "issue:44625|issue:44829", + "issue:43116|issue:45200", + "issue:29127|issue:30990" + ], + "prepared_review_unit_hash": "85fbc7435da8d72d5977ce784cb167d1fe0a76118fb68abaa777baeb5cd46abe", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36541, + "estimated_input_tokens": 9136, + "estimated_eval_tokens": 18528 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a mixed set of unrelated Transformers issues. The soft pairs share broad themes like tokenization, import regressions, or generation behavior, but they point to different concrete bugs and code paths, so none should be deduplicated.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous; if one representative is needed, issue:42886 is the clearest concrete regression report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42886 is a strong representative because it has a specific offline/cache failure mode, version context, and reproducible behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading vs quantized model saving are different failures." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Both mention imports, but one is a specific missing symbol and the other is a broad v5 upgrade regression." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "A torch.split indexing bug and a docstring issue are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different model-specific bugs: RoPE layer count vs tied-weights warning." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Accuracy regression in FineGrainedFP8 is unrelated to a Qwen3-VL-Moe bug." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation cache/state update bug vs missing auxiliary losses in D-FINE are separate issues." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "CLI/env command breakage is unrelated to a sentence-transformers import NameError." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "Different symptoms and models: FP8 accuracy issue vs Qwen3_5 infinite loop on one image." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading and tokenizer decoder type mismatch are different tokenizer problems." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing module import and extra whitespace in tokenization are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration guidance does not match a DSA indexer return-value bug." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer output drift across versions is separate from multimodal assistant mask handling." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping None handling and AutoProcessor kwarg forwarding are different code paths." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang for Sentence-Transformers and a MacOS bus error in CLIP are distinct failures." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Generation kwargs update after sync_gpus is unrelated to recurrent-state reset in qwen3_5." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized model save failure and pipeline-parallelism output corruption are different bugs." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Cannot import one symbol is not the same as multiple import errors after a version update." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility and pipeline-parallelism output are unrelated." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Cached-file kwargs forwarding and chat-template crashes with tool calls are separate behaviors." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset mapping for Qwen3 tokenizer is unrelated to loading a specific model failing." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Buffer read error on position_ids and tokenizer_class mismatch on save_pretrained are different bugs." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Import NameError in sentence-transformers is unrelated to stale Sam3LiteText references in auto_mappings." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer behavior change and batched chat-template crash are not the same defect." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Same general area, but different failure modes and fixes." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Saving dequantized models and GenerationConfig validation are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning and missing attribute on vision encoder output are different SAM3 issues." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "A missing ReLU in DSA indexer and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "Precision/config issue in MLA q_a_layernorm and a docstring correction are not duplicates." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping bug and multimodal assistant mask bug affect different paths." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation and flash_attention_3 training degeneration are distinct issues." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Sentence-Transformers loading stall 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": 17, + "review_unit_index": 38, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:33453", + "issue:36296", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43064", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43452", + "issue:43493", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43825", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44485", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44945", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45753" + ], + "soft_pairs": [ + "issue:43122|issue:44779", + "issue:42907|issue:44964", + "issue:44038|issue:44297", + "issue:44038|issue:44295", + "issue:42907|issue:43452", + "issue:36296|issue:43116", + "issue:43824|issue:44351", + "issue:44038|issue:44222", + "issue:42757|issue:44351", + "issue:45216|issue:45478", + "issue:43066|issue:44779", + "issue:43278|issue:44945", + "issue:43493|issue:45753", + "issue:43122|issue:44977", + "issue:42907|issue:44222", + "issue:43992|issue:44704", + "issue:43784|issue:45446", + "issue:38617|issue:45278", + "issue:42673|issue:45072", + "issue:43937|issue:44485", + "issue:22355|issue:29127", + "issue:42222|issue:43441", + "issue:42371|issue:43874", + "issue:43064|issue:43673", + "issue:43976|issue:45446", + "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:43116|issue:43825", + "issue:44038|issue:45478", + "issue:44222|issue:45478", + "issue:45083|issue:45446", + "issue:43937|issue:45563", + "issue:39401|issue:44779", + "issue:33453|issue:39401" + ], + "prepared_review_unit_hash": "50d59ce4c42ba72670b847d6091f0120c89dc63c13e0acc17574da7e83f326ad", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37377, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of Transformers v5 regressions (imports, tokenizers, save/load, model-specific bugs) rather than one shared duplicate bug. None of the soft edges look like the same concrete issue, so they should all be rejected. The broadest representative issue is the general v5 import-regression report.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45278 is the most umbrella-like report and best represents the broad v5 breakage theme across the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 has the widest scope and is the most suitable cluster representative for triage, even though the underlying reports are not duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they describe different models and failure details; not clearly the same bug." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Different components and symptoms: dequantized save failure vs model load failure." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug vs tokenizer save_pretrained metadata mismatch; different code paths." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Different failures: model-specific v5 breakage vs position_ids buffer access." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving dequantized weights and gguf loading are unrelated behaviors." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug is unrelated to the classification example script issue." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different packages." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "One is Qwen3-VL-Moe runtime breakage; the other is FP8 save_pretrained on moe models." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different modules; no shared underlying defect." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5 v5 regressions, but one is save_pretrained checkpoint corruption and the other is from_pretrained load failure; not clearly the same bug." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong decoder type and wrong tokenization are related broadly, but not the same concrete tokenizer bug." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype mismatch in eval is unrelated to pipeline-parallel wrong output." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and Qwen3_5 infinite loop are different model-specific problems." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Both involve tokenization/flash-attention regressions, but they affect different models and behaviors." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Saving dequantized models and FP8 save_pretrained are separate serialization issues." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens loading and cached_file kwargs forwarding are unrelated load-path bugs." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "Sentence-transformers import breakage and an AuxRequest version-check bug are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing symbol import and broad import-regression umbrella are not the same underlying issue." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM and dtype mismatch CI failure are different bugs." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig validation error and RoPE implementation discussion are different GLM-5 issues." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "ONNX import failure and LayoutLMv3 error-message clarity are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose breakage and Ministral FlashAttention failure are different model-specific regressions." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 control warning and missing image-patch helper are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Optimizer state corruption in trainer/FSDP2 is unrelated to chunked_prefill cache missing." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility issue and AuxRequest import version check are different problems." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Classification script behavior and SigLIP2 implementation mismatch are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer SGD args not used is unrelated to trainer.evaluate() failing after train()." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example script multi-label behavior and pipeline translation warning are different features." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5/Qwen3-Moe on v5, but one is Qwen3-VL-Moe and the other is a from_pretrained issue; not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained moe and Qwen3.5 from_pretrained error are distinct serialization/load failures." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Helper output-length logic in qwen3_omni_moe is unrelated to the PyTorch version-check import bug." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig invalid and stale num_return_sequences warning are different generation issues." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but for different model families and symptom types." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "General tokenizer loading regression and Qwen3 offset_mapping bug are not the same concrete failure." + } + ] + }, + "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": 17, + "review_unit_index": 39, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43450", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43827", + "issue:43906", + "issue:43937", + "issue:43986", + "issue:44075", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44291", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45464", + "issue:45542", + "issue:45561", + "issue:45588", + "issue:45656" + ], + "soft_pairs": [ + "issue:44117|issue:45588", + "issue:43827|issue:45446", + "issue:43937|issue:44977", + "issue:42673|issue:43906", + "issue:45405|issue:45446", + "issue:44291|issue:45071", + "issue:43450|issue:43986", + "issue:45341|issue:45561", + "issue:44117|issue:45081", + "issue:38617|issue:41628", + "issue:39401|issue:43421", + "issue:44410|issue:45310", + "issue:44829|issue:45216", + "issue:42673|issue:43582", + "issue:44484|issue:45310", + "issue:42371|issue:44261", + "issue:42222|issue:45542", + "issue:43408|issue:43742", + "issue:44933|issue:45278", + "issue:42617|issue:45341", + "issue:42673|issue:43577", + "issue:42175|issue:42222", + "issue:43937|issue:44496", + "issue:38617|issue:44220", + "issue:44075|issue:45656", + "issue:44483|issue:45464", + "issue:42371|issue:42831", + "issue:43066|issue:43122", + "issue:43012|issue:43577", + "issue:43673|issue:45341" + ], + "prepared_review_unit_hash": "2a48e2bca3bf7f67f6a7d826e1f208d7fe8e1caafcbc738d591c25e168d28163", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36481, + "estimated_input_tokens": 9121, + "estimated_eval_tokens": 18498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft pairs mostly share only broad subsystem labels (tokenizers, import errors, loading, optimizer, chat API) but not the same concrete bug. I would not merge any of the proposed soft links.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45278 is the best umbrella-style issue here because it is the broadest open report about a wave of import regressions after a version update, but it still does not subsume the specific failures in the soft pairs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most suitable representative issue for this set because it is broad, recent, and open, but the rest of the items are too diverse to treat as true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping fallback vs flash-attention crash on sink-less models; unrelated code paths." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs cleanup about pipeline removals vs a flex_attention import/version check bug." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Different failures: invalid GenerationConfig for GLM-5 vs flash-attention generation bug for Qwen3.5." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 VRAM leak in multithreaded dataloading is not the same as an isolated reproduction of another issue." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version pinning issue vs PyTorch version gating bug; unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights argument mismatch vs PreTrainedConfig type-check regression; different surfaces." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video shape bug vs missing-torchvision crash; both video-related but not the same defect." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Generic testing_utils bug vs xdist race on captured_info.txt; insufficiently specific to be the same bug." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "None-returning tokenizer mapping and Mistral regex patch crash are distinct tokenizer regressions." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Different import errors for different symbols; same broad theme, not the same missing-export bug." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Offset mapping bug vs runtime post-processor update bug; separate tokenizer behaviors." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Layer projections missing in qwen3next vs Qwen3.5 from_pretrained load error; different model-loading problems." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Flash-attention training degeneration vs save_pretrained checkpoint regression; unrelated failures." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in Qwen3 threading vs Apple Silicon caching_allocator_warmup TypeError; different bugs." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "save_pretrained shard-size question vs Qwen3.5 MoE load error; not the same issue." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 configuration warning vs missing rms_norm_eps precision error; both numeric-settings related but not the same defect." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Broken VitPose model loading vs missing TensorFlow backend when only tensorboard is installed; unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Model-type mismatch warning vs MobileLLM key error; different loading problems." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing image_utils import is not enough to call it the same as a broad 'many import errors' report." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "3d_parallel.py runtime failure vs a testing_utils bug; different files and behaviors." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "Qwen3 threading VRAM leak vs BLIP2 dtype staying float32; separate model/runtime issues." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Dependency installation/backend packaging issue vs VitPose model breakage; not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig for a known model vs unrecognized model_type in config.json; different config-loading errors." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for a missing symbol vs audio feature extraction failure; unrelated." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Unused SGD args and double optimizer-step under deepspeed are different optimizer bugs." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Chat/completions request rejection vs streaming inference failure on Qwen3.5-0.8B; different API breakages." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior control and FineGrainedFP8 accuracy issues are both precision-related but not the same change or bug." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Wrong decoder type and changed tokenization output are related to v5 tokenizer behavior but not clearly the same underlying bug." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "PyTorch warning during bfloat16 compilation vs BLIP2 dtype not propagating; distinct dtype-handling issues." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill vs a testing_utils bug; completely different areas." + } + ] + }, + "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": 17, + "review_unit_index": 40, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:38617", + "issue:41628", + "issue:42222", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43335", + "issue:43408", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43824", + "issue:43873", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44368", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44704", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45103", + "issue:45245", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45459", + "issue:45478", + "issue:45542", + "issue:45588" + ], + "soft_pairs": [ + "issue:44410|issue:45478", + "issue:44484|issue:45478", + "issue:43097|issue:44493", + "issue:43874|issue:44261", + "issue:44263|issue:44485", + "issue:42222|issue:45459", + "issue:43012|issue:43408", + "issue:41628|issue:45071", + "issue:43441|issue:44829", + "issue:45071|issue:45542", + "issue:43931|issue:44410", + "issue:45003|issue:45341", + "issue:43931|issue:44038", + "issue:44117|issue:44987", + "issue:43097|issue:43452", + "issue:43335|issue:45071", + "issue:43986|issue:45341", + "issue:43408|issue:44368", + "issue:43986|issue:44484", + "issue:43874|issue:44263", + "issue:43976|issue:45071", + "issue:43531|issue:45341", + "issue:44829|issue:45245", + "issue:43986|issue:45372", + "issue:38617|issue:42757", + "issue:44261|issue:44485", + "issue:43064|issue:44315", + "issue:44410|issue:45375", + "issue:38617|issue:44351", + "issue:43097|issue:45588", + "issue:44062|issue:44117", + "issue:44038|issue:44410", + "issue:44704|issue:45103", + "issue:43502|issue:43992", + "issue:43097|issue:43873", + "issue:43986|issue:44117", + "issue:43824|issue:43986", + "issue:43066|issue:43906" + ], + "prepared_review_unit_hash": "f3b62b55831eac313479b29eb450e0c9d09e41e1e640d10e76c6d6072b220ec0", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 39090, + "estimated_input_tokens": 9773, + "estimated_eval_tokens": 19802 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is thematically similar but not duplicate-tight: the items span unrelated import errors, model-loading regressions, config/runtime warnings, and training issues. None of the soft pairs appear to describe the same underlying bug or change.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45478 is the best representative of the cluster's recurring Transformers v5 model-loading regressions: it is concrete, versioned, and actionable.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45478 is the strongest overall issue candidate because it is specific, clearly reproducible, and closest to the cluster's main loading/regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen-family problems: missing layer projections vs a from_pretrained failure in Qwen3.5 MoE." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Unrelated topics: shard-size documentation/question vs Qwen3.5 MoE loading error." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Different regressions: removal of tie_embeddings_and_encoder_decoder vs unexpected position-id keys." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different subsystems: image-patch counting AttributeError vs MLA rms_norm_eps precision issue." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Both GLM-related, but distinct code paths: torch.split behavior vs RoPE implementation." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "Different bug classes: broken vitpose models vs tokenizer error masking when protobuf is absent." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "A warning during bf16 compilation is unrelated to a model-type mismatch warning." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Importing AutoImageProcessor is unrelated to PretrainedConfig type-checking changes." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Different FlashAttention failures in different models/tasks; not the same concrete bug." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "PretrainedConfig typing and TensorFlow/tensorboard backend detection are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Different Qwen model failures: weight-shape mismatch vs missing attention projections." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to a small testing_utils bug." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Both involve Qwen3-VL, but one is a weight mismatch and the other is a broader version/MoE incompatibility." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping returning None is not the same as failing to load a specific model repo." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removal of a config field is unrelated to gguf_file breaking AutoTokenizer/AutoModelForCausalLM loading." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation is unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "A processor crash without torchvision is unrelated to a testing_utils bug." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "These are different warning conditions with different model/config causes." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Model loading without torchvision is unrelated to a max_shard_size question." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM-related failures: missing image-patch method vs torch.split return handling." + }, + { + "left": "issue:43976", + "right": "issue:45071", + "accept": false, + "reason": "Python version compatibility problem is unrelated to config type checking." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to a testing_utils bug." + }, + { + "left": "issue:44829", + "right": "issue:45245", + "accept": false, + "reason": "FlashAttention training degradation is unrelated to the categorical cardinality runtime error." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision crash is unrelated to the ReasoningEffort import regression." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they come from different packages and symbols." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Different GLM issues: layernorm epsilon precision vs RoPE implementation." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to Liger Kernel not being applied." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing attention projections in qwen3next is unrelated to a missing config field in Qwen3_5MoeVisionConfig." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different import failures from different modules and symbols." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Config-field removal is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer AddedToken argument duplication is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both touch Qwen variants, but one is a broader loading bug and the other is a missing-layer-0 projection issue." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Passing kwargs to cached_file is unrelated to auto_docstring crashes under future annotations." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only network leakage and missing embed_tokens.weight are separate loading bugs." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Embedding-tie removal and offloading/quantization behavior are not the same bug." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision and tokenizer mapping None are different failure modes." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Qwen2.5-VL import failure and video-model processor crash are distinct model-loading problems." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Tokenizer decoder type regression is unrelated to the isolated reproduction of another issue." + } + ] + }, + "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": 17, + "review_unit_index": 41, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43381", + "issue:43475", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43825", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44077", + "issue:44220", + "issue:44230", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44938", + "issue:45216", + "issue:45290", + "issue:45381", + "issue:45479", + "issue:45753" + ], + "soft_pairs": [ + "issue:43097|issue:43901", + "issue:43825|issue:44655", + "issue:43992|issue:44484", + "issue:43352|issue:44380", + "issue:43208|issue:43550", + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:42886|issue:43381", + "issue:42907|issue:45753", + "issue:42907|issue:44655", + "issue:44380|issue:45479", + "issue:43352|issue:43976", + "issue:43064|issue:44380", + "issue:43208|issue:43526", + "issue:36683|issue:43582", + "issue:44483|issue:44514", + "issue:43012|issue:44683", + "issue:43618|issue:44380", + "issue:42757|issue:43824", + "issue:43618|issue:43761", + "issue:38617|issue:43502", + "issue:44038|issue:44230", + "issue:44483|issue:45290", + "issue:43756|issue:44485", + "issue:43299|issue:43867", + "issue:44483|issue:45381", + "issue:43299|issue:45753", + "issue:43901|issue:44655", + "issue:36683|issue:43475", + "issue:42994|issue:44655", + "issue:43867|issue:45216", + "issue:43352|issue:43937", + "issue:44077|issue:44380", + "issue:36010|issue:42757", + "issue:43066|issue:45381", + "issue:36683|issue:44077", + "issue:44220|issue:44655", + "issue:43976|issue:44938" + ], + "prepared_review_unit_hash": "fd0e67524222420abf982f743f8d4a9a103830c10ddbbc45e73057dd0957da18", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 38790, + "estimated_input_tokens": 9698, + "estimated_eval_tokens": 19652 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most items are distinct Transformers regressions or feature/compatibility issues that only share broad theme words like v5, loading, saving, or attention. None of the soft edges look like true duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "issue:44483 is the broadest user-facing v5 regression in the set, but the cluster is too mixed to treat it as a true duplicate canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue must be chosen, issue:44483 is the most generic and broadly impactful, though still not a mergeable umbrella for the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43097", + "right": "issue:43901", + "accept": false, + "reason": "Removed API warning/deprecation vs docs mismatch; same release era, but not the same bug." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Chat template/task support error vs pipeline serialization; different code paths and symptoms." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing model weight on load vs save_pretrained shard sizing question; unrelated save/load problems." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Flash Attention support limitation vs GPT2 attention scaling being ignored; different attention bugs." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers vs Bamba torch.compile/SDPA failure; different models and root causes." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Precision warning during compile vs actual compile failure under SDPA; not the same defect." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "bfloat16 compile warning vs tie_word_embeddings warning; both warnings, but different code paths." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Offline cache loading failure vs gradient checkpointing in eval mode; unrelated behaviors." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Quantized/dequantized save failure vs image-generation infinite loop; different models and failure modes." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Dequantized model saving vs pipeline object serialization; both saving-related, but not the same bug." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention backend scaling bug vs degenerate single-label classification loss; unrelated subsystems." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Missing Flash Attention support vs Python version compatibility failure; no shared root cause." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2 optimizer-state corruption vs attention scaling ignored; different training vs inference issues." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; unrelated components." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config attribute missing vs Apple Silicon TypeError in allocator warmup; unrelated errors." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection vs batched apply_chat_template crash; both v5-facing, but different APIs and code paths." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Compile warning vs compiled flex_attention failure on newer torch; different failure severity and cause." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIPOutput attentions not assigned vs GPT2 attention scaling ignored; both mention attention, but not the same bug." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Missing huggingface_hub symbol vs missing Transformers export; both import errors, but for different modules/symbols." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Missing attentions output vs hidden_states=None regression; same model family, different output fields and fixes." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Missing internal import vs unexpected network access when local_files_only=True; different offline-related problems." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Loading bug in Qwen3-VL-Moe vs fp8 support request for Qwen3-VL/Qwen3.5 MoE; bug vs feature request." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Chat/completions endpoint rejection vs assistant tool-call template crash; different layers of the chat stack." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE layer drop bug vs GLM-5 RoPE discussion; both RoPE-themed, but not the same issue." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL-Moe loading break vs sorted-state_dict load error; both load failures, but different triggers and models." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat/completions request handling vs Qwen2.5-VL video position-ids bug; unrelated failures." + }, + { + "left": "issue:43299", + "right": "issue:45753", + "accept": false, + "reason": "Qwen3VL-Moe loading regression vs Qwen3.5 image infinite loop; different model families and symptoms." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mention stale return_all_scores vs pipeline save_pretrained failure; documentation drift is not the same as serialization bug." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Missing Gemma3Config attribute vs missing SAM3 vision encoder attribute; both are attribute errors, but unrelated models." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model save failure vs pipeline object save_pretrained; both save-related, but not the same code path." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Sorted state_dict load issue vs Qwen3.5 checkpoint regression in save_pretrained; different save/load directions and root causes." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support limitation vs invalid GenerationConfig for GLM-5; unrelated model-specific problems." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init policy issue vs GPT2 attention scaling bug; no shared underlying defect." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different modules; both are import errors, but not the same broken symbol." + }, + { + "left": "issue:43066", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer decoder-type regression vs Qwen2.5-VL video position_ids bug; separate v5 regressions." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3 config missing vocab_size vs patchtsmixer post_init validation; unrelated model/config issues." + }, + { + "left": "issue:44220", + "right": "issue:44655", + "accept": false, + "reason": "Audio feature extraction bug vs pipeline save_pretrained serialization; different subsystems entirely." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Different Python compatibility failures (3.9/3.10 vs 3.14); same general theme, 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": 2, + "component_count": 17, + "review_unit_index": 42, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43352", + "issue:43502", + "issue:43525", + "issue:43575", + "issue:43618", + "issue:43867", + "issue:43931", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44188", + "issue:44242", + "issue:44246", + "issue:44263", + "issue:44355", + "issue:44360", + "issue:44380", + "issue:44484", + "issue:44492", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45593", + "issue:45600", + "issue:45850" + ], + "soft_pairs": [ + "issue:44263|issue:45440", + "issue:36683|issue:43525", + "issue:43352|issue:45593", + "issue:43618|issue:45216", + "issue:44188|issue:44380", + "issue:43064|issue:44492", + "issue:36683|issue:44242", + "issue:43352|issue:43931", + "issue:43010|issue:45278", + "issue:43992|issue:45310", + "issue:43867|issue:45375", + "issue:41669|issue:44246", + "issue:44355|issue:44855", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:44938|issue:45071", + "issue:43867|issue:45357", + "issue:43010|issue:45850", + "issue:43992|issue:45478", + "issue:43010|issue:43976", + "issue:43618|issue:44077", + "issue:44938|issue:45381", + "issue:44360|issue:45440", + "issue:36296|issue:43575", + "issue:28282|issue:33453", + "issue:43867|issue:44484", + "issue:43618|issue:44977", + "issue:43618|issue:44857", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43502|issue:44062" + ], + "prepared_review_unit_hash": "72111cc5762b6f25e8019af46e247fd431d2e7ef024e31105664be3204d2aa37", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36293, + "estimated_input_tokens": 9074, + "estimated_eval_tokens": 18404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly same-subsystem but not the same underlying bug or change; I would reject all of them as duplicates. The broadest issue in the set is the open import-regression report (issue:45278), which is the best triage anchor, but there is no clear duplicate cluster here.", + "confidence": 0.82, + "canonical_issue_reason": "No soft pair is a strong duplicate; issue:45278 is the broadest umbrella report and the most useful stand-in canonical issue for triage.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is open, broad, and phrased as a multi-model regression after upgrade, so it is the best single issue to anchor the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "Different models and failures: a GlmMoeDsaIndexer torch.split issue vs a DeepSeekV3MoE implementation mismatch." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing-config-attribute reports, but for different models and different fields; not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated bugs: FlashAttention support gating for Nemotron vs D-FINE auxiliary loss behavior." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIPOutput attentions assignment bug vs Qwen3.5 save_pretrained checkpoint regression; different code paths." + }, + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "Both involve attention internals, but one is a torch.compile branch divergence and the other is backend scaling config handling." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Training/FSDP optimizer-state corruption is unrelated to a cache-strategy typo." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Different model-loading/config issues: missing vocab_size vs MoE load-balancing loss omission." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "Nemotron FlashAttention support limitation is not the same as Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache update no_grad decoration and broad import errors after upgrade are different regressions." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight is unrelated to Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "A sorted-state_dict loading error is not the same as a missing Qwen3_5MoeVisionConfig field." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both concern import performance, but one is a specific 'import *' cleanup issue and the other is a general slow-import report." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Generic compiled-Python-file errors are not the same as a Python 3.13 @torch.jit.script parsing/IndentationError." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "A missing config attribute in Gemma3 is unrelated to FSDP2/PEFT optimizer-state corruption." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "A missing set_seed in a model test is not the same as a pytest-xdist file-write race." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are different compatibility issues." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Both are model-loading/saving regressions, but they affect different models and different serialization paths." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A specific cache-method change is not the same as a vague report that 5.0 breaks too many models." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different models and symptoms: UMT5 weight omission vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache no_grad decoration and Python version support for Transformers 5.1.0 are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions assignment and PatchTSMixer post_init allowance are unrelated model API issues." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Python 3.14 load failure is not the same as qwen2.5-vl video vision_position_ids being wrong." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "A missing ReLU in a DSA indexer is unrelated to DeepSeekV3MoE divergence from the remote implementation." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and TP-related OOM on model load are different failure modes and likely different causes." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-not-found ImportError and tokenizer-loading regression are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "Model-loading failure is unrelated to the default max_shard_size question." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Different affected components: CLIPOutput attentions vs LwDetrImageLoss AMP crash." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "A flaky test due to missing seeds is unrelated to loading Phi-4 multimodal with latest transformers." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "A seed-related test flake is not the same as a GraniteSpeech stale device override CI failure." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "A flaky integration test is unrelated to auto_mappings.py referencing removed Sam3LiteText configs." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Local-files-only network leakage and AddedToken keyword duplication are different tokenizer/loading bugs." + } + ] + }, + "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": 17, + "review_unit_index": 43, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:42890", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43723", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44075", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45341", + "issue:45375", + "issue:45468", + "issue:45478" + ], + "soft_pairs": [ + "issue:43723|issue:44556", + "issue:42548|issue:44351", + "issue:43010|issue:44075", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:42890|issue:45341", + "issue:43010|issue:43065", + "issue:44355|issue:44749", + "issue:43502|issue:44485", + "issue:45125|issue:45478", + "issue:30990|issue:31795", + "issue:43010|issue:43232", + "issue:45125|issue:45375", + "issue:30333|issue:33453", + "issue:43502|issue:44222", + "issue:43299|issue:44556", + "issue:44315|issue:44360", + "issue:43502|issue:44961", + "issue:42831|issue:43874", + "issue:30333|issue:36296", + "issue:44355|issue:45071", + "issue:44222|issue:44230", + "issue:44355|issue:44655", + "issue:43827|issue:44961", + "issue:33453|issue:34689", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:41628|issue:42548", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:38617|issue:43576", + "issue:43673|issue:44492", + "issue:43576|issue:44938", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:41628|issue:44355", + "issue:36296|issue:43526" + ], + "prepared_review_unit_hash": "a3b9149a31929f5810e68e6fed853491a05f05de43134eab493b1646a8d60a95", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 40, + "serialized_chars": 38441, + "estimated_input_tokens": 9611, + "estimated_eval_tokens": 19478 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly noisy: many issues share Transformer/v5/loading/import wording but appear to describe different bugs. I only see one soft edge that plausibly points to the same underlying model-loading problem.", + "confidence": 0.58, + "canonical_issue_reason": "issue:43576 is the broadest and most central v5 regression report in the set, with inbound references and cluster-wide similarity to other v5 breakage reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43576 is the best representative issue because it is the most umbrella-like and high-signal among otherwise mostly unrelated regression reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading failures, but one is tokenizer loading and the other is checkpoint reload after upgrade; different concrete bugs." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different exported symbols and likely different missing re-exports." + }, + { + "left": "issue:43010", + "right": "issue:44075", + "accept": false, + "reason": "Cache update/no_grad and SGD args not used are unrelated code paths and symptoms." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 OOM from reserved memory and checkpoint reload failure are unrelated regression types." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained bug and missing tensor-parallel plan for Qwen3.5 Moe are different problems." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "Both touch testing, but one is flaky integration tests and the other is a generic testing_utils bug; not the same issue." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache update semantics and a dummy Conv2d in Sam3PixelDecoder are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors and a data-loading slowdown after upgrade are different issues." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only network leakage and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": true, + "reason": "Both point to the same Qwen3.5 Moe loading regression around from_pretrained / tensor-parallel support in v5.4.x." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Stuck loading Sentence Transformers and docs confusion for model.forward are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Different generation/cache bugs with different failing code paths." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Tensor-parallel Qwen3.5 Moe loading and missing deepstack_visual_indexes in vision config are different config issues." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow reporting failure and tokenizer-loading regression are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Network request suppression and FP8 save_pretrained are different bugs." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Qwen3VL Moe loading regression and checkpoint reload after version upgrade are not the same concrete failure." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application and DSA indexer missing ReLU are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "local_files_only behavior and an unrelated issue titled 'racoon' do not match." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FP8 accuracy and missing image-patch count method are different defects." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow failure reporting and tensor-parallel training bug are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled file loading failure and PretrainedConfig type-checking regression are different problems." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained and FP8 inference support are related by theme, but not the same bug or code path." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled-file errors and Pipeline.save_pretrained failure are separate save/load issues." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() and an unrelated issue titled 'racoon' are not duplicates." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer-loading regression and Llama 3.2 vision model loading breakage are different model families and failures." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing huggingface_hub import and broken transformers env command are different surface areas." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression and tensor-parallel training bug are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env CLI and PretrainedConfig type checking are different regressions." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Env command breakage and missing Qwen2_5_VLForConditionalGeneration import are different import/CLI problems." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "AutoImageProcessor and PreTrainedModel import failures are different missing exports." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion and missing layer_type_validation import are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No meaningful relation between the two reports." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "Missing configuration_utils import and broken env command are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing cache during chunked_prefill and a typo in cache strategies are different issues." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env command and Python 3.14 load failure are not the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug and unrelated issue titled 'racoon' do not match." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and unrelated 'racoon' issue are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug and TextClassificationPipeline docs mismatch are different." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "AutoImageProcessor import error and compiled-file loading errors are different problems." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training bug and BeitImageProcessorFast reduce_labels bug are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is cautious and stays within what the packet supports: issue:43576 does have inbound references and several similarity links, and the soft-edge verdicts are generally conservative given only titles/body metadata. The one accepted Qwen3.5 MoE pair also looks plausibly like the same bug family." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 44, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43827", + "issue:43874", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44246", + "issue:44261", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45468", + "issue:45478", + "issue:45542", + "issue:45753", + "issue:45850" + ], + "soft_pairs": [ + "issue:45362|issue:45850", + "issue:44261|issue:44360", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:44273|issue:45003", + "issue:44393|issue:45753", + "issue:44038|issue:45125", + "issue:43576|issue:44556", + "issue:44393|issue:45071", + "issue:43525|issue:45850", + "issue:44556|issue:45092", + "issue:41669|issue:43673", + "issue:44492|issue:45850", + "issue:43572|issue:45850", + "issue:42371|issue:44492", + "issue:45278|issue:45850", + "issue:42890|issue:44908", + "issue:44704|issue:44908", + "issue:43408|issue:44315", + "issue:44246|issue:44273", + "issue:43976|issue:44393", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:44297|issue:45230", + "issue:44908|issue:45071", + "issue:41669|issue:44351", + "issue:44295|issue:45230", + "issue:41669|issue:45341", + "issue:44492|issue:44945", + "issue:43874|issue:45850", + "issue:44393|issue:45478", + "issue:44315|issue:44829", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908", + "issue:43576|issue:44273", + "issue:44655|issue:44861", + "issue:44360|issue:45468", + "issue:44273|issue:44908", + "issue:45230|issue:45310" + ], + "prepared_review_unit_hash": "77a47171854c84aabf4b7dba102d6edbf57c68a096abc230bd058f54e6ad7bdf", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39356, + "estimated_input_tokens": 9839, + "estimated_eval_tokens": 19934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose mix of Transformers v5 regression/model-specific issues, not a tight duplicate cluster. The soft pairs mostly share broad keywords like v5, import, cache, or model names, but they describe different code paths and failure modes.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set (\u201cSince 5.0 version it breaks too many models\u201d), so it best matches the cluster\u2019s overall regression theme, even though it is very vague.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the clearest actionable representative of the main theme: upgrade-to-v5 import breakage after moving from 4.57.0 to 5.5.0.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Specific Qwen3.5 chat crash vs a broad complaint about many models breaking; not the same concrete bug." + }, + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue vs a DSA indexer ReLU issue; unrelated subsystems and symptoms." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "processor.save_pretrained missing files is concrete; the other is just a generic bug report with no matching details." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading failure vs undefined TensorFlow backend error; different features and code paths." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading behavior vs unsafe sys.modules access in modeling_utils; not the same bug." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "2D bounding-box hallucination/error vs infinite loop on a specific image; different failure mode." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Generic Qwen3-VL-Moe bug vs missing _tp_plan for tensor parallelism; related model family but different concrete issue." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Broken `transformers env` command vs checkpoint reload incompatibility after upgrading; different functionality." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issue vs PretrainedConfig type-checking breakage; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Llama4Config pad_token_id AttributeError is a specific config bug, not the same as a broad v5 breakage complaint." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both mention v5 migration, but one is checkpoint reload failure and the other is remote-code/meta-init incompatibility; not the same code-path problem." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star performance issue vs missing GenerationMixin cache during chunked_prefill; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "Typo in cache strategies vs vague broad regression claim; no shared concrete bug." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLmConfig pad_token_idx missing is a specific config regression, not the same as a generic v5 breakage report." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API settings migration vs a cache-strategy typo; unrelated issues." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad v5 breakage reports, but one is import errors and the other is a generic \u201ctoo many models\u201d complaint; too vague to call the same duplicate." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Missing set_seed in tests vs inverse_sqrt scheduler kwargs being ignored; different areas." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding vs scheduler kwargs handling; unrelated code paths." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type warning for sam3_tracker vs Liger Kernel not applied with model_init; unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Slow import sometimes vs lazy loading not working properly; related to imports only at a high level, but not the same bug." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility vs Qwen3-VL bounding-box output error; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support limitation vs Qwen3-VL output hallucination/error; different problem." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command vs scheduler kwargs ignored; no shared code path." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "tokenizer_class mismatch on save_pretrained vs a generic bug report; insufficient evidence of duplication." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs bug vs PretrainedConfig type checking breakage; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-star performance issue vs cannot import HybridCache; both involve imports, but different concrete failures." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer read error vs generic bug report; no basis to merge." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA ReLU discussion vs Gemma4AudioRelPositionalEncoding issue; unrelated." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading issue vs scheduler kwargs bug; different functionality." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Generic bug report vs Qwen3.5 MoE from_pretrained error; too little overlap to treat as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 45, + "review_unit_count": 45, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-393", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:37428", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43618", + "issue:43704", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45092", + "issue:45290", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45850" + ], + "soft_pairs": [ + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:31795|issue:37428", + "issue:44315|issue:45092", + "issue:44485|issue:45468", + "issue:43704|issue:44485", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:44261|issue:45468", + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44861|issue:45290", + "issue:44492|issue:44998", + "issue:44998|issue:45850", + "issue:43618|issue:44998", + "issue:44998|issue:45341", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "4b66ea23f01b572d0b3cc11af7df8b8f2e45910b7564723019f4bc3956d1ab52", + "component_budget": { + "node_count": 393, + "item_count": 393, + "soft_pair_count": 1479, + "serialized_chars": 761590, + "estimated_input_tokens": 190398, + "estimated_eval_tokens": 381052 + }, + "budget": { + "node_count": 36, + "item_count": 36, + "soft_pair_count": 32, + "serialized_chars": 29795, + "estimated_input_tokens": 7449, + "estimated_eval_tokens": 15154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: a few v5/tokenizer/model-loading regressions are superficially similar, but most items are unrelated bug reports. I did not find any soft pair that looks like the same concrete fixable bug, so no duplicate merges are justified.", + "confidence": 0.28, + "canonical_issue_reason": "issue:45850 is the broadest umbrella for the loose 'Transformers 5.0 breaks models' theme, but the cluster is too mixed for a strong canonical issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the closest thing to a cluster-wide summary, but it is too vague to serve as a high-quality duplicate anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Different model bugs: Qwen3VL timestamp calculation vs GLM-5 RoPE implementation." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Unrelated code paths: processor timestamp logic vs inverse_sqrt scheduler kwargs handling." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "One is about import-* cleanup/perf, the other is a VRAM leak in Qwen3ForCausalLM." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion vs a concrete model-loading regression." + }, + { + "left": "issue:31795", + "right": "issue:37428", + "accept": false, + "reason": "Documentation issue vs ImportError from flash attention utils." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Different failures: Liger kernel application in model_init vs remote-code/meta-init incompatibility." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Different architectures and bugs: GLM-5 RoPE vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to GLM-5 RoPE behavior." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings API change is unrelated to a dataloader-thread VRAM leak." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network access bug is unrelated to Qwen3VL timestamp calculation." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp math and processor.save_pretrained missing files are different bugs." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Timestamp calculation bug vs VRAM leak; no shared code-path problem." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return-value issue in GlmMoeDsaIndexer is unrelated to Gemma audio positional encoding." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch is not the same bug as Gemma audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "MLA q_a_layernorm epsilon/config precision issue is unrelated to Gemma audio encoding." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5, but one is broad behavior regression and the other is decoder-type mismatch; not the same concrete bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to the placeholder unemployment issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weights helper crash is unrelated to chat template tool-call handling." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A typo in cache strategies is unrelated to the unemployment issue." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "Generic 'many models break' complaint does not match the unrelated unemployment report." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to the unemployment issue." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "Testing_utils bug and unemployment are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config is unrelated to _get_tied_weight_keys on list inputs." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key handling." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading is unrelated to a configuration_utils ImportError." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 support work is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting after failed training and slow checkpoint loading are different problems." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "from_pretrained performance regression is unrelated to tensor-parallel training bugs." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading slowdown and documentation confusion are not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 inference support request is unrelated to the Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowdown and tokenizer loading regression are different issues." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching audio-model support and qwen fp8 inference support are unrelated features/bugs." + } + ] + }, + "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": 17, + "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": "21e3a71f0c0bd65c709cadf9728c56b6bb17e51d6de26c8457dde6dc6c8133d6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two different community feature requests in the same general area (model/runtime support) but for unrelated backends: Flash Attention 2 vs GGUF. They should not be deduplicated.", + "confidence": 0.95, + "canonical_issue_reason": "Issue #26350 is the more central representative of this cluster because it has substantially higher discussion activity and is the older, broader feature request.", + "canonical_pr_reason": null, + "best_issue_reason": "#26350 is the best issue representative here because it has the most engagement and the topic is a concrete support-extension request, but it is not a duplicate of #33260.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different underlying changes: Flash Attention 2 support and GGUF support are separate architecture/backend requests with different code paths and implementation goals." + } + ] + }, + "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": 17, + "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-27088-4", + "nodes": [ + "issue:27088", + "issue:43834", + "issue:44365", + "issue:44370" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:44365|issue:44370" + ], + "prepared_review_unit_hash": "ec7a7fe249462e7a0f339892d8b7cf27e3df5c7231b6e7b35397688de8e12f27", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4453, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4453, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues all use the same documentation-translation template, but they refer to different languages, so they are not the same underlying task/bug. The Turkish issue is the only concrete one; the others look like placeholder or separate language requests.", + "confidence": 0.9, + "canonical_issue_reason": "issue:27088 is the only concrete, language-specific request ([i18n-TR] Turkish). The others are generic placeholders for different languages, so they should not be merged as duplicates of the Turkish task.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:27088 is the best representative because it is specific and actionable, unlike the placeholder/template issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same translation template, but not the same underlying request: Turkish vs another language placeholder." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Same template wording, but they represent different languages and therefore different tasks." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Both are docs-translation requests, but the language target differs, so they are not duplicates." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Generic language placeholders; no evidence they refer to the same specific language or work item." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Template-matched requests only; different language placeholders, so not the same issue." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Both are placeholder translation issues, but for different languages, so they should remain separate." + } + ] + }, + "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": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-15", + "nodes": [ + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43671|issue:43909", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177" + ], + "prepared_review_unit_hash": "e673687f895e0935c0f9afcf686ae3d52b17c7b25d21532482a32d08c4a52732", + "component_budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 35, + "serialized_chars": 20706, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 + }, + "budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 35, + "serialized_chars": 20706, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "No duplicate cluster here: the items are all separate model-request issues spanning different models, modalities, and architectures. The high soft similarity is driven by templated wording, not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model requests (EuroBert config vs DEIMv2 support); same request template only." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS are unrelated model additions." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert config and RF-DETR are different models and changes." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are distinct model requests." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio are unrelated models/modalities." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and the weight-sparse transformer request are different underlying changes." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are unrelated model additions." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are different models and tasks." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated model requests." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT and the weight-sparse transformer request are different model families." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are unrelated additions." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different models and code paths." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are distinct model additions." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support and Qwen3-ASR are unrelated requests." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 architecture support and the sparse transformer request are different changes." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are distinct model requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture support and Voxtral TTS are unrelated models." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are different model additions." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are different models/tasks." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and BERT-like models with RoPE are not the same change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "Generic RoPE support for BERT-like models and Voxtral TTS are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct modalities and requests." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different models; not one code-path fix." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and the sparse transformer request are different underlying changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model additions." + }, + { + "left": "issue:43837", + "right": "issue:44205", + "accept": false, + "reason": "Qwen3-ASR and SAM3-LiteText are unrelated models." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse transformer support and EUPE are different model requests." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated additions." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and the sparse transformer request are different model families." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different models and modalities." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model additions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 17, + "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-37934-6", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031" + ], + "soft_pairs": [ + "issue:37934|issue:43668", + "issue:42936|issue:43240", + "issue:42936|issue:44031", + "issue:42936|issue:43813", + "issue:42936|issue:43668", + "issue:43240|issue:43668", + "issue:43668|issue:44031", + "issue:43668|issue:43813" + ], + "prepared_review_unit_hash": "90fda2da84ff3715f67316d45749179f936f3ac170881cc4e0ad93361a1b1414", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues only share superficial \u201csmall bug in model/config handling\u201d similarity; they describe different models, parameters, and failure modes, so none of the soft links look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43668 is the most broadly representative of the cluster\u2019s loose theme (a concrete API/config correctness bug), but the set is still too heterogeneous to justify a stronger canonical duplicate grouping.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43668 is the clearest, most self-contained problem statement and the most central by similarity, even though it should not absorb the others as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Different models and different bugs: Llama4TextL2Norm vs ModernBERTConfig type hint; only generic norm-related wording overlaps." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "One is Mask2Former ignore_value not being used; the other is kwargs not forwarded to a loss function. Same broad area, different code paths." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to tokenizer regex warning behavior." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Different subsystems and failure modes; a typo in one quantization attribute is not the same as an unused ignore_value setting." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Both are parameter-related bugs, but they affect different models and different code paths; not the same underlying issue." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Forwarding kwargs to loss calculation is unrelated to ModernBERTConfig norm_eps typing." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Incorrect type hint in a model config and tokenizer regex warnings are unrelated problems." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "A config type-hint typo is not the same bug as a quantization attribute typo." + } + ] + }, + "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": 17, + "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": "d7537a911c297ab83e9b32b938ff362639e2ad3647a9d572e95c93d8a308a23c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are not duplicates: one is about block decoding mechanics, the other is a request for DeepSeek v3.2 support. They refer to different user problems and different likely code paths.", + "confidence": 0.97, + "canonical_issue_reason": "If a representative must be chosen, issue #41084 is the more concrete technical issue; #43037 is a broad model-support request and does not overlap in substance.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #41084 is the better cluster representative because it describes a specific decoding feature/problem rather than a generic support request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different topics and problem domains: block decoding vs DeepSeek v3.2 support. No evidence they describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 17, + "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": "0fb68af1bb2a9b173c9617c778fa1891a4e9cfa52804cd03f448444d66eff91f", + "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 are related to Trainer custom behavior, but they describe different problems: one is about `remove_unused_column` configurability, the other is about FLOPS estimation for custom trainers. They are not the same underlying bug.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue fits well here because the two issues target different code paths and user problems.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a true duplicate; if forced to pick a representative, `Revisit remove_unused_column in Trainer for better customizability` is the clearer broader Trainer customization request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Both mention custom trainers, but one concerns unused-column handling and the other concerns FLOPS estimation. Different behavior, different fix." + } + ] + }, + "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": 17, + "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": "40e0d97fb77b8a0b2489c3aa8f9267e07315879f51947c1e7d0b79ad531e428d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related only at a very high model-family level, but they request different capabilities for different Qwen variants: one is about registering Qwen3-Omni with generic auto classes, the other is about adding sequence classification support for Qwen3.5. Not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more central request: it targets AutoModel/AutoModelForConditionalGeneration registration for Qwen3-Omni, which is a foundational compatibility change.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the better canonical issue because it describes a general registration gap rather than a narrower task-specific head support request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying changes: model registration for Qwen3-Omni auto classes versus sequence classification support for Qwen3.5. They concern different model variants and different code paths, so they should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 17, + "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": "15efe102201ef0e23434d739dcf5af12fba357aecb051706e79084645039133c", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4262, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three closed issues share the exact title \"huggingface \" and look like duplicates; issue #43835 titled \"Fett\" appears unrelated. I would keep the earliest huggingface report as the canonical issue.", + "confidence": 0.84, + "canonical_issue_reason": "Issue #43845 is the earliest of the three near-identical \"huggingface \" reports, so it is the best canonical issue among the likely duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "No richer issue content is available, but #43845 is the earliest clearly matching report and is the best single issue to retain for the duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no evidence they describe the same problem; only weak similarity." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no shared target or file evidence; likely unrelated." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "\"Fett\" does not match the \"huggingface\" reports and there is no substantive overlap." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same exact title and similar minimal metadata; these look like duplicate reports of the same issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and same apparent topic; likely another duplicate report." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and no distinguishing evidence; likely the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: #43845, #43846, and #43859 share the exact title \"huggingface \" and are the only plausible duplicate cluster, while #43835 has a different title and no evidence tying it in. The canonical choice of #43845 is conservative because it is simply the earliest of the matching trio. The pairwise verdicts are also suitably cautious, relying only on identical titles and the absence of distinguishing evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 17, + "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": "13c6327423b956499c64887bbae323eec5630302f837b284486869ae99214b03", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5779, + "estimated_input_tokens": 1445, + "estimated_eval_tokens": 3146 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5779, + "estimated_input_tokens": 1445, + "estimated_eval_tokens": 3146 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like unrelated low-information issue noise/spam, not true duplicates. There isn\u2019t enough substance to merge any pair as the same underlying bug or request.", + "confidence": 0.58, + "canonical_issue_reason": "issue:44418 is the only item with a non-placeholder title, so it is the most usable anchor if one must be chosen; the others are generic spam-like placeholders.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44418 is the most descriptive issue in an otherwise content-poor cluster, but it should still be treated cautiously because the cluster does not show a shared concrete problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Same placeholder-style title, but no body/context to show a shared underlying bug or change." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Identical spam-like titles are not enough to establish the same issue; no concrete evidence ties them together." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is a placeholder/spam-like issue and the other is a generic documentation suggestion; no clear shared problem." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Different titles and no supporting details; cannot treat them as the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Generic documentation suggestion versus placeholder spam; no evidence of the same underlying topic." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Generic documentation suggestion versus placeholder spam; too little information to merge." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Same placeholder title, but identical titles alone are insufficient without any issue content showing the same bug/request." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "No shared concrete subject matter is visible; the titles do not indicate the same underlying issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete overlap beyond being closed issues; not enough to call them duplicates." + } + ] + }, + "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": 17, + "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": "9eb1b482dc991bc2be0dc9c26039713098674912f853eefdefe756a6dd20ee80", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The pair is only weakly similar: same title and comparable body length, but no direct body comparison or other hard signals. With deterministic_accept=false, this should be treated conservatively as a non-duplicate soft match rather than an accepted duplicate.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45145 is the best single representative if one must be chosen: it is earlier and has the only inbound reference, but this does not by itself justify merging the pair as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "Identical title alone plus soft similarity is not enough here; there is no body-level comparison, target, filename, or other deterministic evidence that these are the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate call is too strong for the packet. The evidence only supports soft similarity: identical title and similar body length, but no direct body comparison, links, targets, filenames, or deterministic accept signal. The summary also overstates by saying the bodies are likely the same and that there are no distinguishing signals. Canonicalization to #45145 is fine only if the pair itself is accepted, but here the soft-edge verdict should remain conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 17, + "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": 14, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45672-2", + "nodes": [ + "issue:45672", + "issue:45674" + ], + "soft_pairs": [ + "issue:45672|issue:45674" + ], + "prepared_review_unit_hash": "8bb320c344d835e1306d73607cde92e67523e8eff5967b31517c17f3b666f9a0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two different 4-bit inference bugs: one is a Gemma4/audio-module uint8 dtype TypeError from `torch.finfo()`, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an AssertionError. They are related only at a broad subsystem level, not the same underlying issue.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #45674 is the better anchor because it describes a specific, reusable configuration regression with a clear failure mode and reproduction path. #45672 appears to be a separate Gemma4/audio-module dtype bug.", + "canonical_pr_reason": null, + "best_issue_reason": "#45674 is the strongest representative issue in this cluster because it is narrowly defined and actionable; however, the cluster itself looks heterogeneous rather than duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Different root causes and failure paths: #45672 is a Gemma4 audio-module `torch.finfo()` TypeError on uint8 weights, while #45674 is a BitsAndBytesConfig `llm_int8_skip_modules` / `lm_head` exclusion regression causing AssertionError. Same broad area, 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": 15, + "component_count": 17, + "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": "313c85982daf129cf286dc5af6243be86df942233c1215e9ee6683019e0ae5ed", + "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": "Two closed issues with empty bodies and unrelated personal-name titles; they do not appear to describe the same underlying bug or change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good canonical representative: both are low-signal, unrelated entries with no description, targets, or evidence of a shared problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Reject: the titles refer to different people and dates, and there is no body or other evidence indicating 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": 16, + "component_count": 17, + "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": "39e5d44a72d7ed6f82f415b83f1ee459873bc381a21267e8f4b558e71fe31d2a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues only share a loose text similarity in their titles and do not describe the same bug, request, or code change. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Unrelated personal-name/date titles with no shared problem statement, code context, or issue target; not the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 17, + "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": "203baefa561837a9b95002ae37334d4facbfe89f7517ecbfa88cacd8f9f60b84", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster contains two clear duplicate pairs: 45870/45871 and 45872/45873. Issue 45869 is about a dataset offering and does not match the example/scraping-service requests, so it should stay separate.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue covers the whole cluster; the items split into two separate duplicate groups. If one representative must be chosen, 45872 is the strongest because it is the more specific, concrete request and has an exact duplicate in 45873.", + "canonical_pr_reason": null, + "best_issue_reason": "45872 is the best representative issue overall: it is specific, concrete, and clearly has an exact duplicate partner (45873).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different requests: dataset bundle advertisement vs a financial time-series example. Same general finance theme, but not the same underlying issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different requests: dataset bundle advertisement vs a financial time-series example. Not the same bug/change." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Same title and same example request; these are duplicates describing the same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both are finance-related examples, but 45872 adds a scraping-service component and is a different request from the example issue." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Different request content: example time-series prediction vs time-series plus scraping service. Related topic, not the same issue." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Same reasoning as 45870/45872: overlapping finance theme, but different concrete request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different request content: example time-series prediction vs time-series plus scraping service. Not a duplicate." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Exact title match and matching request content; these are duplicates of the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: 45870/45871 and 45872/45873 are the only clearly identical pairs by title/body similarity, and 45869 is visibly different in topic. The soft-edge rejections stay conservative, and the overall \u2018no single canonical issue\u2019 conclusion fits the split cluster structure." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-13", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:45795" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "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:44827", + "pull_request:43995|pull_request:45213", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:45795", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43823|pull_request:45795", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "a5980b02fed89ceb74128a8df1753ec43bb8a1abd42b4998de4dd3a887294399", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 19, + "serialized_chars": 35206, + "estimated_input_tokens": 8802, + "estimated_eval_tokens": 17860 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 19, + "serialized_chars": 35206, + "estimated_input_tokens": 8802, + "estimated_eval_tokens": 17860 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated PRs across docs, new model adds, and small refactors. The only true duplicate pair is the two `fixed_cross_entropy` PRs that both add `weight`/`label_smoothing` support for the same helper in `loss_utils.py` and target the same issue. Everything else is only superficially similar by subsystem or broad theme.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical PR for the loss-utils duplicate pair: it directly targets issue 43240, makes the same concrete `fixed_cross_entropy` change, and is the more explicit fix rather than the alternative phrasing in 43254.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest representative artifact in this cluster because it has the clearest problem statement, explicit issue linkage, and the most directly scoped code change. The rest of the cluster does not form a single cohesive fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Both are unrelated feature/test PRs; they do not fix the same code-path problem and cannot plausibly be merged as one change." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "One is tokenizer API cleanup, the other is adding a new model; shared model-adjacent wording is not enough for duplication." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "A tokenizer method addition and an executorch export test are unrelated changes." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch processor/model auto plumbing in a broad sense, but they address different models and different bugs." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss helper kwargs support and adding a new model are unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "These are different fixes in different areas; the overlap is only that both are small code changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "An executorch dynamic-shape test and Mistral4 test fixes are not the same underlying change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "A test for exportability and a new model addition are not duplicates." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "A new model implementation and a class-annotation resolution fix are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "Adding MobileLLM and refactoring Falcon output collection are different model work, not one bug fix." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Both touch model internals, but they fix different models and different code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactoring and a model-creation skill are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and a CLIP-like weight renaming fix are different changes with no shared concrete defect." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45795", + "accept": false, + "reason": "Mistral4 fixes and documentation for audio/video processors are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Different bugs in different subsystems; no evidence they are the same underlying issue." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "A model test fix and a draft model-creation skill are not the same change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "Adding MobileLLM is not the same as creating an internal model-creation skill." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45795", + "accept": false, + "reason": "New model code versus documentation updates for processors are not duplicates." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs modify `src/transformers/loss/loss_utils.py` to pass the same extra `cross_entropy` kwargs (`weight` and `label_smoothing`) through `fixed_cross_entropy`; they fix the same concrete bug and are merge-equivalent." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The lone accepted duplicate pair (43251/43254) is well supported by the shared `loss_utils.py` edit and shared issue target 43240, and the other soft-edge rejections do not overstate similarity. The cluster characterization as mostly unrelated PRs is also supported by the listed diffs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "7f37e71834fcb93ae91f3ab81798b0252fabb5dfea0cefccce65ad1df0abd942", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related VibeVoice PRs, but they are not the same change: one adds the base VibeVoice integration, while the other adds VibeVoice Realtime and additional tokenizer/feature-extraction pieces. Shared filenames reflect the same subsystem, not duplicate work.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 40546 is the broader foundational VibeVoice implementation and is the better representative of the cluster than the later realtime-focused extension.", + "best_issue_reason": null, + "best_pr_reason": "PR 40546 is the best single PR to represent the cluster because it introduces the core model integration, docs, and auto-registry wiring.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Different underlying change: 40546 implements VibeVoice, while 42781 adds VibeVoice Realtime and related tokenizer/feature-extraction support. They touch overlapping files but not the same concrete bug or mergeable patch." + } + ] + }, + "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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "4ad4ae34e24f6dee3e60e33f812b00bff48f542c737e2be1fd1d6cb667559c01", + "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 pull requests are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They should not be deduplicated or clustered as the same change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different underlying changes and code paths: one adds covariate-aware forecasting to TimesFM, the other adds DINOv3 image classification support. They do not look mergeable into one PR and are not the same bug/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": 4, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-3", + "nodes": [ + "pull_request:41797", + "pull_request:45186", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186", + "pull_request:45186|pull_request:45630" + ], + "prepared_review_unit_hash": "771256700e3448515692b8cc95d84cec66d98e727adbe10dbfe99534234a9c7b", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10429, + "estimated_input_tokens": 2608, + "estimated_eval_tokens": 5472 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10429, + "estimated_input_tokens": 2608, + "estimated_eval_tokens": 5472 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three PRs are separate model-addition efforts (DeepSeek OCR, Isaac, and Kimi2-6). They share Transformers integration boilerplate, but they do not describe the same bug or change, so no duplicate/merge is warranted.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR among these: each PR introduces a different model integration and changes different model-specific files.", + "best_issue_reason": null, + "best_pr_reason": "PR #45630 is the most complete and actively reviewed model-addition PR in the set, so it is the best representative if one must be chosen, but it is still not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "DeepSeek OCR vs Isaac are different model integrations with different model-specific files and code paths; shared auto-registration/docs boilerplate is not enough." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Isaac vs Kimi2-6 are distinct model additions; they only overlap on common Transformers wiring, 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": 6, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "9d18f724c64066a3188ebcf35a73cd036a231fc1112bd21d0d57c789cc01c39d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two independent workflow-maintenance PRs: one upgrades checkout/cache/upload-artifact for Node 24 compatibility, the other upgrades docker/build-push-action and related workflow actions. They overlap in GitHub Actions files but address different action families and are not the same underlying change.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #42975 is the broader, more representative workflow-tooling upgrade and explicitly targets Node 24 compatibility across many workflows.", + "best_issue_reason": null, + "best_pr_reason": "PR #42975 is the best cluster representative because it spans more workflow files and captures the general GH Actions runtime compatibility migration.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "These are different upgrades: #42975 updates checkout/cache/upload-artifact, while #42976 updates docker/build-push-action. Same subsystem, but not the same concrete fix and not something that clearly belongs in one merged PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-7", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:43973", + "pull_request:44601", + "pull_request:45490", + "pull_request:45838" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490", + "pull_request:43973|pull_request:45838", + "pull_request:44601|pull_request:45838", + "pull_request:45490|pull_request:45838" + ], + "prepared_review_unit_hash": "b28d53845c1daa33beba3ad11df7e18b78edf60589bd0113b937c067de5a341b", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 23515, + "estimated_input_tokens": 5879, + "estimated_eval_tokens": 12014 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 23515, + "estimated_input_tokens": 5879, + "estimated_eval_tokens": 12014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of mostly unrelated model-addition PRs plus one distributed loading PR. The shared filenames and similarity scores are driven by repository templates and auto-mapping boilerplate, not the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: each PR adds a different model or subsystem change (ViT NEPA, Molmo, Molmo2, lfm2 audio, CTSM, PP-OCRv6, and native PP loading), so they are not duplicates of one another.", + "best_issue_reason": null, + "best_pr_reason": "There is no strong global best PR for deduping because the items do not describe one common change; if forced to pick a representative, any choice would be arbitrary.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models (ViT NEPA vs Molmo); similar docs/auto-registry edits are template overlap, not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different models (ViT NEPA vs Molmo2) with distinct code paths and files; not the same underlying work." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA model addition is unrelated to CTSM time-series model addition." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are separate model integrations with different implementations; they are related by naming, not by being the same PR." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are different model families; overlap is only in shared framework boilerplate." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are distinct model additions and do not fix or add the same code-path problem." + }, + { + "left": "pull_request:43973", + "right": "pull_request:45838", + "accept": false, + "reason": "Lfm2 audio support and PP-OCRv6 text recognition are different model integrations with no shared concrete fix." + }, + { + "left": "pull_request:44601", + "right": "pull_request:45838", + "accept": false, + "reason": "Native pipeline-parallel loading infrastructure is unrelated to PP-OCRv6 model support; not mergeable as one PR." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45838", + "accept": false, + "reason": "CTSM model addition and PP-OCRv6 recognition models are different changes in different subsystems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-4", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:45892" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:45892", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:45892", + "pull_request:44815|pull_request:45892" + ], + "prepared_review_unit_hash": "a7144b309434a0cab1768e797298f5b8c7cc8f62ca61249e9241515cf6ce4ad5", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10131, + "estimated_input_tokens": 2533, + "estimated_eval_tokens": 5322 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10131, + "estimated_input_tokens": 2533, + "estimated_eval_tokens": 5322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly unrelated PRs that all use generic \u201cfix\u201d wording. Two are test-only changes, one is a Whisper tokenizer/test change, one is a FP8 dequantization change, and one is a DeepSeek v4 attention fix. They do not look like the same underlying bug or change cluster.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR stands out here; if one representative must be chosen, 44815 is the most substantive code fix and the broadest patch in the set, while the others are either test-only or different model-family bug fixes.", + "best_issue_reason": null, + "best_pr_reason": "44815 is the best representative because it contains a concrete implementation fix across core loading and FP8 integration, whereas 43665 is test noise and 43291/45892 are narrower model-specific changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer/test changes vs unrelated CLIP/ViT test edits; no shared underlying bug or code path." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper decoding bug/test work is unrelated to FP8 dequantization and model loading." + }, + { + "left": "pull_request:43291", + "right": "pull_request:45892", + "accept": false, + "reason": "Different subsystems and models: Whisper tokenizer vs DeepSeek v4 attention logic." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "43665 is clearly test-only/assert noise; 44815 is a real FP8 dequant fix, so they are not the same change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:45892", + "accept": false, + "reason": "Test-only CLIP/ViT edits are unrelated to the DeepSeek v4 model bug fix." + }, + { + "left": "pull_request:44815", + "right": "pull_request:45892", + "accept": false, + "reason": "Both are real fixes, but they target different code paths and model families; not mergeable 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": 9, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "312df79426bba27e475205f3e76e6a7fa33b4f4d511d4932e1c6f19cea78a485", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one is a setup/version/dependency tweak, the other is a CI/workflow and testing-utilities change. The shared Vit test file looks incidental, not evidence of the same fix.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43532 is the more representative change set: it spans CI workflow and testing utility code, while 43488 is a small setup/dependency bump plus a test touch. Neither is a duplicate of the other.", + "best_issue_reason": null, + "best_pr_reason": "PR 43532 is the better cluster representative because it contains the broader substantive code-path change; 43488 appears to be a narrow bot/setup-oriented edit.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Different underlying changes. 43488 updates setup/versioning and adds a pytest dependency, while 43532 rewires CI/workflow behavior and touches testing utilities/notifications. The shared Vit test file is not enough to treat them as the same PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "4f60e83b520119d58447bddf8120a30a53cc5926ac8b4e471c535f11ce084ac7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are distinct feature PRs: one adds Promptable Concept Segmentation and the other adds Promptable Visual Segmentation. They overlap on shared pipeline/docs scaffolding, but they implement different pipelines, different model integrations, and different test suites, so they should not be treated as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No issue is present. If a representative PR is needed, #43613 is the broader/more complete feature addition, with more model integrations and a larger test surface.", + "best_issue_reason": null, + "best_pr_reason": "No issue is present. #43613 is the better cluster representative because it spans the larger concrete change set and central pipeline registration work, but it is not a duplicate of #43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Reject: they add different pipelines for different tasks (concept vs visual segmentation). Shared docs/registry files are just common scaffolding, and the implementation/test/model files are distinct, so 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": 11, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "f95d398d7589f5073a33ef9c7d9ce2644d8bad37299db0481a9cf9d69edb973c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate Trainer extension PRs touching different code paths in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT. They are not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the two changes address different behaviors and should not be merged as one duplicate candidate.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other, so there is no global best PR for this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different underlying changes: 43651 alters training loss scaling during gradient accumulation, while 43663 adds an overridable helper for signature-column filtering. Shared file alone is not enough to treat them as the same bug or 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": 12, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-3", + "nodes": [ + "pull_request:43656", + "pull_request:43836", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842", + "pull_request:43836|pull_request:43842" + ], + "prepared_review_unit_hash": "9635ed14b57f971fc1bc28589d67d341ab2c2172268a37be941ddc014316d5c5", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5991, + "estimated_input_tokens": 1498, + "estimated_eval_tokens": 3252 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5991, + "estimated_input_tokens": 1498, + "estimated_eval_tokens": 3252 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is centered on a `serve.py` annotation-related `TypeAdapter` NameError. `43836` and `43842` are tightly aligned on the same tracker (`43824`) and the same concrete fix area, while `43656` is related but broader and aimed at a different issue target (`43576`).", + "confidence": 0.84, + "canonical_issue_reason": "`43824` is the clearest shared issue proxy because it is explicitly targeted by both `43836` and `43842`, and it matches the concrete `TypeAdapter` NameError in `serve.py`.", + "canonical_pr_reason": "`43842` is the best canonical PR because it is the most focused fix for the `TypeAdapter` NameError in `serve.py`, with a minimal change and an explicit link to the shared issue.", + "best_issue_reason": "`43824` is the best issue representative for the cluster because it is the common tracker for the most tightly matching PR pair and corresponds to the specific runtime failure being fixed.", + "best_pr_reason": "`43842` is the strongest representative PR: it directly and narrowly addresses the `TypeAdapter` import-time NameError in `serve.py` without adding the broader annotation cleanup present in `43656`.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Related in the same file and both touch annotation evaluation, but `43656` targets a different issue (`43576`) and applies a broader `from __future__ import annotations`/annotation cleanup fix rather than the focused `TypeAdapter`-specific bug in `43842`. Not clearly the same underlying change." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": true, + "reason": "Both PRs target the same issue (`43824`) and address the same concrete `TypeAdapter` NameError in `serve.py`, differing mainly in how narrowly they express the fix. They are plausibly mergeable into one bugfix PR." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is a bit too strong on the soft edge for 43656 vs 43842. 43656 targets a different issue (`43576`), and its fix is broader (`from __future__ import annotations` plus wider annotation cleanup) than the focused `TypeAdapter`-specific change in 43842. The packet supports saying they are in the same file and plausibly related, but not that they clearly fix the same bug or are comfortably mergeable into one PR for the same bugfix. The 43836/43842 pairing is better grounded because they share issue `43824`, but the 43656/43842 accept is overstated." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": 14, + "component_count": 27, + "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:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44086", + "pull_request:44018|pull_request:44085", + "pull_request:44071|pull_request:44085", + "pull_request:44068|pull_request:44086", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request: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:44059|pull_request:44071", + "pull_request:44073|pull_request:44161", + "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:44056|pull_request:44161", + "pull_request:43997|pull_request:44018", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44071|pull_request:44129", + "pull_request:44010|pull_request:44018", + "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": "34a2a01bc9aad95611920483ea6b3b9ff708b495fa3e77d064cfd7d0a1ab79a5", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84895, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75092, + "estimated_input_tokens": 18773, + "estimated_eval_tokens": 37802 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of model-specific output-tracing refactors around issue #43979. The only true duplicate subset is GPT-J/CodeGen: 44066, 44085, and 44722 overlap on the same GPT-J output-tracing code path. The GPT-Neo pair is only thematically similar and should not be treated as a duplicate.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best canonical PR for the duplicate subset because it is explicit about the shared issue, spans the shared GPT-J/CodeGen refactor, and is the broadest representative of the overlapping patch set.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the strongest representative overall: it directly covers the shared GPT-J/CodeGen output-tracing change and cleanly anchors the actual duplicate cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families; only the general output-tracing theme overlaps." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model implementations and different concrete refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing code path with overlapping file changes; these could be merged into one PR." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet and EfficientNet are separate model code paths; same pattern, not the same change." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are unrelated concrete changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models and files; only a broad refactor theme matches." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTaV2 are unrelated implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "Same output-tracing pattern, but different model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen versus MPT is only a theme match, not the same bug or patch." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-Neo pairing is only soft similarity; no shared issue target or concrete duplicate signal." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different model scope; the overlap is just the standardized output-tracing pattern." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different concrete code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated model changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet share only the broad refactor style." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are separate model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different model changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo refactor is separate from the GPT-J/CodeGen duplicate set." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different concrete changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are separate model refactors." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are not the same underlying change." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are different model implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are unrelated concrete changes." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 refactor is unrelated to GPT-J/CodeGen output tracing." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT only share the high-level modernization pattern." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are different model refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are different model implementations." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert/XLM are separate model changes." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert share only the refactor theme." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different concrete implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are not the same underlying change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBERT and MPT are unrelated model paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBERT and VisualBert are different implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTaV2 and SpeechT5 are different model refactors." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTaV2 and ViTS are not the same change." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44074", + "accept": false, + "reason": "SegFormer and TextNet are separate model code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "MPT and SpeechT5 are unrelated changes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and SegFormer are only loosely similar refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are different model code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are unrelated implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing refactor on the same file; these overlap enough to be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different models and changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet are different model refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated model paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTaV2 and SpeechT5 are different concrete changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTaV2 and TextNet are unrelated model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen standardized output-tracing refactor with direct file overlap; this is a true duplicate/overlap." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly grounded, but it overstates the GPT-Neo pairing. In the packet, pull_request:44018 vs pull_request:44068 is only a soft-similarity match; there is no deterministic accept signal or shared issue target, so calling them a clear duplicate subset is too strong. The GPT-J/CodeGen duplicate claims are well supported, but the GPT-Neo verdict should be more conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 27, + "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": "f1e620e090e23fda40c46d4d4eecda9d49a3128a880cb4ed434e7d2317f46c33", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84895, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17124, + "estimated_input_tokens": 4281, + "estimated_eval_tokens": 8818 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs are all variants of a standardized output-tracing refactor, but they touch different model implementations and code paths rather than the same concrete fix.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44073 is the clearest representative: it explicitly targets the standardized output-capture refactor, adds the new tracing hooks, and shows the core pattern used across the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 44073 is the best single exemplar because its title and diff most directly reflect the shared refactor theme and it includes both the decorator import changes and output-capture hook setup.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Different model families (FocalNet vs ImageGPT) and separate implementations; same refactor style, but not the same underlying bug or mergeable fix." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert vs TextNet; both migrate output tracing, but they modify different model code paths and are not one concrete change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert are unrelated model implementations; these are parallel refactors, not duplicate fixes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "Both are output-tracing refactors, but for different architectures and files, so they should not be merged as one PR." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert address different model internals; shared decorator imports do not indicate the same fix." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are distinct code paths; the similarity is only the output-capture migration pattern, not a duplicate change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "71cf12ecd24f9b0a0d704cceddddfbfcfbd0f5be6d298703ce86f7bed31080ba", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40162, + "estimated_input_tokens": 10041, + "estimated_eval_tokens": 20338 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40162, + "estimated_input_tokens": 10041, + "estimated_eval_tokens": 20338 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of similar-but-independent output-tracing refactors across many unrelated model implementations. They share a naming pattern and plumbing changes, but not the same concrete bug or change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the best representative: it makes the most substantial tracing-related plumbing changes, adds output-capturing support, and spans two closely related model files (mamba/falcon_mamba).", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the strongest single representative of the cluster because it combines the core tracing infrastructure change with model-specific updates; the other PRs are narrower, mostly single-model refactors.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths; only the refactor theme is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Both are output-tracing refactors, but for unrelated model architectures and distinct forward paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Same general plumbing theme, but not the same underlying bug or mergeable change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Timm backbone and mamba/falcon_mamba are unrelated code paths; shared tracing terminology is not enough." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different models and mechanisms; no common concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Different models; both change tracing behavior, but not the same underlying issue." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Separate model implementations with distinct forward logic." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families; similar refactor pattern only." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "MobileNet V1 and vision encoder-decoder wrappers are unrelated code paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different architectures; no evidence of one concrete shared bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Unrelated models; same refactor motif but not the same fix." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different code paths and model types." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Vision-text dual encoder vs mamba/falcon_mamba are unrelated implementations." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models with no shared concrete bug." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Unrelated model families; only the refactor style matches." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "No common code path; one is UnivNet, the other is Mamba-family tracing." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models and different output-tracing implementations." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and forward paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "UPerNet and Mamba-family changes are not the same underlying change." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Mamba-family tracing and SuperPoint tracing are unrelated implementations." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different models and different refactor scope; no mergeable shared fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Timm backbone and CodeGen are different code paths; same theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen are unrelated models." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Distinct models with distinct forward semantics." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder and CodeGen are unrelated; not the same bug." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different wrapper models and different call chains; only both are tracing refactors." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Depth Anything and Speech Encoder-Decoder are unrelated model families." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "No shared concrete problem; different model implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Different models and different output-capturing changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Depth-related tracing and DPR tracing are separate changes." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Vision and speech encoder-decoder wrappers are parallel but distinct code paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models with no common underlying bug." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Distinct architectures; shared tracing theme is too broad." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and implementation details." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech encoder-decoder and SuperPoint are unrelated." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different code paths and model families." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Separate models; not the same fix." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint and RWKV have unrelated forward logic." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different models; no evidence of a shared concrete issue." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are unrelated implementations." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-3", + "nodes": [ + "pull_request:44662", + "pull_request:45075", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401", + "pull_request:44662|pull_request:45075", + "pull_request:45075|pull_request:45401" + ], + "prepared_review_unit_hash": "3f16e3926c01faf250a3a8aed226b646b1b464af9a03ececae777f6006bb1efb", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-addition PRs, not duplicates. They share Transformers boilerplate (docs, auto registries, model package wiring), but each targets a different model family and code path: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45075 is the best standalone representative of the cluster because it is a non-draft, substantial model integration with a dedicated conversion script and broad test coverage; however, it does not subsume the other PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR #45075 is the strongest overall example of a complete model-implementation PR in this set, but it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models and different concrete functionality: PenguinVL vs Voxtral TTS. The overlap is only shared framework wiring and docs, which is not enough to treat them as the same change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "Both are model additions, but they implement different model families and code paths (PenguinVL vs DeepSeek-OCR-2). Shared auto-registration and docs are boilerplate, not duplicate evidence." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model integrations with different modalities and implementation details. They are not plausibly mergeable into one PR 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": 17, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "04945a5ba076074e47e1e356f76ff4936fa82738d4af99bdbae2655ff9ca8f40", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4406, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4406, + "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 conceptually related around module/kernel fusion, but they address different layers: one adds a generic Module Fusion API, the other implements n-to-1 kernel fusion via KernelConfig and model-loading integration.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45363 is the more concrete, end-user-facing implementation: it wires fusion through KernelConfig and model loading across multiple integration files, whereas #44979 is a lower-level API scaffold for module fusion.", + "best_issue_reason": null, + "best_pr_reason": "PR #45363 is the best representative of the cluster because it targets a specific fusion mechanism and concrete code paths; #44979 is broader infrastructure and not the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related subsystem, but not the same underlying change: #44979 adds a general Module Fusion API, while #45363 implements n-to-1 kernel fusion through KernelConfig and model integration. They would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "68467cf68eee70c106b636b9f24be93bb1bbefd9027d5d1e1d1f8556f01d9d00", + "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 fix the same Mistral regex patch AttributeError in the same file and target the same issue; 45317 is the stronger representative because it includes tests and a fuller patch.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:45317 is the better canonical PR because it covers the same fix as 45086 but adds test coverage and appears to be the more complete implementation.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:45317 is the best PR representative since it addresses the same concrete bug, is more complete, and is more suitable as the merged duplicate target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Accept: both PRs patch the same AttributeError in _patch_mistral_regex, touch the same code path/file, and are plausibly mergeable into one PR; 45317 is just the more complete 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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "28bd75a0e7a00f69a785681104862b0b7acfdebcd6bea13f8389a5a0ffe9b692", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch NemotronH internals, but they are not duplicates: one is a refactor/conversion-mapping update for GraniteMoeHybrid compatibility, while the other introduces dedicated NemotronHDense/NemotronHSparse model variants, docs, and tests.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the broader, more representative change: it adds the new dense/sparse model split, documentation, auto-mapping updates, and tests, whereas #45115 is a narrower compatibility/refactor patch.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the best standalone representative for the NemotronH work because it covers the actual feature/model surface change rather than only the conversion and inheritance refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They overlap on NemotronH files, but the concrete changes differ: #45115 focuses on conversion-key renaming and GraniteMoeHybrid-style inheritance, while #45569 adds separate dense/sparse model variants plus docs/tests. Same subsystem, 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": 20, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "fe64f2824af359f9044892150d7fa09c31366e290213da6c119c29dc143c9316", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related feature PRs for heterogeneous model/config support, but they are not duplicates: one adds per-layer config plumbing, the other adds broader modeling/runtime support on top of that.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the broader, more representative change: it covers heterogeneous modeling/runtime integration across cache, masking, model utils, and GPT-OSS code, not just config serialization/plumbing.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the better cluster representative because it spans the end-to-end heterogeneous support path and subsumes the narrower config-only groundwork in #45333.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related but not the same underlying change: #45333 focuses on per-layer config support, while #45332 extends that into model/runtime behavior. They are complementary pieces of a feature, not duplicate PRs that could be safely merged into one without losing scope separation." + } + ] + }, + "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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "4aafcdcbd3673122ffdb763b4dc94f9ce85fb34f2c183204b43494bea4461a81", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are about type annotations/type-checking, but they address different changes: one is a broad typing sweep across many files, the other adds specific typing protocols/aliases for modeling utilities and PEFT-related code. They are related in theme, not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45415 is the broader repository-wide typing sweep and is the better representative if this cluster is grouped by topic, but it is not the same concrete change as 45425.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the more focused, concrete typing-infrastructure change; it is a better standalone representative than the broad type-check cleanup in 45415, but still not a duplicate of it.", + "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 while 45425 adds specific protocols/aliases for different code paths; they do not fix the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45487-4", + "nodes": [ + "pull_request:45487", + "pull_request:45817", + "pull_request:45836", + "pull_request:45843" + ], + "soft_pairs": [ + "pull_request:45487|pull_request:45843", + "pull_request:45487|pull_request:45836", + "pull_request:45817|pull_request:45843", + "pull_request:45836|pull_request:45843" + ], + "prepared_review_unit_hash": "ba76cf9517786da69ba67cd2809dc680236c43e87c97904d7d3543c5798ae770", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 10388, + "estimated_input_tokens": 2597, + "estimated_eval_tokens": 5450 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 10388, + "estimated_input_tokens": 2597, + "estimated_eval_tokens": 5450 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all loosely about model-parallel or model-related test issues, but they target different models and different failure modes. They should not be deduplicated into one change.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45843 is the best representative because it contains substantive source changes for a real model-parallel fix, while the others are either different model families or test-skips only.", + "best_issue_reason": null, + "best_pr_reason": "PR 45843 is the strongest standalone PR in the set: it fixes a concrete model-parallel implementation issue in DeIMv2 and is broader/more central than the test-only or model-specific fixes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45487", + "right": "pull_request:45843", + "accept": false, + "reason": "Both mention model parallel, but they fix different model families and different code paths (AltCLIP/ChineseCLIP token type handling vs DeIMv2 backbone no-split propagation)." + }, + { + "left": "pull_request:45487", + "right": "pull_request:45836", + "accept": false, + "reason": "45487 is a source-code model-parallel fix; 45836 only skips failing tests for MiniCPM-V 4.6/Qwen3.5 cache compatibility. Not the same change." + }, + { + "left": "pull_request:45817", + "right": "pull_request:45843", + "accept": false, + "reason": "Gemma4 and DeIMv2 are unrelated implementations; these are separate model-parallel bugs, not one shared underlying defect." + }, + { + "left": "pull_request:45836", + "right": "pull_request:45843", + "accept": false, + "reason": "One is a test skip for invalid quant-cache coverage, the other is an implementation fix for model parallelism. Different underlying issues." + } + ] + }, + "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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "3667855bcea49dca5032169456bd7a1909d91eabe8b34f76490abc2d0cb844af", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4505, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4505, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs both touch expert-parallel loading, but they fix different failure modes: one targets Accelerate/DeepSpeed ZeRO-3 config handling, the other targets FSDP2/rank-0 broadcast overwriting experts. They are related in subsystem, not the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the stronger canonical PR because it addresses a concrete overwrite bug during EP model loading with broader code-path coverage and a clearer end-to-end fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the best representative of the cluster since it fixes a specific runtime corruption issue in the loading path, whereas #45548 is a narrower config/launch integration fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Both concern expert parallelism, but #45548 fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch, while #45662 fixes EP + FSDP2 experts being overwritten by rank-0 broadcast. Different concrete bugs and different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "e4d0fad74a13f12cbf26e2b5211882bb31651573dbade4a13548f2e743e1598e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist/captured_info collision in testing debug logs; PR 45645 is the broader, more complete change and subsumes the narrower 45639 fix.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the better canonical PR because it includes the core xdist-safe logging fix in testing_utils plus the CI workflow adjustment needed to consume multiple captured_info files.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 most fully represents the underlying change: it fixes the collision, preserves debug log collection in CI, and covers the same affected paths as 45639.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: xdist collisions for captured_info/debug logs. 45645 directly incorporates the same fix and adds the workflow follow-up, so they could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45683-2", + "nodes": [ + "pull_request:45683", + "pull_request:45734" + ], + "soft_pairs": [ + "pull_request:45683|pull_request:45734" + ], + "prepared_review_unit_hash": "eba4ca0f26b2d3be0c00e316270a9341d31af2208b44dbab31b1d3a8e81e7538", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4469, + "estimated_input_tokens": 1118, + "estimated_eval_tokens": 2492 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4469, + "estimated_input_tokens": 1118, + "estimated_eval_tokens": 2492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch the same quantization helper, but they fix different bugs: one special-cases audio submodules, the other changes how default skip lists are combined with user-supplied skips. They should not be clustered as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "45734 is the better representative PR because it fixes the core skip-module composition logic in `get_modules_to_not_convert`, which is broader and more central than the narrow audio-module special case in 45683.", + "best_issue_reason": null, + "best_pr_reason": "45734 is the strongest standalone PR in the set: it addresses a concrete quantization failure mode in the main code path and its fix is directly about the helper\u2019s behavior rather than a model-specific workaround.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45683", + "right": "pull_request:45734", + "accept": false, + "reason": "Both touch quantization skip logic, but 45683 adds audio-tower/embed_audio exclusions while 45734 changes additive handling of `skip_modules`; these are different underlying bugs 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": 26, + "component_count": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": 27, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45858-2", + "nodes": [ + "pull_request:45858", + "pull_request:45890" + ], + "soft_pairs": [ + "pull_request:45858|pull_request:45890" + ], + "prepared_review_unit_hash": "6e3fb1ec61a4376dc12129a1db8ff42f107e2641bbbb7194020f031c654f1640", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4140, + "estimated_input_tokens": 1035, + "estimated_eval_tokens": 2326 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4140, + "estimated_input_tokens": 1035, + "estimated_eval_tokens": 2326 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are XPU test-expansion changes, but they target different model test files and different concrete expectations, so they are related in theme but not duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No single PR is a clear duplicate representative here: each changes a different model test and adds distinct XPU expectations for a different code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45858", + "right": "pull_request:45890", + "accept": false, + "reason": "They both add XPU expectations, but for different models/tests (hy_v3 vs exaone4_5) and different expected outputs. This is a shared testing pattern, not the same underlying bug or change, and they would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/manifest.json b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..57452be6e31d1c5114a7cce51f17d2f5e70a87ed --- /dev/null +++ b/snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260511t120023z", + "artifacts": { + "hybrid": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-05-11T12:10:27Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260511T120023Z", + "variant": "hybrid" +} diff --git a/snapshots/20260511T120023Z/manifest.json b/snapshots/20260511T120023Z/manifest.json index 6fb548f36f22fb6b516811ea36019856b7816349..ebd889209d68238514b29849fdc473fde35d2603 100644 --- a/snapshots/20260511T120023Z/manifest.json +++ b/snapshots/20260511T120023Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 936 }, "extracted_at": "2026-05-11T12:00:23Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260511t120023z", + "runs": { + "hybrid-model-20260511t120023z": { + "analysis_id": "hybrid-model-20260511t120023z", + "artifacts": { + "hybrid": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260511T120023Z/analysis-runs/hybrid-model-20260511t120023z/manifest.json", + "model": null, + "published_at": "2026-05-11T12:10:27Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260511T120023Z", "watermark": {