diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index be8ee6a65b5926c94d770b7a7cb198be23591a44..8d563ac2f7b11dbc60a5b82a4314e0fc0c944a05 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": "20260617T180027Z", - "generated_at": "2026-06-17T18:13:20Z", + "snapshot_id": "20260618T000034Z", + "generated_at": "2026-06-18T00:09:35Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1739,6 +1739,77 @@ } ] }, + { + "cluster_id": "cluster-46562-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46562, + "issue_numbers": [], + "pr_numbers": [ + 46562, + 46658 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46562, + "right_pr_number": 46658, + "code_similarity": 0.168, + "size_similarity": 0.401, + "file_overlap": 0.032, + "area_overlap": 0.0, + "patch_similarity": 0.523, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 441, + 447 + ], + [ + 591, + 598 + ], + [ + 823, + 829 + ], + [ + 883, + 889 + ], + [ + 1082, + 1088 + ], + [ + 1205, + 1211 + ] + ], + "right_ranges": [ + [ + 362, + 368 + ], + [ + 795, + 801 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-44018-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -2123,6 +2194,15 @@ ], "target_issue_number": 46305, "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46562-2", + "canonical_pr_number": 46562, + "duplicate_pr_numbers": [ + 46658 + ], + "target_issue_number": null, + "reason": "PRs in cluster-46562-2 are treated as duplicates because they share soft_similarity evidence." } ], "best_issue": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 9d18b66d76947c4dd0bdfda38d2917b05983493f..a381346b4d56ac13c0eb962994198b8a8e36b533 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": "20260617T180027Z", - "generated_at": "2026-06-17T18:13:20Z", + "snapshot_id": "20260618T000034Z", + "generated_at": "2026-06-18T00:09:35Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -130,334 +130,328 @@ "issue:45237|issue:46015", "issue:43039|issue:44805" ], - "prepared_review_unit_hash": "83352bdae897901beb4dfd76982b3f7860b33d589783d26723b4d1f3a2dab196", + "prepared_review_unit_hash": "7c86c1552f899ba60896cee129b60c46ec4b166293331562afb6bb8f15ec6dd4", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 62, - "serialized_chars": 45721, - "estimated_input_tokens": 11431, - "estimated_eval_tokens": 23118 + "serialized_chars": 45731, + "estimated_input_tokens": 11433, + "estimated_eval_tokens": 23122 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated issue reports with only broad thematic overlap. The clearest duplicates are the DeepSpeed ZeRO-3 BertModel index error pair, and a likely torch 2.4.1 incompatibility pair. No PRs are present.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43596 is the best canonical issue because it is the broader, cleaner statement of the ZeRO-3/BertModel index error; issue:43638 reads like the same bug on a narrower non-pretrained Bert variant.", + "summary": "This cluster is mostly heterogeneous runtime/docs/model-compatibility issues. I found one clear duplicate pair: the two DeepSpeed ZeRO-3 + BERT index-out-of-bounds reports (43596/43638). The rest are only loosely similar and should stay separate.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43596 is the cleanest representative of the only true duplicate pair in the set: both it and 43638 report the same DeepSpeed ZeRO-3 / BERT index-0-out-of-bounds failure during model init.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is the strongest representative of an actual duplicate cluster in this set and has the clearest shared bug description with issue:43638.", + "best_issue_reason": "issue:43596 is the best issue anchor because it captures the duplicate subcluster most clearly and is the cleaner title of the matching pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different models and failure modes: missing Qwen3_5 config field vs chat/completions streaming error." + "reason": "Different failures: Qwen3.5 config attribute access vs chat/completions streaming inference error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different bugs: AMD GPU failure for GPT-OSS vs an MPS histogram operator issue." + "reason": "Different models/backends: GPT-OSS on AMD GPUs vs MoE MPS histogram unsupported for Int." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different subsystems and symptoms: image preprocessor dtype mismatch vs GPU-specific model execution failure." + "reason": "Different problem classes: image preprocessor dtype mismatch vs GPT-OSS AMD GPU failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS-on-AMD." + "reason": "Autocast dtype mismatch is not the same bug as GPT-OSS-20B on AMD GPUs." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Different models and likely different broken paths; too broad to treat as the same bug." + "reason": "One is GPT-OSS on AMD GPUs; the other is a broken Nemotron example. Not the same underlying issue." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Not the same underlying issue: AMD execution failure vs invalid method call in GraniteMoEHybrid." + "reason": "Different model-specific failures; no shared concrete code path." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "Collective mismatch under DeepSpeed Zero3 is unrelated to ONNX export errors." + "reason": "Deepspeed collective mismatch vs ModernBERT ONNX export error are unrelated." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3Moe FineGrainedFP8 are unrelated problems." + "reason": "Both are model export/runtime issues, but different models and different failure modes." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export failure and CUDA assert with auto device mapping are different code paths." + "reason": "ModernBERT ONNX export and Qwen3 auto device mapping are unrelated." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crashes and outdated RAG docs are unrelated." + "reason": "Both concern docs/examples, but they target different features and different doc problems." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are v5 docs cleanup items, but they concern different removed references and are not the same issue." + "reason": "Both are v5 docs cleanup issues, but they reference different removed commands/tasks and are not the same change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export error is unrelated to the GenerationMixin import regression." + "reason": "ModernBERT ONNX export and GenerationMixin import error are different regressions." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT embedding unpacking and resume-training/trust-remote-code are different issues." + "reason": "Different trainer/resume scenarios and different reported failures." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask shape mismatch is not the same as Gemma4 device mismatch during forward." + "reason": "Generic mask/tensor shape mismatch is not the same as Gemma4 device-map forward mismatch." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both mention index errors, but the reported failing tensor shapes and setups differ." + "reason": "Both are index/shape errors, but the model paths and root causes differ." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "DeepSpeed ZeRO-3 Bert index error is unrelated to resume-training with remote code." + "reason": "DeepSpeed ZeRO-3 BERT init error is unrelated to resume-training with trust_remote_code." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different error classes and contexts; no evidence they share the same root cause." + "reason": "Similar error family, but not the same code path or model setup." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask/index shape mismatch is unrelated to Qwen3.5 video chat template token mismatch." + "reason": "Mask shape mismatch and video token/feature mismatch are different bugs." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different failures: mask indexing error vs GPT-OSS TP-only issue." + "reason": "Generic tensor mask mismatch is not the same as GPT-OSS MoE tensor-parallel failure." }, { "left": "issue:46165", "right": "issue:46468", "accept": false, - "reason": "Model resume/trust-remote-code issue is unrelated to the Nemotron example bug." + "reason": "Resume-training with remote code and a broken Nemotron example are unrelated." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration error and model resume problem are unrelated." + "reason": "Video StopIteration in chat template is unrelated to resume-training issues." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Same broad training area but different bugs and code paths." + "reason": "DeepSpeed ZeRO-3 BERT init failure is unrelated to resume-training with remote code." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." + "reason": "BarkProcessor voice_preset and ModernBERT ONNX export are unrelated." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "DeepSpeed collective mismatch and Bert Zero-3 index error are different failures." + "reason": "Different distributed failures; only superficial similarity in deepspeed wording." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same underlying DeepSpeed ZeRO-3/Bert index-out-of-bounds bug; 43638 is just a narrower Bert variant." + "reason": "Same underlying bug: DeepSpeed ZeRO-3 + BERT initialization hits the same index-0-out-of-bounds failure." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy being called is unrelated to GPT-OSS tensor-parallel failure." + "reason": "Liger Kernel cross_entropy call path is unrelated to GPT-OSS tensor-parallel issues." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Incorrect detokenization and LoRA merge/reload output mismatch are different bugs." + "reason": "Different model/tokenization issue vs LoRA merge/save/reload mismatch." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen3-8B NVFP4 runtime and LoRA merge/reload mismatch are unrelated." + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA reload inconsistency are unrelated." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Processor failure for Voxtral and remote-code model_class/config_class error are different issues." + "reason": "Processor failure for Voxtral and remote-code config_class error are different problems." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed weight dimensionality error and collective mismatch are not the same bug." + "reason": "Both mention DeepSpeed, but one is a weight-shape RuntimeError and the other is a collective mismatch." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation checkpointing/FSDP recomputation mismatch is unrelated to mask shape indexing." + "reason": "Activation checkpointing/FSDP recompute mismatch is not the same as a generic mask/index shape error." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Both touch Qwen3_5MoeTextRotaryEmbedding, but one is CPU-offload incompatibility and the other is uninitialized-memory after to_empty materialization." + "reason": "Same component family, but different reported bugs: CPU offload incompatibility vs uninitialized memory after materialization." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index error and GPT-OSS TP-only failure are unrelated." + "reason": "BERT ZeRO-3 init failure and GPT-OSS MoE TP failure are unrelated." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "FSDP activation checkpointing mismatch is unrelated to CustomTrainer compute_loss signature breakage." + "reason": "FSDP checkpointing recompute mismatch and CustomTrainer.compute_loss signature break are unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different models and symptoms; not the same underlying bug." + "reason": "Different models and code paths; no duplicate evidence." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Mistral dtype mismatch and GraniteMoEHybrid invalid method are different issues." + "reason": "Different model-specific device/dtype issues." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different Qwen/GPT-OSS failures with no shared concrete code path." + "reason": "Different GPT/Qwen model/runtime failures and different hardware paths." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Gemma4 inputs_embeds reverse lookup and Qwen3.5 video token mismatch are unrelated." + "reason": "Different multimodal bugs: reverse embedding lookup vs video feature/token mismatch." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 sliding_window_pattern regression is unrelated to the Mistral regex patch crash." + "reason": "Gemma3 config attribute error and Mistral regex patch crash are unrelated." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Gemma4 device mismatch and Gemma4 reverse-embedding lookup are different bugs." + "reason": "Gemma4 device mismatch and reverse embedding lookup are different bugs." }, { "left": "issue:45800", "right": "issue:46515", - "accept": true, - "reason": "Both report the same torch 2.4.1 incompatibility; one is just a narrower versioned restatement." + "accept": false, + "reason": "Both mention torch compatibility, but they are different incompatibility reports." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "General torch incompatibility and Qwen-3-8B NVFP4 runtime failure are not the same issue." + "reason": "Torch 2.10 incompatibility and Qwen-3-8B-NVFP4 runtime error are unrelated." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP checkpointing mismatch and auto device mapping CUDA assert are unrelated." + "reason": "FSDP activation checkpointing and Qwen3 auto device mapping are different failures." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to an MPS int histogram operator error." - }, - { - "left": "issue:43638", - "right": "issue:45161", - "accept": false, - "reason": "Duplicate of a different pair's reasoning: Bert ZeRO-3 index error is not the same as GPT-OSS TP-only failure." + "reason": "Autocast dtype mismatch is not the same as MPS histogram not implemented for Int." }, { - "left": "issue:30277", - "right": "issue:36331", + "left": "issue:43828", + "right": "issue:44292", "accept": false, - "reason": "DeepSpeed collective mismatch and custom trainer signature change are unrelated." + "reason": "Different models and different failure modes." }, { - "left": "issue:34928", - "right": "issue:36010", + "left": "issue:43541", + "right": "issue:43716", "accept": false, - "reason": "Activation checkpointing/FSDP issue is unrelated to GenerationMixin import failure." + "reason": "Mixtral torch dynamo grouped_mm tracing error is unrelated to image preprocessor dtype mismatch." }, { - "left": "issue:45237", - "right": "issue:46015", + "left": "issue:43716", + "right": "issue:46468", "accept": false, - "reason": "GPT-OSS AMD failure and outdated RAG examples are unrelated." + "reason": "Different model families and unrelated failures." }, { - "left": "issue:43039", - "right": "issue:44805", + "left": "issue:43716", + "right": "issue:45685", "accept": false, - "reason": "Liger Kernel cross-entropy call and mask/index shape mismatch are different code paths." + "reason": "Different model/backend issues; no same underlying bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted duplicates are well supported by the issue titles/body themes (43596/43638 and 45800/46515), and the rejected soft-edge pairs are mostly appropriately marked as unrelated. No PR pair issue is present." + "feedback": "Grounded overall: the summary stays conservative, and the only accepted soft pair (issue:43596/issue:43638) is well supported by the titles. The other verdicts mostly avoid overclaiming and do not look like obvious duplicates from the packet." }, "evaluator_used": true, "retried": false, @@ -552,8 +546,8 @@ "issue:43854|issue:43866", "issue:43425|issue:45070", "issue:43975|issue:45901", - "issue:43596|issue:45800", "issue:43792|issue:45070", + "issue:43596|issue:45800", "issue:45362|issue:45464", "issue:44292|issue:45901", "issue:39290|issue:45823", @@ -570,262 +564,274 @@ "issue:43296|issue:43366", "issue:44387|issue:45005" ], - "prepared_review_unit_hash": "bc2807fc994148c1894b1e09e78ca9d84dde1d26b86b5b766012a460d8cbd7d2", + "prepared_review_unit_hash": "26b5b8b76dd8f0aa7f4295f1f0d464629f541f93743428f673d7b3e955fd9d0a", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39793, - "estimated_input_tokens": 9949, - "estimated_eval_tokens": 20154 + "serialized_chars": 39792, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose set of unrelated bug reports, with only a couple of clear duplicate-like pairs. The strongest duplicate pair is the AutoTokenizer v5 regression reports; the DeepSpeed ZeRO-3 Bert index errors also look like the same underlying issue.", - "confidence": 0.58, - "canonical_issue_reason": "issue:45920 is the broadest and most central issue here: it describes the AutoTokenizer v5 regression across multiple model families, and it subsumes the narrower Granite-only report.", + "summary": "This cluster is mostly a set of unrelated issue reports; only two soft pairs look like true duplicates: the AutoTokenizer v5 token-ID regression and a DeepSpeed ZeRO-3 index-out-of-bounds loading failure.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45920 is the broad, current umbrella for the AutoTokenizer v5 token-ID regression across multiple model families, so it best represents the duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the best representative issue because it is open, broader in scope than the Granite-specific report, and matches the shared tokenizer-regression theme most closely.", + "best_issue_reason": "issue:45920 is the most useful triage target: open, broad, and already framed as the shared tokenizer regression rather than a narrow model-specific instance.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Different models and failure modes: checkpoint corruption vs missing text encoder weights during loading." + "reason": "Different model families and failure modes: corrupted-checkpoint/load failure vs missing text-encoder weight loading." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Both mention version compatibility, but they are different incompatibilities with different symptoms and code paths." + "reason": "Both are compatibility-related, but one is a torch version incompatibility and the other is support for USE_DISTRIBUTED=0." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL loading failure in vLLM is unrelated to the USE_DISTRIBUTED=0 PyTorch support breakage." + "reason": "PaddleOCR-VL/vLLM model load failure vs distributed-build support regression; unrelated code paths." }, { "left": "issue:41762", "right": "issue:43596", "accept": true, - "reason": "Both report DeepSpeed ZeRO-3 loading-time IndexErrors on BERT-like models, likely the same underlying initialization bug." + "reason": "Same DeepSpeed ZeRO-3 init/loading failure class: index 0 out of bounds from empty/size-0 tensors while loading models." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 config attribute mismatch with vLLM and Qwen3 auto-device-map cuda asserts are different bugs." + "reason": "Different bugs: Gemma3 sliding_window_pattern config issue vs Qwen3 MoE auto device-map CUDA assert." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are docs-related, but they concern different removed/changed docs content and not the same issue." + "reason": "Both are docs-related, but they cover different APIs and different stale references." }, { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the distributed-build compatibility regression." + "reason": "Float8 storage lookup error vs distributed support regression; not the same underlying bug." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "AMD GPU support failure and table-question-answering crashes are unrelated." + "reason": "AMD GPU support failure vs table-question-answering task crash; unrelated." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross-entropy dispatch and Qwen3-VL video StopIteration are different problems." + "reason": "Liger cross-entropy path issue vs Qwen3-vl video StopIteration; different subsystems." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Detokenization regression and outdated RAG examples are unrelated." + "reason": "Tokenizer detokenization regression vs outdated RAG examples; unrelated." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Gemma4 device-mapping mismatch and Qwen3.5 video token/feature mismatch are different model-specific bugs." + "reason": "Gemma4 device-mapping mismatch vs Qwen3.5 video/chat-template mismatch; different code paths." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral and Gemma4 device mismatch are unrelated." + "reason": "Processor failure for Voxtral vs Gemma4 PLE device mismatch; not the same issue." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and RAG docs examples are unrelated." + "reason": "NemotronH checkpoint-loading bug vs docs/examples update; unrelated." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote-code model breakage and USE_DISTRIBUTED=0 support are different regression classes." + "reason": "Remote-code model breakage vs distributed-build support regression; different failures." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Both describe the same AutoTokenizer v5 wrong-token-ID regression; 45812 is just a narrower Granite-specific instance of the broader issue." + "reason": "Same AutoTokenizer v5 token-ID regression; 45920 is the broader umbrella report spanning more affected model families." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Checkpoint corruption and checkpoint loading failure for a different model are not the same bug." + "reason": "Corrupted checkpoint/load issue vs NemotronH implementation cannot load checkpoints; different models and symptoms." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes Int8Params API incompatibility is unrelated to distributed support breakage." + "reason": "bitsandbytes Int8Params API incompatibility vs distributed support regression; unrelated." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and outdated RAG examples are unrelated." + "reason": "Qwen-3-8B-NVFP4 runtime error vs RAG docs issue; unrelated." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window, but they are different model families and different concrete failures." + "reason": "Both mention sliding_window, but Gemma3TextConfig/vLLM and Qwen3-MoE are different model/config bugs." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexErrors, but the shapes, triggers, and model contexts differ too much to treat as the same bug." + "reason": "Both are ZeRO-3 index errors, but one is a mask/tensor shape mismatch and the other is an empty-index out-of-bounds during BertModel init." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper loading failure and table-question-answering crash are unrelated." + "reason": "Whisper model load failure vs table-question-answering crash; unrelated." }, { - "left": "issue:43975", - "right": "issue:45901", + "left": "issue:43854", + "right": "issue:43866", "accept": false, - "reason": "Detokenization and table QA runtime failure are unrelated." + "reason": "GLM-4.7-Flash test loading failure vs Ovis2 checkpoint corruption; different model issues." }, { - "left": "issue:43596", - "right": "issue:45800", + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 incompatibility vs pydantic PretrainedConfig field regression; different incompatibilities." + }, + { + "left": "issue:43975", + "right": "issue:45901", "accept": false, - "reason": "DeepSpeed ZeRO-3 BERT loading issue and torch/transformers incompatibility are different regressions." + "reason": "Tokenizer detokenization bug vs table-question-answering crash; unrelated." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper model failure and pydantic PretrainedConfig regression are unrelated." + "reason": "Whisper load failure vs pydantic model field regression; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Both are version-sensitive issues, but ZeRO-3 empty-index init failure and torch/transformers compatibility are not the same bug." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both touch Qwen3.5, but one is a general chat crash and the other is a streaming API failure; not enough evidence they are the same bug." + "reason": "Different Qwen3.5 failures: chat crash vs streaming inference API failure." }, { "left": "issue:45685", "right": "issue:46304", "accept": false, - "reason": "MPS histogram int error and distributed-build support regression are unrelated." + "reason": "MPS histogram int error in MoE vs distributed support regression; unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Grouped_mm tracing failure in Mixtral and MPS histogram integer error are different code paths." + "reason": "Both are MoE-related, but grouped_mm tracing and MPS histogram failures are different code paths." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3 sliding_window_pattern missing are unrelated." + "reason": "ImportError for GenerationMixin vs Gemma3TextConfig sliding_window_pattern; unrelated." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both involve MoE/dtype issues, but the concrete failure mechanisms and model paths differ." + "reason": "Grouped_mm tracing failure vs autocast dtype mismatch; different MoE bugs." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Untied lm_head weights in Mistral3 and a corrupted Ovis2 checkpoint are unrelated." + "reason": "Tied lm_head weights in Mistral3 vs Ovis2 checkpoint corruption; unrelated." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure and 'Can't compile non template nodes' are unrelated." + "reason": "Processor load failure vs compile non-template nodes; unrelated." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are IndexErrors, but they involve different shapes, contexts, and likely different root causes." + "reason": "Both are ZeRO-3 index errors, but the reported mechanics differ: mask-shape mismatch vs empty-index out-of-bounds." }, { "left": "issue:45399", "right": "issue:46661", "accept": false, - "reason": "Flash-attn fallback gating and remote-code config_class AttributeError are unrelated." + "reason": "flash-attn fallback gating vs remote-code model_class/config_class error; unrelated." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and a pydantic PretrainedConfig field regression are different issues." + "reason": "StableLm pad_token_idx regression vs pydantic PretrainedConfig field regression; different config bugs." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch compatibility breakage and Gemma4 device-mapping mismatch are unrelated." + "reason": "Torch/transformers compatibility issue vs Gemma4 device mismatch; different failures." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL loading and GGUF gpt-oss support are different model-support requests/problems." + "reason": "PaddleOCR-VL/vLLM loading failure vs GGUF gpt-oss support request; unrelated." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 memory/OOM regression and tied-weights translation-model issue are unrelated." + "reason": "Int4 CUDA memory regression vs tied-weights translation-model issue; different bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays mostly conservative. The open, broader AutoTokenizer regression issue:45920 is a reasonable canonical representative for issue:45812, and the other pairwise judgments are framed cautiously. The DeepSpeed ZeRO-3 pair is a bit speculative, but not overstated enough to reject." + "feedback": "The triage is grounded in the packet: the canonical issue choice is well-supported by the open, broader AutoTokenizer v5 regression report, and the soft-edge verdicts stay conservative overall. The two accepted pairs both have strong title-level alignment (shared exact regression/error class), while the rejected pairs are mostly kept separate for sensible model- or subsystem-specific differences." }, "evaluator_used": true, "retried": false, @@ -921,26 +927,26 @@ "issue:44360|issue:44512", "issue:44863|issue:45005", "issue:44492|issue:44512", - "issue:45183|issue:46463", "issue:44918|issue:45507", + "issue:45183|issue:46463", "issue:46489|issue:46491", "issue:43975|issue:44292", "issue:44292|issue:45507" ], - "prepared_review_unit_hash": "fbc5f6c6d69ca733621b9e8ece249db0313285c6572fc61de221fd4b88f087e3", + "prepared_review_unit_hash": "5b3e321af2af67b518dfbbf07c0092ec498de914d803c8df5e35d584e9832298", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36256, + "serialized_chars": 36254, "estimated_input_tokens": 9064, "estimated_eval_tokens": 18384 }, @@ -949,207 +955,207 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers/model-specific reports. The only clearly merge-worthy duplicate-like thread is the Qwen3.5 video-input regression pair; the remote-code and input-validation pairs are too inference-heavy to accept conservatively.", - "confidence": 0.57, - "canonical_issue_reason": "Issue 46294 is the strongest anchor because it is open, concrete, and directly matches the accepted Qwen3.5 video-input/token mismatch thread with 44479.", + "summary": "A few pairs look like the same underlying regression, but most are only superficially similar and should remain separate. No PRs are present.", + "confidence": 0.67, + "canonical_issue_reason": "issue:44479 is the clearest representative of an accepted duplicate-like set: it describes the broad v5.3.0 video-input regression across several Qwen VL / Qwen3.5 variants and names the concrete broken path.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46294 is the best representative issue: it is specific enough to stand for the only solid duplicate-like subthread in the set and is more actionable than the broader sibling report.", + "best_issue_reason": "Best overall issue to anchor triage because it states a concrete regression with a wide affected surface and is the most representative of the accepted video-input bug family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both are IndexError-flavored, but the titles point to different surfaces: generic mask-shape mismatch vs TRL/Qwen3.5 embedding unpacking." + "reason": "Both mention mask/shape failures, but one is a generic mask-index mismatch and the other is a TRL SFT/input-embedding failure on Qwen3.5; too little evidence of the same root cause." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "MoE is the only real overlap; grouped_mm torch-dynamo tracing and GPT-OSS tensor-parallel failure look like different bugs." + "reason": "Different models and different failures: grouped_mm tracing error in Mixtral vs tensor-parallel-only failure in GPT-OSS MoE." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch and chat/completions streaming failure are different code paths and symptoms." + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a streaming chat/completions failure on Qwen3.5-0.8B. Different code paths." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support is unrelated to the Qwen3.5 streaming inference failure." + "reason": "GGUF architecture support for gpt-oss is unrelated to a Qwen3.5 streaming API failure." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "init_empty_weights constructor-arg regression and PretrainedConfig pydantic break are different underlying issues." + "reason": "`init_empty_weights`/_is_hf_initialized construction error is a different bug from the pydantic `PretrainedConfig` field regression." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and the pydantic config regression do not describe the same bug." + "reason": "Qwen3-MoE sliding-window behavior and a pydantic model-field regression are unrelated." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Grouped_mm tracing failure and missing Float8 storage are distinct runtime problems." + "reason": "Grouped_mm tracing on Mixtral and missing Float8 storage deserialization are different runtime failures." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Both are multimodal, but image-preprocessor dtype mismatch is not the same as video feature/token mismatch." + "reason": "Mistral-3 image-preprocessor dtype mismatch is unrelated to Qwen3.5 video token/feature mismatch." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code breakage and torch/transformers version incompatibility are too broad to treat as one bug." + "reason": "Remote-code breakage and torch/transformers version incompatibility are too broad to be the same concrete bug." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Mask-size mismatch and deepspeed zero3 index-out-of-bounds are different failures." + "reason": "Both are index errors, but one is a mask-length mismatch and the other is a zero3/BertModel empty-tensor index error; different failure modes." }, { "left": "issue:44479", "right": "issue:46294", "accept": true, - "reason": "Both describe the same Qwen3.5-family video-input regression, with the same video feature/token mismatch theme." + "reason": "Same Qwen video-input regression family: both describe mismatches between video features and video tokens/chat templating on Qwen3.5 variants." }, { "left": "issue:43054", "right": "issue:43994", - "accept": false, - "reason": "Both involve SigLIP2 quality, but the reported symptoms and likely root causes differ." + "accept": true, + "reason": "Both are SigLIP2-specific reports of bad/nonsensical outputs, suggesting the same underlying model implementation regression." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Composite-config unwrapping and uninitialized rotary embedding memory are unrelated bugs." + "reason": "Composite-config unwrapping in AutoModelForCausalLM and uninitialized rotary embedding memory after to_empty() are different bugs." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 and remote-code config_class errors are unrelated." + "reason": "Flash Attention 2 integration and remote-code model_class/config_class lookup are separate failures." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Image-preprocessor dtype mismatch and TRL embedding unpacking are different code paths." + "reason": "Different models and symptoms: Mistral-3 image dtype mismatch vs Qwen3.5 input-embedding/TRL packing failure." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-dimension error is unrelated to the Qwen3.5/TRL unpacking failure." + "reason": "FSDP weight-shape initialization error is unrelated to Qwen3.5 input-embedding unpacking." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "High perplexity on gpt-oss and a GLM5 issue do not match as the same bug." + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are not the same bug." }, { "left": "issue:45632", "right": "issue:45698", - "accept": false, - "reason": "Both are remote-code/custom-module related, but the titles indicate different concrete failures: cache-path collision versus wrong module loaded after save_pretrained." + "accept": true, + "reason": "Both point to remote-code/custom-module resolution loading the wrong artifact because the cache/path identity is not unique enough." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Jupyter custom-model initialization and flash-attn2 fallback logic are unrelated." + "reason": "Notebook custom-model initialization in v5 and Flash-Attn2 fallback gating are unrelated." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but the model families and failure modes are too different." + "reason": "Untied lm_head weights in Mistral3 are not the same as translation-model tied-weight issues in v5." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3Moe FP8 failure and flash-attn2 fallback checks are different concerns." + "reason": "Qwen3-MoE FineGrainedFP8 failure and Flash-Attn2 fallback gating are different subsystems and failure modes." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is not the tied-weights translation-model issue." + "reason": "PagedAttentionCache linear_attention crash and tied-weights translation-model regressions do not share a concrete code path." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior and tied-weight handling are distinct bugs." + "reason": "Offloading/quantization behavior and tied-weight translation-model regressions are too far apart." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "DSA indexer ReLU discussion and a docs typo are unrelated." + "reason": "A DSA indexer ReLU discussion is unrelated to a documentation typo about the removed run command." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading and translation-model tied weights are different issues." + "reason": "NemotronH checkpoint loading and translation-model tied weights are different defects." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A cache-strategy typo and removed run-command docs are not the same underlying change." - }, - { - "left": "issue:45183", - "right": "issue:46463", - "accept": false, - "reason": "Both concern validation/error messaging, but the packet does not show a shared helper or the same underlying bug; the titles point to different validation surfaces." + "reason": "Cache-strategy typo and removed-doc-command mismatch are just documentation issues, not the same one." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 TRL embedding unpacking and GraniteMoEHybrid invalid method calls are unrelated." + "reason": "Qwen3.5 input-embedding unpacking and GraniteMoEHybrid invalid-method calls are unrelated." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": true, + "reason": "Both are input-validation error-message bugs in the shared decoder/input-checking path; T5 is a concrete instance of the broader XOR validation problem." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and different round-trip failures." + "reason": "Different tokenizer regressions on different model families; same general area, but not the same underlying bug." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Tokenizer detokenization and init_empty_weights argument handling are unrelated." + "reason": "DeepSeek-Coder v1 detokenization regression and an NVFP4 loading/runtime error are unrelated." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "init_empty_weights TypeError and GraniteMoEHybrid invalid method are different bugs." + "reason": "NVFP4 runtime failure and GraniteMoEHybrid invalid method are different bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The overall summary is mostly grounded, but at least one soft-edge accept overstates the evidence. In particular, issue:45183 vs issue:46463 are both about validation/error messaging, but the packet does not show they share a helper or the same bug; the titles point to different validation surfaces (input_ids/inputs_embeds XOR logic vs T5 decoder input validation). The remote-code pair issue:45632 vs issue:45698 is plausible, but also relies on inference beyond the packet. Because the instructions require conservative soft-edge verdicts, I can\u2019t accept this as-is." + "accept": true, + "feedback": "Grounded overall. The summary matches the packet, the accepted soft edges are plausible from the titles alone, and most non-matches are conservatively rejected. No PRs are present." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1237,9 +1243,9 @@ "issue:44403|issue:44451", "issue:43975|issue:45520", "issue:44403|issue:44488", - "issue:42947|issue:43856", "issue:24643|issue:28530", "issue:43643|issue:45020", + "issue:42947|issue:43856", "issue:41762|issue:42915", "issue:44991|issue:45399", "issue:43824|issue:43872", @@ -1252,244 +1258,244 @@ "issue:42491|issue:44368", "issue:43828|issue:45823" ], - "prepared_review_unit_hash": "3ef5615aab35e28f1292ae3021e53346bfff546f9b7d3138241b0fec8be9801d", + "prepared_review_unit_hash": "e555f838f81ea1a151821e8da191de576f1c86f0fb68392264a188b9d61cab82", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38113, - "estimated_input_tokens": 9529, - "estimated_eval_tokens": 19314 + "serialized_chars": 38112, + "estimated_input_tokens": 9528, + "estimated_eval_tokens": 19312 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous, but there is a small duplicate set around the transformers 5.x non-persistent-buffer regression. The strongest match is 43950 as the umbrella issue, with 44534 and 45902 fitting it closely.", - "confidence": 0.68, - "canonical_issue_reason": "43950 is the best canonical issue: it states the broad 5.x non-persistent-buffer corruption regression, and the nearby reports (44534, 45902) are direct variants of that same bug.", + "summary": "This cluster is mostly heterogeneous: many unrelated model-specific bugs, tokenizer/loading regressions, and training/runtime issues. The only clearly duplicate subcluster is the Transformers v5 non-persistent-buffer regression; the tokenizer mis-tokenization pair also looks like the same underlying v5 tokenizer bug family.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43950 is the clearest root report for the Transformers 5.x non-persistent-buffer regression, and it best anchors the strongest duplicate subcluster (with 43644/44534/45902).", "canonical_pr_reason": null, - "best_issue_reason": "43950 is the most reusable triage target because it captures the underlying buffer-materialization regression rather than just one model-specific symptom.", + "best_issue_reason": "issue:43950 is the best representative because it describes the core bug directly, rather than a downstream symptom or a broader follow-on report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying 5.x regression: non-persistent buffers are being filled/corrupted with junk." + "reason": "Same Transformers v5 non-persistent-buffer regression; 44534 is the same corruption/junk-buffer bug described in 43950." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but one is a doc/example mismatch+quantization issue while the other is a quality/regression complaint about text embeddings." + "reason": "Both mention SigLIP2, but one is a docs/example mismatch and quantization failure while the other is degraded text embeddings; not the same bug." }, { "left": "issue:43644", "right": "issue:43950", - "accept": false, - "reason": "Different bug class: remote_code config-field loss vs non-persistent buffer corruption." + "accept": true, + "reason": "Same v5 non-persistent-buffer corruption bug; 43644 and 43950 describe the same failure mode with different wording." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Unrelated failures: custom compute_loss signature mismatch vs DeepSpeed ZeRO-3 loading index error." + "reason": "Different failures: a custom `compute_loss` signature mismatch vs a DeepSpeed ZeRO-3 loading index error." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different incompatibilities and different error surfaces; not the same code-path bug." + "reason": "Different quantization/storage incompatibilities (`_is_hf_initialized` vs missing `Float8_e4m3fnStorage`), not the same code path." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Both involve dtype/position handling, but the concrete failures and affected paths differ." + "reason": "Different model/runtime problems: autocast dtype mismatch on Phi-tiny-MoE vs RLHF instability from Qwen3.5 3D `position_ids` handling." }, { "left": "issue:45920", "right": "issue:46489", - "accept": false, - "reason": "Related tokenizer regressions, but 45920 is a broad multi-model report while 46489 is a model-specific follow-up; not clearly the same duplicate." + "accept": true, + "reason": "Same tokenizer regression family in v5+; 46489 is a specific DeepSeek-Coder instance of the broader wrong-token-IDs bug." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption is not the same issue as noisy loading output." + "reason": "A corrupted checkpoint and a loading-noise issue are unrelated." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "One is a SigLIP2 embedding-quality issue; the other is a tokenizer change/regression across versions." + "reason": "One is a SigLIP2 embedding-quality complaint, the other is a tokenizer-version behavior change; too broad to be the same bug." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both concern SigLIP2 behavior, but the symptoms are different enough that they do not look like the same underlying bug." + "reason": "Both involve SigLIP2 output quality, but they describe different symptoms and no shared concrete failure path." }, { "left": "issue:43950", "right": "issue:45902", "accept": true, - "reason": "Same underlying non-persistent-buffer materialization bug; 45902 is a concrete model-specific manifestation." + "reason": "Same non-persistent-buffer materialization regression; 45902's uninitialized `inv_freq` is a concrete manifestation of 43950's bug." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different models and different failure modes: tokenizer loading error vs missing attribute on a causal LM." + "reason": "Unrelated issues: bad AutoTokenizer error for Voxtral vs missing `all_tied_weights_keys` on MolmoForCausalLM." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Both mention Qwen3.5/MoE, but illegal-memory-access from 3D position_ids is not the same as CPU-offload incompatibility." + "reason": "Different Qwen3.5 issues: FlashAttention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Both are about SigLIP2, but one is model quality and the other is implementation discrepancy with JAX." + "reason": "Related to SigLIP2, but one reports poor embeddings and the other reports HF/JAX parity mismatch; not clearly the same concrete bug." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is gradient checkpointing/LoRA behavior and the other is int4 quantization OOM." + "reason": "Different training issues: ineffective gradient checkpointing with PEFT LoRA vs higher reserved memory under int4 quantization." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states rebuild issue is unrelated to buffer initialization corruption." + "reason": "Different root causes: Gemma4 FSDP2 `shared_kv_states` KeyError vs Qwen3_5Moe `inv_freq` uninitialized memory." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and different classes of load-time failures." + "reason": "Different model-loading failures with no shared concrete code path." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "General loading noise vs a specific model load failure are not the same underlying bug." + "reason": "A generic loading-noise report and a specific tokenizer/model loading failure are not duplicates." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Tokenizer detokenization bug and Python 3.13 flash_attn import failure are unrelated." + "reason": "DeepSeek-Coder detokenization and a Python 3.13 flash_attn import error are unrelated." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Same as above: loading noise is not the same as a concrete model load failure." - }, - { - "left": "issue:42947", - "right": "issue:43856", - "accept": false, - "reason": "Different performance/memory problems with different triggers and code paths." + "reason": "Unnecessary load noise vs a specific model loading failure; not the same bug." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "Training DeepSpeed weight-shape error is unrelated to early stopping metric selection." + "reason": "Training with DeepSpeed causing a weight-dimension error is unrelated to early stopping metric resolution." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "The latter is a broad remote_code regression umbrella; the former is one specific symptom, but the shared root cause is not clear enough to call them duplicates." + "reason": "Both touch `remote_code`, but one is a missing-fields config bug and the other is a broad breakage claim; not the same concrete failure." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different memory-related training problems with different mechanisms." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 loading error and Qwen3 MoE FP8 failure are different issues." + "reason": "Gemma3 ZeRO-3 loading error and Qwen3 MoE/LoRA compatibility issue are unrelated." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure and flash-attn2 fallback gating are unrelated." + "reason": "Tokenizer loading failure and flash-attn fallback gating are separate subsystems and separate bugs." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Importing a missing model class and a bitsandbytes parameter incompatibility are separate problems." + "reason": "Importing a missing model class and a bitsandbytes Int8Params constructor incompatibility are unrelated." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram support error and Gemma4 device mismatch are not the same bug." + "reason": "MPS histogram implementation gap vs Gemma4 device-map mismatch are different errors." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only loading failure is unrelated to GLM5." + "reason": "Different model-specific loading issues with no shared code path." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model-loading failures with different missing pieces." + "reason": "Different model-loading failures: GLM4V load failure vs missing `all_tied_weights_keys`." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote_code regression umbrella vs Gemma4 PLE device mismatch are unrelated." + "reason": "Broad remote-code breakage and a Gemma4 device mismatch are not the same issue." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer-mapping registration failure is not the same as remote_code breakage." + "reason": "Tokenizer-mapping helper failure and remote_code breakage are different problems." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral tokenizer regex patch crash and flash-attn fallback gating are unrelated." + "reason": "Mistral tokenizer regex patch crash and flash-attn2 fallback gating are unrelated." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "LoRA compatibility across versions and tie_word_embeddings warning are not the same issue." + "reason": "Both mention Qwen3/MoE/LoRA, but one is a version-compatibility break and the other is a configuration warning." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Different incompatibilities: autocast dtype mismatch vs device placement mismatch." + "reason": "Different runtime/device issues: dtype mismatch vs device-map placement mismatch." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted duplicate set around issue 43950 is well-supported by the titles/body text, and the other soft-edge rejections do not overstate similarity." + "feedback": "Grounded overall: the cluster is clearly heterogeneous, and the duplicate calls are mostly conservative. The non-persistent-buffer v5 regression subcluster is well supported by the titles, and the tokenizer regression pair is also reasonably framed as a related v5-family issue rather than overclaimed as a universal duplicate." }, "evaluator_used": true, "retried": false, @@ -1571,15 +1577,15 @@ "issue:45020|issue:46661", "issue:43792|issue:44220", "issue:43257|issue:45305", - "issue:42947|issue:45923", "issue:41553|issue:44991", + "issue:42947|issue:45923", "issue:36331|issue:41093", "issue:43940|issue:45313", "issue:43854|issue:44863", "issue:44479|issue:44560", "issue:41553|issue:44488", - "issue:44928|issue:45588", "issue:41093|issue:44805", + "issue:44928|issue:45588", "issue:24643|issue:30064", "issue:44863|issue:46402", "issue:44530|issue:44863", @@ -1589,14 +1595,14 @@ "issue:45325|issue:45902", "issue:45663|issue:45823" ], - "prepared_review_unit_hash": "96534d11481da0029457fbb8d0d354d7ddb832732c2afda8d0446c47075e135d", + "prepared_review_unit_hash": "49c326d8de779fa026a1f13201a85b863a11a0777f31c6b98b9001c1b09ecdb2", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, @@ -1611,186 +1617,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous. I found a few true duplicate pairs: the v5 docs/pipeline removal references, the `_is_hf_initialized` loading TypeError, the Whisper fbank helper failure, the Qwen video-input regression, and the mask/tensor shape mismatch. The rest are related by subsystem but not the same underlying bug.", - "confidence": 0.69, - "canonical_issue_reason": "If a single representative must be chosen, issue 41093 is the cleanest canonical target: 44805 repeats the same mask/tensor shape mismatch with only the dimensions elided.", + "summary": "This set is mostly heterogeneous, but a few pairs look like true duplicates: stale v5 docs references, DeepSeek tokenizer regressions, the `_is_hf_initialized` init-empty-weights/typeerror path, and the identical mask-shape IndexError. Most other pairs are only loosely related by subsystem or model family and should stay separate.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here: it describes a general recent-version regression affecting remote-code models, which makes it the best anchor for the cluster even though most other items are narrower and not duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 41093 is the earliest and most exact report of the mask/indexing error, so it is the best duplicate anchor among the issues here.", + "best_issue_reason": "Issue 45020 is the most representative issue for the overall cluster because it captures the common theme of v5 regressions breaking previously working model-loading paths, while the rest are mostly model-specific edge cases.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both are transformer regressions, but 45020 is a broad remote_code breakage report while 45356 is a specific Kimi tokenizer codec/warning regression; not the same bug." + "reason": "Related regression theme, but not the same bug: one is a broad remote_code breakage, the other is a specific Kimi-K2.5 tokenizer/codec regression." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but they hit opposite conversions (`ids->tokens` vs `tokens->ids`) and different hot paths." + "reason": "Both are tokenizer performance regressions, but they hit different methods (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and different code paths." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs problem: v5 removed pipeline tasks, but the docs still reference them. 44509 is a later restatement of the same documentation mismatch." + "reason": "Same docs cleanup problem: both report stale references to pipeline/tasks removed in v5." }, { "left": "issue:45488", "right": "issue:46489", - "accept": false, - "reason": "Both involve tokenizer correctness on DeepSeek-family models, but the model families and failure modes differ enough that this does not look like one underlying bug." + "accept": true, + "reason": "Same tokenizer regression family: DeepSeek tokenization is broken because v5/LlamaTokenizer no longer respects the tokenizer.json pre-tokenizer behavior." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "One is a tied-weight bug in Mistral3; the other is a from_pretrained buffer regression. Related area, different code paths." + "reason": "Different bugs: tied `lm_head` weights in Mistral3 vs corruption of non-persistent buffers during `from_pretrained()`." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Both touch AutoModel config/loading, but one is about unwrapping `text_config` for composite configs and the other is a remote-code `config_class` attribute error; different defects." + "reason": "Both involve model-loading APIs, but one is config unwrapping for composite multimodal configs and the other is a remote-code `config_class` attribute failure." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same underlying `_is_hf_initialized` kwarg regression during model initialization/loading; the symptoms differ, but the bug is the same." + "reason": "Same underlying `_is_hf_initialized` incompatibility during empty-weight/model materialization." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Different failures: FineGrainedFP8/Qwen3Moe vs bitsandbytes `_is_hf_initialized` incompatibility." + "reason": "Both mention model-init/dtype/FP8-related failures, but the concrete bugs and code paths are different." }, { "left": "issue:45020", "right": "issue:46661", "accept": false, - "reason": "Both are broad loading regressions, but they affect different surfaces: remote_code model behavior vs missing `config_class` on AutoModel.from_pretrained." + "reason": "Both are broad model-loading regressions, but the specific failure modes differ: remote_code breakage vs missing `config_class` on `from_pretrained()`." }, { "left": "issue:43792", "right": "issue:44220", - "accept": true, - "reason": "Same Whisper audio preprocessing failure; 44220 names the specific `_torch_extract_fbank_features()` helper behind the generic 'can't run' report." + "accept": false, + "reason": "Too generic to merge confidently; the titles suggest a Whisper run failure and a specific fbank feature-extraction issue, but not the same concrete bug." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both involve DeepSpeed, but one is a weight-conversion/load issue and the other is gradient averaging with `model_accepts_loss_kwargs=True`; not the same bug." + "reason": "Both involve DeepSpeed and Qwen3-family models, but one is weight conversion/loading and the other is gradient averaging with `model_accepts_loss_kwargs=True`." }, { - "left": "issue:42947", - "right": "issue:45923", + "left": "issue:41553", + "right": "issue:44991", "accept": false, - "reason": "Both mention gradient checkpointing, but one is about PEFT LoRA not taking effect and the other is a missing support flag on a trust_remote_code model." + "reason": "Both are tokenizer/loading regressions, but they affect different models and different failure points." }, { - "left": "issue:41553", - "right": "issue:44991", + "left": "issue:42947", + "right": "issue:45923", "accept": false, - "reason": "Both are tokenizer loading problems, but they concern different models and different failure modes." + "reason": "Both touch gradient checkpointing, but one is a PEFT/LoRA effectiveness bug and the other is a missing capability flag on a trust-remote-code model." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "`compute_loss` signature mismatch vs a tensor/mask indexing error; unrelated bugs." + "reason": "Different trainer/runtime failures: an unexpected `num_items_in_batch` kwarg versus a tensor mask-shape mismatch." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are ZeRO-3 weight-loading failures in Qwen models, but the model targets and missing-weights patterns differ; not enough to treat as one bug." + "reason": "Similar DeepSpeed ZeRO-3 symptom, but the model variants and reported missing-weight patterns differ enough that this is not safely the same bug." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different model-loading failures for different architectures; only the general symptom is shared." + "reason": "Both are checkpoint/model-loading failures, but they concern different architectures with different load paths." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Same 5.3.0 Qwen VL video-input regression; 44560 is a specific StopIteration manifestation of the broader video path failure." + "accept": false, + "reason": "Both are video-input regressions, but they affect different model families and different processing stages." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer-related, but one is a bad error message for Voxtral and the other is a specific model-loading failure for sleng-bert." - }, - { - "left": "issue:44928", - "right": "issue:45588", - "accept": false, - "reason": "Both mention Qwen3.5/attention, but one is an RLHF NaN/SDPA fallback issue and the other is a flash-attention crash on `s_aux=None`." + "reason": "Both are tokenizer/model-loading issues, but the failures and targets are different." }, { "left": "issue:41093", "right": "issue:44805", "accept": true, - "reason": "Same mask/tensor length mismatch in indexing; 44805 is essentially the same error with the concrete sizes elided." + "reason": "Same underlying IndexError pattern: a mask/tensor length mismatch at index 0 with the same shape-mismatch symptom." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Different problems: RLHF NaNs from Qwen3.5 position-id/SDPA behavior versus a flash-attention AttributeError on `s_aux=None`." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Training with DeepSpeed and void segmentation map preprocessing are unrelated problems." + "reason": "Completely unrelated: DeepSpeed training weight-shape error versus image-processor void segmentation-map handling." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Different model-loading failures for different model families; no shared concrete bug path." + "reason": "Different model families and different loading failures; not the same checkpoint bug." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "PagedAttentionCache/Qwen3.5 attention-group crash vs NemotronH checkpoint loading failure; unrelated." + "reason": "Same broad area of model support, but one is a PagedAttentionCache runtime crash and the other is a checkpoint-loading failure." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF gpt-oss support and Phi-tiny-MoE autocast dtype mismatch are unrelated." + "reason": "Both are model/runtime compatibility issues, but one is GGUF architecture support and the other is a dtype mismatch under autocast." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "Zero-3/BertModel init crash vs torch version incompatibility; not the same underlying issue." + "reason": "No concrete overlap: ZeRO-3/Bert init failure versus torch-version incompatibility." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "GGUF architecture support vs Qwen3-MoE sliding_window behavior; different bugs." + "reason": "Different Qwen3-related bugs: GGUF architecture support versus sliding_window behavior." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both involve rotary/position-related code, but one is a Qwen2.5-VL video/image scaling bug and the other is uninitialized memory in Qwen3_5Moe rotary embedding." + "reason": "Both involve positional/rotary embeddings, but the symptoms and code paths are distinct." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is an FSDP2 attention KeyError and the other is a device mismatch with `device_map=\"auto\"`; separate defects." + "reason": "Same Gemma4 family, but one is an FSDP2 shared-kv-state KeyError and the other is a device-mapping mismatch in forward." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall and conservative. The accepted duplicate pairs are well supported by the titles/metadata: 43827\u219444509, 43872\u219444291, 41093\u219444805, and the Qwen video/doc/Whisper pairs are plausible without overclaiming. The rejected pairs stay appropriately separate by failure mode or code path." + "feedback": "The cluster summary is grounded in the packet, and the representative issue choice is reasonable. The soft-edge decisions are mostly conservative: the accepted pairs are well-supported by near-identical titles/root causes (docs pipeline removals, `_is_hf_initialized`, mask-shape IndexError) or a clearly shared tokenizer regression family, while the rejected pairs are appropriately kept separate when the concrete failure modes differ. No obvious overstatement stands out." }, "evaluator_used": true, "retried": false, @@ -1870,15 +1876,15 @@ "issue:42915|issue:43957", "issue:44560|issue:45685", "issue:43645|issue:43646", - "issue:43550|issue:44910", "issue:44560|issue:44918", + "issue:43550|issue:44910", "issue:44683|issue:46515", "issue:42617|issue:43366", "issue:36010|issue:42915", "issue:45812|issue:46489", "issue:45823|issue:46515", - "issue:44560|issue:44805", "issue:44661|issue:45823", + "issue:44560|issue:44805", "issue:38175|issue:43994", "issue:43632|issue:45834", "issue:44877|issue:45030", @@ -1887,8 +1893,8 @@ "issue:43317|issue:44756", "issue:44661|issue:44805", "issue:42491|issue:43720", - "issue:38175|issue:39692", "issue:43644|issue:44534", + "issue:38175|issue:39692", "issue:41720|issue:42915", "issue:42915|issue:44756", "issue:42491|issue:45276", @@ -1896,244 +1902,232 @@ "issue:44960|issue:44977", "issue:46459|issue:46512" ], - "prepared_review_unit_hash": "82486f339bfd6ce12e0e0fac45c4fca5531e99b318320ce466930bd76def32e0", + "prepared_review_unit_hash": "96260b857ebdddf979cce0768b06c9525effbc7295edda1bdd54886b22fb8880", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37773, - "estimated_input_tokens": 9444, - "estimated_eval_tokens": 19144 + "serialized_chars": 37772, + "estimated_input_tokens": 9443, + "estimated_eval_tokens": 19142 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers regression reports; only a few pairs look like true duplicates of the same underlying bug. The clearest duplicate groups are custom model initialization, non-persistent buffer corruption, SigLIP2 inference regression, and BERT model loading failures.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43646 is the broadest, most reusable summary of the custom-model initialization regression, with issue:43645 as a narrower notebook repro.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers regressions. The only clear duplicate pair is the non-persistent buffer corruption reports (43644/44534); a SigLIP2 output issue and a custom-model-init regression also look duplicate-like, but most other pairs are only broadly similar and should stay separate.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43644 is the strongest canonical anchor: it\u2019s the earliest, clearest statement of the non-persistent-buffer regression, and issue:44534 is an exact wording follow-up of the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best representative issue because it describes the underlying v5 custom model initialization breakage in general terms rather than a single environment-specific reproduction.", + "best_issue_reason": "issue:43644 is the best overall issue to keep as the representative duplicate target because it has the cleanest, most specific bug description and an exact near-duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43638", "right": "issue:46515", "accept": false, - "reason": "Different failures: deepspeed zero3 empty-tensor index error vs a torch 2.4.1 incompatibility report." + "reason": "Different failures: Deepspeed ZeRO-3/Bert shape issue vs a torch 2.4.1 incompatibility report." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Unrelated: GGUF gpt-oss support request vs a Float8 storage lookup error." + "reason": "Unrelated topics: GGUF gpt-oss support vs Float8 storage lookup failure." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different code paths and symptoms: image preprocessor/model dtype mismatch vs Qwen3-VL video StopIteration." + "reason": "Different model families and code paths: Mistral-3 image dtype mismatch vs Qwen3-VL video StopIteration." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo batched track leakage and Whisper empty-transcription handling are distinct bugs." + "reason": "AudioFlamingo track embedding leak and Whisper empty transcription are different bugs." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 long-OCR indexing." + "reason": "Non-persistent buffer loading regression vs Kosmos2.5 long-OCR index error; not the same bug." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Import failure for Qwen2.5-VL is not the same as meta-device loading regressions." + "reason": "Importing a missing class and meta-device loading failures are different code paths." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and meta-device loading issues are different problems." + "reason": "Qwen3-MoE FP8 failure and meta-device loading breakage are separate issues." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Qwen3-VL StopIteration during video handling is unrelated to the MPS histogram int error." + "reason": "Qwen3-VL video StopIteration and MoE MPS int-histogram error are unrelated." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying custom-model initialization regression; 43645 is just a notebook-specific repro of the broader 43646 bug." + "reason": "Both describe the same v5 regression breaking custom model initialization; the notebook mention is just a narrower manifestation." }, { - "left": "issue:43550", - "right": "issue:44910", + "left": "issue:44560", + "right": "issue:44918", "accept": false, - "reason": "Different models and mechanisms: Bamba SDPA compile crash vs Qwen3.5 flash-attention packed-position bug." + "reason": "Different Qwen3.5/TRL failures and different symptoms; not the same underlying bug." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:43550", + "right": "issue:44910", "accept": false, - "reason": "Both involve Qwen3.5/Qwen3-VL, but one is video StopIteration and the other is TRL embedding unpacking." + "reason": "Bamba torch.compile/SDPA failure is unrelated to Qwen3.5 flash-attention illegal access." }, { "left": "issue:44683", "right": "issue:46515", "accept": false, - "reason": "Compiled flex_attention on torch >= 2.9 is unrelated to the generic torch 2.4.1 incompatibility." + "reason": "Compiled flex_attention on torch>=2.9 and torch 2.4.1 incompatibility are different version-specific problems." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py execution failure and GGUF gpt-oss support are unrelated." + "reason": "3d_parallel.py run failure is unrelated to GGUF gpt-oss support." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Importing GenerationMixin and Qwen3Moe FP8 execution are different bugs." + "reason": "ImportError for GenerationMixin and Qwen3-MoE FP8 failure are unrelated." }, { "left": "issue:45812", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer-related, but Granite token IDs and DeepSeek-Coder v1 output are separate regressions." + "reason": "Both concern tokenization, but they target different models and likely different tokenizer bugs." }, { "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "Gemma4 PLE device mismatch is not the same as a broad torch version incompatibility report." + "reason": "Gemma4 device-mismatch during forward is not the same as a generic torch 2.4.1 incompatibility." }, { - "left": "issue:44560", - "right": "issue:44805", + "left": "issue:44661", + "right": "issue:45823", "accept": false, - "reason": "Different failures: video StopIteration vs mask/tensor shape mismatch." + "reason": "add-new-model-like validation failure and Gemma4 PLE device mismatch are different issues." }, { - "left": "issue:44661", - "right": "issue:45823", + "left": "issue:44560", + "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping command failure is unrelated to Gemma4 device placement." + "reason": "Qwen video StopIteration and a mask shape IndexError are not the same bug." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 inference regression on google/siglip2-base-patch16-224; both describe the model producing bad/degenerate outputs." + "reason": "Same SigLIP2 model and same wrong-output symptom; the second report is a rephrasing/expansion of the first." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "_is_hf_initialized flag breakage and Kosmos2.5 OCR indexing are unrelated." + "reason": "_is_hf_initialized flag breakage and Kosmos2.5 OCR index errors are unrelated." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Strict config loading for granite_speech and glm4v validation changes are different config bugs." + "reason": "Strict config blocking granite_speech and glm4v config validation are separate config problems." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "deepspeed zero3 Bert init failure is unrelated to the tokenizer-mapping add-new-model-like issue." + "reason": "ZeRO-3/Bert empty-shape index error and add-new-model-like tokenizer-mapping failure are different code paths." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Zero3 BertModel indexing and tokenizer-mapping command behavior are distinct issues." + "reason": "Both are index/config style failures, but the Bert ZeRO-3 error is not the same bug as tokenizer-mapping validation." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "gpu+cpu offload dequantization and Strix Halo mmap OOM are different loading problems." - }, - { - "left": "issue:44661", - "right": "issue:44805", - "accept": false, - "reason": "Tokenizer-mapping validation and mask shape indexing are unrelated." + "reason": "gpu+cpu offload dequantized-model loading and Strix Halo mmap OOM are different loading issues." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Qwen3_moe LoRA version incompatibility and packed-weight accelerate loading are different defects." - }, - { - "left": "issue:38175", - "right": "issue:39692", - "accept": false, - "reason": "SigLIP2 bad outputs and documentation example errors/quantization failures are not the same bug." + "reason": "Qwen3-MoE LoRA incompatibility and BitNet packed-weight loading are unrelated." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same non-persistent buffer corruption regression in Transformers 5.0.0; the titles are essentially the same bug." + "reason": "Exact same bug: Transformers v5 fills non-persistent buffers with junk; 44534 is just a later re-report." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but 39692 is a docs example/model-processor mismatch and quantization issue, not the same output-corruption bug." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and Qwen3Moe FP8 config failure are different runtime issues." + "reason": "Qwen3 auto device mapping cuda assert and Qwen3-MoE FP8 failure are different failures." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Qwen3Moe FP8 failure and Strix Halo mmap/OOM are unrelated." + "reason": "Qwen3-MoE FP8 and Strix Halo mmap OOM are unrelated." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "LoRA qwen3_moe compatibility and Gemma4 resize_token_embeddings are separate bugs." + "reason": "Qwen3-MoE LoRA hf4.x/hf5.x incompatibility and Gemma4 resize_token_embeddings are different regressions." }, { "left": "issue:44451", "right": "issue:44488", - "accept": true, - "reason": "Both report the same latest-version model-loading regression for closely related BERT models; the failure mode is the same." - }, - { - "left": "issue:44960", - "right": "issue:44977", "accept": false, - "reason": "Generic GLM5 issue and Qwen3.5 flash-attention generation failure are not the same problem." + "reason": "Both are model-loading failures, but the reports name different BERT repos and don\u2019t clearly point to one concrete shared bug." }, { "left": "issue:46459", "right": "issue:46512", "accept": false, - "reason": "Silent dtype handling when loading multimodal checkpoints is distinct from stale config.dtype after from_config()." + "reason": "Both mention dtype, but one is silent dtype loss on checkpoint load and the other is stale config.dtype after from_config; different bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays mostly conservative: the true-duplicate calls are well supported by the titles/descriptions, and the many false calls are reasonably separated by bug family. The summary also matches the evidence that this is a mixed cluster with only a few clear duplicate pairs." + "feedback": "The judgments are conservative and grounded in the packet. The three accepted soft pairs are plausible near-duplicates, and the rejected pairs mostly stay on the cautious side given the differing titles and symptom details. The canonical choice of issue:43644 is well supported by the exact follow-up issue:44534." }, "evaluator_used": true, "retried": false, @@ -2218,233 +2212,240 @@ "issue:44960|issue:46050", "issue:45488|issue:45812", "issue:44164|issue:44291", - "issue:45276|issue:45335", "issue:45362|issue:45406", + "issue:45276|issue:45335", "issue:44910|issue:45588", "issue:44402|issue:44488", "issue:43404|issue:43454", "issue:43425|issue:43957", "issue:43873|issue:44387", "issue:39692|issue:44960", - "issue:42947|issue:45704", "issue:41720|issue:44155", + "issue:42947|issue:45704", "issue:45663|issue:45759", "issue:44291|issue:44756", "issue:36010|issue:39404", + "issue:44387|issue:44756", "issue:45663|issue:45859" ], - "prepared_review_unit_hash": "bdd8634fc86ffdbf9d8169e2e3d38822c44424b8a7c50ba553e8e4b6f8920d6a", + "prepared_review_unit_hash": "9dd6c2b5f3fc619eaf3dab73aa30fbade302cbce61e10e5c992d15535ed5ce80", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35908, - "estimated_input_tokens": 8977, - "estimated_eval_tokens": 18210 + "soft_pair_count": 30, + "serialized_chars": 36205, + "estimated_input_tokens": 9052, + "estimated_eval_tokens": 18360 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly distinct issues that only share broad subsystem or symptom overlap. I found one strong duplicate-like pair around resize_token_embeddings in composite models; the rest should stay separate.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45276 is the broader, more representative report of resize_token_embeddings not propagating to nested/output embedding modules; issue:45335 looks like a model-specific instance of the same failure mode.", + "summary": "The cluster is still mostly heterogeneous. The two resize_token_embeddings and lm_head-weight-tying pairs are only model-family similarities in the packet, not strong enough to treat as confirmed duplicates or one mergeable fix, so they should remain separate.", + "confidence": 0.52, + "canonical_issue_reason": "Issue 45276 is the clearest anchor because it is a concrete, narrowly scoped resize_token_embeddings propagation bug; 45335 is related but not strong enough evidence of the same exact defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45276 is the best representative issue here because it describes the underlying resizing bug more generally and maps cleanly to the duplicate-like t5gemma report.", + "best_issue_reason": "45276 remains the best representative issue because it is the most specific and technically well-scoped item in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different regressions: one is Qwen3.5 ZeRO-3 weight loading, the other is incorrect visual-encoder keys on save_pretrained." + "reason": "Both are Qwen3.5 multimodal issues, but one is ZeRO-3 loading for language_model and the other is incorrect visual-encoder save keys; different code paths." }, { "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "Different models and failures; only broad AI/model-example overlap, not the same bug." + "reason": "No concrete shared bug or fix path is visible; they concern different models and different failures." }, { "left": "issue:44488", "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 regressions, but the model-specific symptoms point to different underlying paths." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Qwen3 MoE training memory inefficiency vs quantization/offloading behavior; not the same concrete issue." + "reason": "These are both memory/quantization adjacent, but one is training memory inefficiency and the other is offloading with quantization." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5.0 regressions, but the symptoms and code paths are unrelated." + "reason": "One is about non-persistent buffers, the other about notebook custom-model initialization; not the same bug." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Different memory issues: Qwen3 MoE training vs int4 quantization reserved-memory/OOM regression." + "reason": "Both mention memory, but the reported failures are different enough that a single fix is not supported by the packet." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different models and load-time failures; no clear shared underlying bug." + "reason": "Different models and different loading failures; no shared concrete defect is evident." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different model families and different loading/offload paths." + "reason": "These are unrelated loading/conversion failures across different models and subsystems." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Same family (SigLIP2) but one is a docs/example mismatch plus quantization failure, the other is bad inference outputs; not clearly one bug." + "reason": "Both involve SigLIP2, but one is a docs example with multiple errors and the other is bad AutoModel/pipeline outputs; not clearly the same bug." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA/Qwen3 MoE hf4.x vs hf5.x compatibility is unrelated to generic resize_token_embeddings propagation." + "reason": "A Qwen3 MoE compatibility issue is not the same as a resize_token_embeddings propagation bug." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume step math and SGD args handling are separate problems." + "reason": "Trainer checkpoint-resume math and SGD argument handling are separate issues." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different model/version compatibility issues with no shared code-path indicated." + "reason": "These are different Qwen3-related failures with different symptoms and likely different fixes." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "Different GLM5/model-loading and quantization persistence bugs." + "reason": "A model title and a quantized save/load regression do not indicate the same underlying bug." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizer classes/model families and are not clearly the same defect." + "reason": "Both are tokenizer regressions, but the affected models and likely root causes differ." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "save/from_pretrained extra_state handling and init_empty_weights/_is_hf_initialized are separate load/save failures." - }, - { - "left": "issue:45276", - "right": "issue:45335", - "accept": true, - "reason": "Both report resize_token_embeddings failing to update nested/output embeddings in composite models; 45335 looks like a narrower variant of the same bug." + "reason": "Saving/loading extra_state and init_empty_weights argument handling are unrelated serialization issues." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Different models and crash causes: chat crash for Qwen3.5 vs Gemma4Processor missing _tokenizer in serve." + "reason": "Different models and different crashes; no strong evidence of one shared fix." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "They share a resize_token_embeddings bug class, but the packet only shows Gemma4 vs T5Gemma model-specific variants, not enough to confirm one bug or one mergeable fix." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both involve flash_attention.py, but one is a packed-sequence/position_ids bug and the other is a sink-less s_aux=None AttributeError." + "reason": "Both touch flash_attention, but one is a Qwen3.5 memory access crash and the other is an AttributeError on s_aux=None; different paths." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "One is tokenizer vocab-size inconsistency; the other is a model-loading failure for a different tokenizer/model." + "reason": "Tokenizer vocab-size mismatch and tokenizer loading failure are different problems." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both mention untied lm_head weights, but they are different model classes with no clear proof of the same code-path bug." + "reason": "They both involve lm_head tying, but the packet only shows Mistral3 vs AyaVision variants; that is not enough to confidently call them the same bug or one fix." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "General Torch 2.10 incompatibility is not the same as meta-device loading breakage for specific models." + "reason": "Torch version incompatibility and meta-device loading errors are distinct regressions." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Quantization/offload inefficiency and int4 reserved-memory/OOM are related only at a high level, not the same bug." + "reason": "Both are memory-related, but the specific failures differ too much to merge as one bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 documentation/model issues are unrelated to GLM5 loading problems." + "reason": "No meaningful overlap beyond both being model-related issues." }, { - "left": "issue:42947", - "right": "issue:45704", + "left": "issue:41720", + "right": "issue:44155", "accept": false, - "reason": "Gradient checkpointing/PEFT LoRA and an apex RMSNorm memory leak are unrelated." + "reason": "A Qwen3 device-mapping crash and an AudioFlamingo3 batched inference leak are unrelated." }, { - "left": "issue:41720", - "right": "issue:44155", + "left": "issue:42947", + "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device mapping crash and AudioFlamingo3 track-leak inference bug are different failures." + "reason": "Gradient checkpointing with PEFT LoRA and an Apex RMSNorm memory leak are separate problems." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma-4 FSDP2 attention state bug vs Qwen composite config unwrapping bug; different models and code paths." + "reason": "Gemma4 FSDP2 attention-state rebuilding and Qwen multimodal text_config unwrapping are different initialization bugs." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected init arg during init_empty_weights is unrelated to mmap OOM on Strix Halo." + "reason": "Unexpected _is_hf_initialized handling and Strix Halo mmap OOM are unrelated." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin and Whisper pipeline return_language regression are separate issues." + "reason": "GenerationMixin import failure and Whisper pipeline return_language regression are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 reserved-memory OOM and disabling mmap on Strix Halo are different memory issues with different causes." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma-4 FSDP2 attention state handling and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different bugs." + "reason": "Gemma4 FSDP2 KeyError and Qwen3.5 MoE CPU-offload incompatibility are separate bugs." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative overall. The one accepted pair, issue:45276 \u2194 issue:45335, is well supported by the titles and the high similarity signal: both describe resize_token_embeddings failing to propagate to nested/output embeddings in composite models. The rest of the soft-edge verdicts are appropriately cautious and do not overclaim shared root causes based on broad subsystem overlap alone." + "accept": false, + "feedback": "The summary is mostly grounded, but the two accepted soft edges overstate the evidence. In both accepted pairs, the titles only show a shared bug class across different models, not enough to confidently say they are the same bug or a single mergeable fix. In particular, 45276\u219445335 is Gemma4 vs T5Gemma, and 43404\u219443454 is Mistral3 vs AyaVision; those are plausible similarities, but not conservative enough to mark true from the packet alone. The rest of the verdicts look appropriately cautious." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -2474,11 +2475,11 @@ "issue:43716", "issue:43828", "issue:43854", + "issue:43906", "issue:43957", "issue:44155", "issue:44164", "issue:44292", - "issue:44387", "issue:44402", "issue:44451", "issue:44479", @@ -2512,10 +2513,9 @@ ], "soft_pairs": [ "issue:44292|issue:44912", - "issue:44387|issue:44756", "issue:43957|issue:44756", - "issue:44155|issue:45823", "issue:43335|issue:45663", + "issue:44155|issue:45823", "issue:42947|issue:43381", "issue:44863|issue:46468", "issue:45276|issue:45292", @@ -2544,234 +2544,247 @@ "issue:43425|issue:45823", "issue:43425|issue:43716", "issue:44479|issue:44805", - "issue:44589|issue:45084" + "issue:44589|issue:45084", + "issue:43906|issue:44530" ], - "prepared_review_unit_hash": "f04df2b56683c7ed236d1fa38fe3c692b088629108fc88e4118891912a37b290", + "prepared_review_unit_hash": "d0a88a5bd7125803c3028500bfede436526631cf852999066dad252097fa008c", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37888, - "estimated_input_tokens": 9472, - "estimated_eval_tokens": 19200 + "serialized_chars": 37882, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers bug reports. The only clear duplicate pocket is around `resize_token_embeddings` not propagating changes to output/token embeddings across related model variants; most other soft matches are just same subsystem, same model family, or same symptom class, not the same concrete bug.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45292 is the cleanest general statement of the duplicate set: it describes the core `resize_token_embeddings` bug against `output_embeddings`, and it sits at the center of the accepted near-duplicates.", + "summary": "This cluster is mostly a mix of unrelated bug reports. The only clear duplicate family is the resize_token_embeddings/output embedding regression affecting multiple model variants; the rest should stay separate.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45292 is the cleanest, most generic report of the shared resize_token_embeddings/output_embeddings bug, and it aligns with the model-specific variants in the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best global representative because it is concise, model-agnostic, and captures the shared code-path problem better than the more model-specific variants.", + "best_issue_reason": "issue:45292 is the best cluster representative because it names the core code-path problem in model-agnostic terms and sits at the center of the accepted duplicate family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Both are quantization/load failures, but they affect different models and different quantization formats (`NVFP4` vs `MXFP4`); too broad to be the same bug." + "reason": "Both are quantization/loading regressions, but they involve different models, quantization formats, and failure modes; not the same concrete bug." }, { - "left": "issue:44387", + "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "One is an int4 quantization memory regression; the other is Strix Halo mmap/OOM behavior. Same symptom, different underlying cause." + "reason": "Meta-device loading breakage and Strix Halo mmap OOM are different root causes and fixes." }, { - "left": "issue:43957", - "right": "issue:44756", + "left": "issue:43335", + "right": "issue:45663", "accept": false, - "reason": "`torch.device('meta')` model-loading breakage is unrelated to the Strix Halo mmap/OOM issue." + "reason": "Different model/configuration bugs: sparse-layer construction vs FSDP2 KeyError in Gemma4 attention." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 track leakage and Gemma4 device-mapping mismatch are different forward-path bugs." - }, - { - "left": "issue:43335", - "right": "issue:45663", - "accept": false, - "reason": "Different models and different failures: SwitchTransformers config construction vs Gemma4 FSDP2 cache rebuilding." + "reason": "AudioFlamingo3 track leakage and Gemma4 PLE device mismatch are unrelated inference bugs." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is about PEFT/LoRA effectiveness and the other about eval-mode usage." + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA behavior and the other is eval-mode usage; different issues." }, { "left": "issue:44863", "right": "issue:46468", "accept": false, - "reason": "Both are Nemotron-related, but one is checkpoint loading and the other is an example script failure; not the same code-path bug." + "reason": "NemotronH checkpoint loading failure and official Nemotron example breakage are related by model family only, not the same bug." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Both report `resize_token_embeddings` not updating output-side embeddings; `45276` is a model-specific variant of the same underlying bug." + "reason": "Same underlying resize_token_embeddings regression: Gemma4-specific failure to update embeddings matches the generic output_embeddings bug." }, { "left": "issue:45685", "right": "issue:46468", "accept": false, - "reason": "Unrelated failures: MPS histogram type error in MoE vs a Nemotron example breakage." + "reason": "MPS histogram unsupported for MoE and a broken Nemotron example are different failures." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression for Qwen VL models is unrelated to strict config loading for granite_speech." + "reason": "Video-input regression and strict-config granite_speech loading are unrelated." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Save/load `extra_state` handling is unrelated to mmap-induced OOM on Strix Halo." + "reason": "save/from_pretrained extra_state handling and mmap OOM avoidance are not the same code-path." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Gradient checkpointing in eval mode is not the same as Qwen3.5 RLHF NaNs from 3D `position_ids` and SDPA fallback." + "reason": "Eval-mode checkpointing and RLHF NaN explosions from 3D position_ids are different bugs." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Model-test loading failure for GLM-4.7-Flash and Qwen3.5 save regression are different code paths." + "reason": "Different model-loading regressions: unit test loading of GLM-4.7-Flash vs Qwen3.5 save_pretrained correctness." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve custom-model save/load, but one is a permissions propagation bug and the other is wrong-module loading." + "reason": "Both involve custom-model save/load, but one is file permissions and the other is wrong module resolution after save_pretrained." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention sink handling and Gemma4 FSDP2 shared KV state are unrelated." + "reason": "Flash-attention sink-less crash and Gemma4 FSDP2 KeyError are distinct attention-path issues." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Both are tokenizer issues, but one is BPE-marker handling and the other is `continue_final_message` logic." + "reason": "Tokenizer BPE-marker handling and continue_final_message behavior are different tokenizer regressions." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same underlying regression: `resize_token_embeddings` fails to propagate to output/decoder embeddings across model variants." + "reason": "Same resize_token_embeddings bug, with 45335 as a model-specific report that decoder.embed_tokens is not updated." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Both mention mismatches, but one is autocast dtype behavior and the other is meta-device loading." + "reason": "Both are dtype-related, but one is autocast mismatch and the other is meta-device loading failure." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization in notebooks and Trainer resume/max_steps calculation are unrelated bugs." + "reason": "Custom model initialization in notebooks and Trainer resume max_steps calculation are unrelated." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Both are model/tokenizer loading issues, but they concern different models and different failure modes." + "reason": "Tokenizer vocab-size mismatch and ScandiBERT loading failure are different tokenizer/model issues." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model initialization and `extra_state` save/load handling do not share the same concrete defect." + "reason": "Notebook custom-model initialization and extra_state save/load handling are separate bugs." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to mmap/OOM behavior." + "reason": "Float8 storage lookup failure and mmap OOM avoidance are unrelated runtime problems." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "A broad Torch 2.10 incompatibility report is not the same bug as the Phi-tiny-MoE autocast dtype mismatch." + "reason": "Torch 2.10 incompatibility and autocast dtype mismatch are not the same concrete bug." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer regressions, but one is wrong token IDs and the other is slower `convert_tokens_to_ids`." + "reason": "Wrong Granite token IDs and slow convert_tokens_to_ids are separate tokenizer regressions." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "Tokenizer performance regression and CodeLlama leading-space round-trip regression are different defects." + "reason": "Performance regression in convert_tokens_to_ids and CodeLlama space-roundtrip regression are different issues." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline `return_language` regression is unrelated to a generic tokenization change report." + "reason": "Whisper return_language pipeline regression and version-dependent tokenization changes are unrelated." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both concern custom model loading after v5, but one is about `base_model_prefix` and the other about Jupyter-defined classes; too different to merge." + "reason": "Both concern custom models, but one is base_model_prefix loading and the other is notebook initialization semantics." }, { "left": "issue:39404", "right": "issue:46665", "accept": false, - "reason": "Whisper pipeline language return and `from_pretrained` inside Python classes are unrelated." + "reason": "Whisper pipeline return_language and from_pretrained-in-classes behavior are different bugs." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache group-type failure and non-template-node compile error are not the same code-path problem." + "reason": "PagedAttentionCache group-type crash and compile non-template-nodes error are unrelated." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch with `device_map='auto'` are unrelated." + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are different failure modes." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility and Mistral-3 image-preprocessor/model dtype mismatch are different bugs." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Qwen VL video-input regression and a generic mask-shape IndexError are not the same bug." + "reason": "Video-input regression and mask-shape IndexError are distinct code-path failures." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup and compile-node type errors are unrelated." + "reason": "Float8 storage error and compile non-template-nodes error are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Isolated reproduction of issue 38071 and PagedAttentionCache crash are not the same underlying bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The duplicate pocket around resize_token_embeddings (45276/45292/45335) is well supported by the titles and similarity scores, and the other soft-edge rejections do not overclaim sameness beyond what the packet shows." + "feedback": "The cluster summary is grounded in the packet and stays conservative: most pairs are reasonably rejected as different bugs, and the accepted resize_token_embeddings/output_embeddings family is well supported by the titles. The chosen representative issue:45292 is a sensible generic center for that duplicate family." }, "evaluator_used": true, "retried": false, @@ -2842,7 +2855,6 @@ "issue:46489" ], "soft_pairs": [ - "issue:43906|issue:44530", "issue:43901|issue:44509", "issue:43425|issue:44918", "issue:43994|issue:44079", @@ -2871,233 +2883,224 @@ "issue:44661|issue:44683", "issue:43643|issue:44530", "issue:44451|issue:44987", - "issue:42915|issue:44530", - "issue:43541|issue:45084" + "issue:43541|issue:45084", + "issue:42915|issue:44530" ], - "prepared_review_unit_hash": "150636bc4a963e822dec0c6c4435b1922171d8cf7ae0483fcf0780b1c7aa0ef4", + "prepared_review_unit_hash": "d61d21249559251a0e8f7783990b87f8cc78d864ecd5fdb0b6d87321dd8bc57c", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37109, - "estimated_input_tokens": 9278, - "estimated_eval_tokens": 18812 + "soft_pair_count": 30, + "serialized_chars": 36807, + "estimated_input_tokens": 9202, + "estimated_eval_tokens": 18660 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue pairs by title/metadata. The one clear duplicate-like match is the Qwen3-VL MoE loading pair (43299/43931); the rest look like separate bugs or separate documentation/model-specific regressions.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43299 is the broadest and most central report in the only strong duplicate-like pair: a Transformers 5.0.0.dev0 regression breaking Qwen3-VL MoE model loading.", + "summary": "This cluster is very heterogeneous: most items are unrelated model-, tokenizer-, export-, or runtime-specific regressions, so there is no single strong duplicate center. If one issue must represent the cluster, issue 43611 is the broadest loading-regression report.", + "confidence": 0.42, + "canonical_issue_reason": "Issue 43611 is the broadest and most central representative here because it describes a general Transformers 5.0 loading regression affecting models with `base_model_prefix`, which is closer to the common theme than the narrower model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43299 best represents the shared Qwen3-VL MoE loading failure described by the matching pair, while issue:43931 reads as a more specific symptom of the same problem.", + "best_issue_reason": "Issue 43611 is the best global representative for this cluster since it captures the widest class of failures among the items; the rest are mostly distinct edge cases or model-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43906", - "right": "issue:44530", - "accept": false, - "reason": "The titles point to different areas: an isolated reproduction of an old issue versus a Qwen3.5 PagedAttentionCache crash." - }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs-related, but the titles describe different documentation topics and different pipeline task mentions." + "reason": "Both are docs-related, but one is about `return_all_scores` and the other about removed pipeline tasks; different documentation issues, not the same bug." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "The titles describe different failures: Torch version incompatibility versus a Qwen3.5 SFT trainer embedding issue." + "reason": "One is a Torch 2.10 compatibility issue, the other is a TRL SFT trainer/model unpacking failure; no shared code-path bug." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "The titles suggest different bug types: SigLIP2 inference/output issues versus ModelOutput key assignment." + "reason": "`siglip2` bad outputs and `ModelOutput` key assignment are different failure modes in different areas." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model regressions in v5, but the titles refer to different models and different loading failures." + "reason": "CamemBERT prediction regression and tokenizer loading for `EMBIDDIA/est-roberta` are unrelated tokenizer/model issues." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "The titles describe different problems: torch.export export failure versus a Perceiver image classification issue." + "reason": "`torch.export` failing on `torch_compilable_check` is a compiler/export problem; Perceiver non-default resolution is a separate model-shape issue." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "The titles point to different loading bugs: a T5 tokenizer conflict and a gguf/from_pretrained breakage." + "reason": "T5 tokenizer method conflict and `gguf_file` loading breakage are different tokenizer/loading bugs." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "The titles describe different regressions: Whisper pipeline return_language and MXFP4 quantization loading/fallback." + "reason": "Whisper pipeline language return regression and MXFP4 quantization fallback are unrelated." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "The titles describe different save/load or config-handling problems: extra_state versus composite Qwen multimodal configs." + "reason": "`extra_state` save/load handling and `from_config` unwrapping of `text_config` are different serialization/configuration issues." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "The titles point to different bugs: quantized save/load corruption and a mistral_common tokenizer behavior issue." + "reason": "Quantized `sinq` save/load breakage and mistral_common `continue_final_message` behavior are distinct bugs." }, { "left": "issue:43299", "right": "issue:43931", - "accept": true, - "reason": "The titles are closely aligned on Qwen3-VL MoE model loading; one is the broad regression and the other is a specific weight-shape mismatch symptom." + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a dev0 regression in Moe model loading and the other is a weight-shape mismatch for a specific checkpoint; not the same concrete defect." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "The titles describe different issues: an isolated reproduction of an old issue and a tokenizer continue_final_message bug." + "reason": "An isolated reproduction of an older issue and a mistral_common prompt-format bug are not the same underlying change." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "The titles describe different tokenizer-loading failures on different models." + "reason": "Different tokenizer failures: loading `est-roberta` vs Mistral regex patching with `fix_mistral_regex=True`." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "The titles point to different topics: Qwen3 MoE training memory usage versus Strix Halo mmap/OOM handling." + "reason": "One is training memory inefficiency in Qwen3 MoE, the other is mmap-related OOM on Strix Halo; same broad symptom, different causes." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "The titles describe different failures: a deepspeed zero3/BertModel index error and a compiled flex_attention torch version issue." + "reason": "Deepspeed zero3/BertModel indexing error and compiled flex_attention torch version failure are unrelated runtime/compiler issues." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "The titles describe different behaviors: extra_state save/load handling and a Qwen VL video-input regression." + "reason": "Save/load `extra_state` and video-input regression in Qwen VL models are separate problems." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "The titles indicate different tokenizer loading problems for different models." + "reason": "Loading `sleng-bert` and Mistral tokenizer regex patching are different tokenizer paths." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "The titles describe different loading regressions for different model names and version ranges." + "reason": "`sleng-bert` loading failure and `physical-intelligence/fast` loading failure are separate model-loading regressions." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "The titles point to different v5 breakages: base_model_prefix loading and non-persistent buffers." + "reason": "Both are Transformers 5 loading regressions, but one is about `base_model_prefix` handling and the other about non-persistent buffers; different bugs." }, { "left": "issue:44462", "right": "issue:46489", "accept": false, - "reason": "The titles describe different tokenizer-related bugs: ignoring tokenizer.json and a DeepSeek-Coder v1 output regression." + "reason": "Ignoring repository `tokenizer.json` and DeepSeek-Coder tokenizer output regression are related to tokenizers but not the same defect." }, { "left": "issue:42796", "right": "issue:43404", "accept": false, - "reason": "The titles refer to different problems: tokenizer BPE marker handling and Mistral3 weight tying." + "reason": "Ministral tokenizer BPE-marker handling and Mistral3 weight tying are different model/tokenizer issues." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "The titles point to different failures: a T5 tokenizer load error and a Qwen3 MoE FineGrainedFP8Config issue." + "reason": "T5 tokenizer load conflict and Qwen3Moe FP8 config failure are unrelated." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "The titles describe different bugs: image preprocessor dtype mismatch and a template compilation TypeError." + "reason": "Image-preprocessor dtype mismatch and template compilation error are unrelated." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "The titles point to different model-structure/loading issues affecting different models." + "reason": "CLIP structure inconsistency and Sam3Model text-encoder loading are both model-loading regressions but on different architectures and code paths." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "The titles describe different errors: Trainer GPU broadcast failure and Whisper return_language behavior." + "reason": "Trainer GPU broadcast NoneType error and Whisper return_language regression do not share a concrete underlying bug." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "The titles both mention device/offload concerns, but they describe different models and different failures." + "reason": "XLNet CPU device-placement bug and Qwen3_5Moe CPU-offload incompatibility are different device-management problems." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "The titles point to unrelated topics: TOKENIZER_MAPPING_NAMES behavior and flex_attention compilation." + "reason": "Tokenizer-mapping codegen failure and flex_attention torch version failure are unrelated." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "The titles describe different bugs: missing config fields with trust_remote_code and a Qwen3.5 cache crash." + "reason": "Missing fields with `trust_remote_code=True` and PagedAttention cache crashes on `linear_attention` are different issues." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "The titles indicate different loading regressions for different models." + "reason": "Different model-loading failures for different checkpoints; no shared code-path evidence." }, { - "left": "issue:42915", - "right": "issue:44530", + "left": "issue:43541", + "right": "issue:45084", "accept": false, - "reason": "The titles describe different Qwen3-related issues: FineGrainedFP8Config versus PagedAttentionCache linear_attention." + "reason": "Grouped_mm tracing failure in Mixtral and template-node compilation failure are different compiler/runtime problems." }, { - "left": "issue:43541", - "right": "issue:45084", + "left": "issue:42915", + "right": "issue:44530", "accept": false, - "reason": "The titles point to different code-path failures: torch dynamo grouped_mm tracing and template compilation." + "reason": "Both mention Qwen3-family models, but one is FP8 fine-grained config loading and the other is paged-attention cache group-type handling; not the same bug." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall summary is plausible, but several soft-edge explanations go beyond what the packet supports. With only titles/metadata available, claims like \u201cdifferent subsystems,\u201d \u201cno shared code path is evident,\u201d or specific root-cause distinctions are a bit too strong. The Qwen3-VL MoE pair (43299/43931) is the only clearly plausible duplicate-like match from the packet, but the rest of the verdicts should stay more explicitly title-based and tentative." - }, - "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 }, @@ -3180,8 +3183,8 @@ "issue:44987|issue:44991", "issue:43262|issue:43377", "issue:43746|issue:46326", - "issue:44403|issue:45081", "issue:42915|issue:45084", + "issue:44403|issue:45081", "issue:43883|issue:44991", "issue:44464|issue:45987", "issue:43646|issue:44534", @@ -3198,232 +3201,226 @@ "issue:42915|issue:44589", "issue:44062|issue:44987" ], - "prepared_review_unit_hash": "ab9dd1879980ea9690688a1a90289bab50b8ce22508bab53c9c93b90a6c6a5fb", + "prepared_review_unit_hash": "5e64b6a05cf15e73449095944cf32a49920c920a2fc747d3ce14c29b8a83aff7", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37705, - "estimated_input_tokens": 9427, - "estimated_eval_tokens": 19110 + "serialized_chars": 37698, + "estimated_input_tokens": 9425, + "estimated_eval_tokens": 19106 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose family of Transformers v5-era loading/initialization regressions, but the soft pairs are generally too different to merge as exact duplicates. The best representative issue is the broad custom-model initialization regression.", - "confidence": 0.54, - "canonical_issue_reason": "issue:43646 is the broadest, most central report in the set: it describes a Transformers 5.0.0 regression in custom model initialization, which is the closest umbrella theme for several nearby v5 loading/init failures.", + "summary": "This cluster is noisy and mostly contains unrelated bug reports that only share broad Transformer/v5/loading terminology. I did not find any soft pair that looks like the same underlying issue, so all soft-edge candidates are rejected. The broadest representative issue is the generic custom-model initialization regression.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43646 is the broadest, most representative Transformers v5 regression in the set: a generic custom-model initialization breakage, which best matches the cluster\u2019s overall similarity theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best cluster anchor because it is generic enough to represent the recurring v5 initialization/loading breakages without being overly model-specific.", + "best_issue_reason": "Issue 43646 is the best single issue to anchor this cluster because it is the most general symptom among the v5 regression reports and is more representative than the highly specific model- or backend-specific failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Different bugs: perplexity/model-eval behavior vs tokenizer post-processor updates at runtime." + "reason": "Different problems: perplexity/eval behavior vs tokenizer backend post-processor updates." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Unrelated failures: Apple Silicon warmup TypeError vs `_is_hf_initialized` flag breakage." + "reason": "Unrelated failures: Apple Silicon warmup TypeError vs `_is_hf_initialized` flag regression." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are v5 regressions, but one is about non-persistent buffers being corrupted and the other about custom model initialization; not the same concrete bug." + "reason": "Both are v5 regressions, but they describe different bugs: non-persistent buffer serialization vs custom model initialization." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Different subsystems and symptoms: token counting vs multimodal iterable dataset handling." + "reason": "Trainer token accounting and multimodal iterable-dataset loading are unrelated." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Both mention backend packaging, but the specific issues differ: missing tensorflow in extras vs an unsupported `tensorflow_text` backend name." + "reason": "Both mention backends, but one is packaging extras and the other is an unsupported backend mapping entry." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Related audio code, but the reported failures are different and not clearly the same root cause." + "reason": "Same audio area, but different concrete code paths and symptoms." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both involve Whisper/special tokens, but one is about `batch_decode` ignoring a flag and the other is empty-transcription generation after token alignment." + "reason": "Both involve Whisper, but one is `batch_decode` skipping special tokens and the other is an empty-transcription/alignment failure." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Unrelated: missing tied-weights metadata in one model vs a separate model-loading failure." + "reason": "Different model-loading errors with different missing attributes and failure modes." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "Different model families and code paths: multimodal iterable dataset loading vs mistral_common tokenizer behavior." + "reason": "Different models and different features: multimodal iterable dataset loading vs `continue_final_message` handling." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Different problems: tokenizer method-name conflict vs Jupyter/custom model initialization regression." + "reason": "Tokenizer class-name conflict and Jupyter custom-model initialization are separate bugs." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both are GGUF-related, but one is a generic `gguf_file` loader break and the other is a specific Qwen2.5-GGUF load failure; not enough evidence they are the same bug." + "reason": "Both involve GGUF loading, but one is a generic API break and the other is a specific Qwen2.5-GGUF regression." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Different loading paths: device_map/offload dequantized model loading vs GGUF model loading." + "reason": "Different loading contexts: device-map dequantized offload vs Qwen2.5-GGUF failure." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Different failures: template compilation error vs an invalid method call in GraniteMoEHybrid." + "reason": "Template compilation error and GraniteMoEHybrid invalid-method call are unrelated." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are model/tokenizer loading regressions, but they affect different models and likely different root causes." + "reason": "Both are loading regressions, but they affect different models and likely different tokenizer/model code paths." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio-processing bugs, but one is chat-template sampling-rate handling and the other is batch padding-mask behavior." + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask batching are distinct issues." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "Different features and code paths: PEFT/local checkpoint loading vs mistral_common continue-final-message behavior." + "reason": "PEFT local-checkpoint loading and mistral_common `continue_final_message` are unrelated." }, { - "left": "issue:44403", - "right": "issue:45081", + "left": "issue:42915", + "right": "issue:45084", "accept": false, - "reason": "Noise on load vs a specific Mistral tokenizer regex patch crash; not the same bug." + "reason": "FP8 config loading failure and template compilation failure do not share the same bug." }, { - "left": "issue:42915", - "right": "issue:45084", + "left": "issue:44403", + "right": "issue:45081", "accept": false, - "reason": "Completely different failures: FP8 config loading vs template compilation." + "reason": "Unnecessary loading noise and a Mistral regex patch crash are different issues." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Different model-specific load failures with no shared concrete code-path evident." + "reason": "No substantive overlap beyond both being model-loading failures." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Generation inconsistency with compiled forward vs cache length return-type contract; related area, but not the same issue." + "reason": "Chunked generation inconsistency and `StaticCache.get_seq_length()` type mismatch are different code paths." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Both are v5 regressions, but one is about custom model init and the other about non-persistent buffer serialization." + "reason": "Both are v5-related custom-model regressions, but one is initialization and the other is buffer serialization." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both mention lm_head/tied weights, but one is a missing tie in a specific model and the other is serialization varying by device." - }, - { - "left": "issue:43645", - "right": "issue:44534", - "accept": false, - "reason": "Different symptoms and code paths: notebook model init vs buffer serialization corruption." + "reason": "Related to tied weights, but one is untied weights in a model class and the other is device-dependent serialization behavior." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Different Qwen model bugs: hidden-state output vs rope/position-id scaling." + "reason": "Different Qwen model bugs with different mechanisms." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Both touch audio/codec models, but one is padding-mask batching and the other is STE/forward mismatch in DAC." + "reason": "MIMI batching/padding and DAC latent/STE behavior are unrelated." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different model-loading failures in unrelated tokenizers/models." + "reason": "Different model loading failures with different root causes." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Both involve FP8-ish issues, but the reported errors and likely code paths are not the same." + "reason": "Both touch v5/init behavior, but the failures are different: FP8 config vs unexpected `_is_hf_initialized` argument." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Different models and failures: nonsensical outputs vs processor/model input-size mismatch." + "reason": "Different models and different symptoms: nonsensical predictions vs image-size mismatch." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Unrelated model-loading regressions with different failure modes." + "reason": "Different loading bugs for different models/tokenizers." }, { "left": "issue:28530", "right": "issue:34567", "accept": false, - "reason": "Completely different trainer features: early stopping metric selection vs token counting state updates." + "reason": "Early stopping configuration and token-count tracking are unrelated." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Different tokenizer/model loading errors with no clear shared root cause." + "reason": "Tokenizer load conflict and MobileLLM key error are different failures." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "Both mention FP8/float8-adjacent failures, but the concrete bug surfaces are different and not clearly mergeable." + "reason": "Both involve newer v5/FP8-related regressions, but the specific errors and affected subsystems differ." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "Different tokenizer construction/loading errors for different models." + "reason": "AddedToken keyword conflict and model-loading failure for a specific model are not the same bug." } ] }, @@ -3523,17 +3520,17 @@ "issue:43482|issue:43525", "issue:35707|issue:40444", "issue:43482|issue:45538", - "issue:43632|issue:44534", - "issue:41553|issue:43582" + "issue:41553|issue:43582", + "issue:43632|issue:44534" ], - "prepared_review_unit_hash": "18512ec71675f07834add63213c4e214ef6874d732b2ce9bf5b2ba3d96ade4fc", + "prepared_review_unit_hash": "eff243808bda245775a5482ee86bfa192b114af6b302f128646db0275f40be7b", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, @@ -3548,195 +3545,195 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous issue cluster. The only clear duplicate pair is the Qwen3.5 save_pretrained regression; the Qwen VL video vs still-image rope-index reports are related but not the same bug and should stay separate.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45216 is the best anchor because it states a concrete Qwen3.5 save_pretrained regression with a clear version boundary, and issue:45357 reads like the same checkpoint-key bug persisting in a later release.", + "summary": "Most items are unrelated; only the Qwen3.5 `save_pretrained` regression pair looks like a true duplicate. The rest span distinct models, APIs, and failure modes, so they should stay separate.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45216 is the best canonical representative: it describes the broader Qwen3.5 `save_pretrained` checkpoint regression, and 45357 looks like a narrower follow-up on the same serialization path (incorrect visual encoder keys).", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the most representative and actionable issue in the set: it has a specific failure mode, version regression, and a strong duplicate candidate in issue:45357.", + "best_issue_reason": "45216 is the strongest issue candidate overall because it is the broader, earlier regression report and subsumes the narrower 45357 symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation sync/dispatch overhead is unrelated to StaticCache.get_seq_length() returning a tensor instead of an int." + "reason": "Both mention generation/caching behavior, but one is a broad per-token performance complaint and the other is a `get_seq_length()` contract issue; not the same bug." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export with torch_compilable_check is a different failure mode from OmDet-Turbo's processor/model image-size mismatch." + "reason": "`torch.export` with `torch_compilable_check` is unrelated to a processor/model input-size mismatch." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Both are multimodal failures, but one is a Llava Next load crash and the other is a Qwen3OmniMoe video-processing ValueError." + "reason": "Different models and different failures: Llava Next loading segfault vs Qwen3OmniMoe video-input `ValueError`." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "They are both Qwen VL-related, but one is a video-input regression across several models while the other is a still-image temporal position_ids scaling issue in qwen2.5-vl." + "reason": "Both are Qwen-VL/video-related, but one is a broad video regression across several models while the other is a still-image `position_ids` scaling bug in Qwen2.5-VL." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Both describe the same Qwen3.5 save_pretrained checkpoint-layout regression affecting the visual encoder keys." + "reason": "Same model and same `save_pretrained` serialization path; 45357 reads like a narrower symptom of the broader Qwen3.5 checkpoint corruption reported in 45216." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "A T5 tokenizer method-name conflict is unrelated to Qwen3 auto device mapping causing cudaErrorAssert." + "reason": "Tokenizer method-name conflict is unrelated to the Qwen3 auto-device CUDA assert." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are different bugs in different model families." + "reason": "Sliding-window handling for Qwen3-MoE is unrelated to strict config loading for GraniteSpeech." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "One is a tokenizer loading failure and the other is a tokenizer performance regression; they are not the same bug." + "reason": "Both involve tokenizers in v5, but one is a load failure for a specific model and the other is a performance regression in `convert_tokens_to_ids`." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch 2.10 incompatibility is unrelated to noisy model-loading warnings." + "reason": "Torch version incompatibility is unrelated to loading noise during transformer initialization." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM is unrelated to Qwen3-MoE sliding_window issues." + "reason": "PaddleOCR-VL loading in vLLM is unrelated to Qwen3-MoE sliding-window behavior." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio feature extraction and compiled flex_attention failures are separate code paths." + "reason": "Audio feature extraction failure is unrelated to compiled flex attention on newer torch versions." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "base_model_prefix loading breakage and non-persistent buffer corruption are both v5 regressions but not the same underlying issue." + "reason": "`base_model_prefix` loading failure and non-persistent buffer corruption are distinct v5 regressions with different code paths." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen2.5-VL-related, but one is training with multiple images per prompt and the other is temporal position-id scaling." + "reason": "Multi-image IterableDataset finetuning failure is not the same as the still-image temporal position-id scaling bug." }, { "left": "issue:45812", "right": "issue:46491", "accept": false, - "reason": "Different tokenizer regressions on different model families and different symptoms." + "reason": "Different tokenizers and different symptom classes: wrong Granite token IDs vs CodeLlama leading-space round-trip regression." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Loading noise and tokenizer load failure are not the same bug." + "reason": "General loading noise is not the same as a concrete tokenizer load failure." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "An _is_hf_initialized regression and a Strix Halo mmap OOM are unrelated." + "reason": "`_is_hf_initialized` regression is unrelated to mmap/OOM behavior on Strix Halo." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "An eval-mode checkpointing misuse is different from a missing supports_gradient_checkpointing flag." + "reason": "Eval-mode gradient checkpointing behavior is not the same as a missing `supports_gradient_checkpointing` flag." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper loading/runtime failure and missing Float8 storage object are distinct failures." + "reason": "Whisper runtime failure is unrelated to the Float8 storage-object error." }, { "left": "issue:42796", "right": "issue:43746", "accept": false, - "reason": "Tokenizer BPE-marker handling and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Tokenizer BPE-marker handling is unrelated to GraniteSpeech PEFT checkpoint loading." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen VL training issues, but they affect different models and different failure modes." + "reason": "Both are Qwen-VL training issues, but one is multi-GPU training and the other is multi-image prompt handling; not the same bug." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and Float8 storage lookup are separate compatibility bugs." + "reason": "Unexpected `_is_hf_initialized` argument during init-empty-weights is unrelated to missing Float8 storage registration." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated." + "reason": "`ModelOutput` key assignment bug and Gemma-4 FSDP2 `shared_kv_states` `KeyError` are different problems." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "High perplexity on GPT-OSS is unrelated to Whisper batch_decode skipping special tokens." + "reason": "High perplexity on a benchmark model is unrelated to Whisper `batch_decode` ignoring `skip_special_tokens`." }, { "left": "issue:42796", "right": "issue:46710", "accept": false, - "reason": "A tokenizer BPE-marker bug is unrelated to a DeepSeek-R1 accuracy regression after upgrade." + "reason": "Tokenizer BPE-marker handling is unrelated to a downstream accuracy regression in DeepSeek-R1-Distill-Llama-8B." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Qwen2.5-GGUF loading and missing pad_token_id on Llama4Config are different loading issues." + "reason": "Qwen2.5-GGUF loading under v5 and Llama4 `pad_token_id` missing are distinct model/config loading issues." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "inputs_embeds/past_key_values progressive generation is unrelated to Qwen2.5-VL multi-image IterableDataset finetuning." + "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to finetuning with multiple images per prompt." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading under v5 and CLIPTokenizer max-length behavior are unrelated." + "reason": "GGUF loading failure and CLIPTokenizer `model_max_length` misuse are different tokenizer/model-loading bugs." }, { - "left": "issue:43632", - "right": "issue:44534", + "left": "issue:41553", + "right": "issue:43582", "accept": false, - "reason": "A broken _is_hf_initialized flag is not the same as non-persistent buffers being filled with junk." + "reason": "AutoTokenizer error messaging for Voxtral is unrelated to Apple Silicon `caching_allocator_warmup` `TypeError`." }, { - "left": "issue:41553", - "right": "issue:43582", + "left": "issue:43632", + "right": "issue:44534", "accept": false, - "reason": "AutoTokenizer error messaging for Voxtral and AppleSilicon caching_allocator_warmup TypeError are unrelated." + "reason": "These are separate v5 regressions affecting different internals: HF init flags vs non-persistent buffers." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but one soft-edge verdict is too strong. `issue:44479` and `issue:45325` are related in topic/version, yet the packet only shows a broad overlap around Qwen VL / temporal-position handling; it does not clearly establish they are the same bug, especially since one is a video-input regression across several models and the other is a still-image temporal position_ids scaling issue in qwen2.5-vl. That pair should stay conservative/rejected unless there is stronger evidence. The Qwen3.5 `save_pretrained` pair (`45216`/`45357`) looks plausible, and the rest of the rejects are well supported." + "accept": true, + "feedback": "Grounded overall and conservative in the soft-edge calls. The only accepted pair, issue:45216 vs issue:45357, is reasonably supported by the titles as the same Qwen3.5 save_pretrained regression path, and the rest of the rejections stay on the safe side without overclaiming duplicates." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -3808,8 +3805,8 @@ "issue:42175|issue:45542", "issue:43421|issue:44568", "issue:44829|issue:45479", - "issue:41720|issue:45081", "issue:44079|issue:44534", + "issue:41720|issue:45081", "issue:43818|issue:46402", "issue:44462|issue:45812", "issue:43818|issue:45357", @@ -3819,8 +3816,8 @@ "issue:42175|issue:43296", "issue:43749|issue:44079", "issue:44265|issue:45072", - "issue:43454|issue:44466", "issue:43792|issue:43957", + "issue:43454|issue:44466", "issue:45325|issue:45381", "issue:43493|issue:43994", "issue:40444|issue:45381", @@ -3829,24 +3826,24 @@ "issue:44291|issue:44877", "issue:43632|issue:45072", "issue:43873|issue:45701", - "issue:42796|issue:45381", "issue:43701|issue:45325", + "issue:42796|issue:45381", "issue:43296|issue:45800" ], - "prepared_review_unit_hash": "97b57966be4512734e9325e140dba72cf1664dec325fb169b37e39439e87a3b5", + "prepared_review_unit_hash": "81a98c3c684600504f736b889d2648128c660c7f77e7009956b62e44545ca7c5", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36009, + "serialized_chars": 36010, "estimated_input_tokens": 9003, "estimated_eval_tokens": 18262 }, @@ -3855,196 +3852,178 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous regression reports with two plausible duplicates: the SigLIP2 output-mismatch pair, and a Qwen2.5-VL positional-index pair that is similar but not strong enough to merge as one bug. The cluster\u2019s best representative remains the broad Transformers v5 initialization/loading regression theme.", - "confidence": 0.54, - "canonical_issue_reason": "Issue 43646 is the broadest representative of the dominant cluster theme: Transformers v5 initialization/loading regressions impacting model setup behavior across many models.", + "summary": "This cluster is mostly heterogeneous; the soft links are largely topical rather than true duplicates. The safest anchor is the high-activity quantization/offloading report #43873, but the only clearly conservative stance on the proposed soft edges is to reject them absent stronger body-level overlap.", + "confidence": 0.74, + "canonical_issue_reason": "#43873 has the strongest discussion/inbound-reference signal and is a broad, representative issue in an otherwise mixed cluster, so it is the best canonical issue anchor for triage.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 best captures the cluster\u2019s central pattern as a broad v5 initialization regression, rather than a narrow model-specific symptom.", + "best_issue_reason": "#43873 is the highest-signal issue here; it is broad enough to serve as the cluster representative even though the set is not a tight duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Different failures: device-map/offload loading error vs MoE training memory inefficiency." + "reason": "Different bugs: device_map/offload loading failure vs Qwen3 MoE training memory usage." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding-window bug vs empty-weights init argument mismatch; unrelated code paths." + "reason": "Both are v5 regressions, but they affect different code paths and symptoms; not the same concrete bug." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TensorFlow packaging, but one is missing TF in the torch extra and the other is a tensorboard-only backend error." + "reason": "Both mention TF/backends, but one is packaging/install coverage and the other is a runtime backend error; not a duplicate." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime post-processor refresh bug vs tokenizer special-token insertion bug; different layers." + "reason": "Related tokenizer/runtime symptoms, but the packet does not establish the same underlying bug; keep this conservative." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Different training regressions: flash-attention degeneration vs single-label zero-loss behavior." + "reason": "Different failure modes: flash_attention_3 training degeneration vs single-label zero-loss behavior." }, { - "left": "issue:41720", - "right": "issue:45081", + "left": "issue:44079", + "right": "issue:44534", "accept": false, - "reason": "Auto device mapping cuda assert vs tokenizer regex patch AttributeError; unrelated." + "reason": "ModelOutput key assignment and non-persistent buffer junk filling are unrelated regressions." }, { - "left": "issue:44079", - "right": "issue:44534", + "left": "issue:41720", + "right": "issue:45081", "accept": false, - "reason": "ModelOutput key assignment bug vs non-persistent buffer initialization junk; different internals." + "reason": "Different models and code paths: auto device-map cuda assert vs Mistral regex patch crash." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Video-LLaVA vision-tower/temporal-attention issue vs Sam3 text-encoder weight loading regression." + "reason": "Video-LLaVA temporal-attention/weight issue is unrelated to Sam3Model text-encoder loading." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Repository tokenizer.json selection bug vs Granite token-id regression; not the same underlying issue." + "reason": "Tokenizer-loading and tokenizer-id regression are only loosely related; titles/metadata do not establish a shared concrete bug." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Different models and failures: Video-LLaVA architecture/weights vs Qwen3.5 visual-encoder save_pretrained keys." + "reason": "Different model-specific regressions: Video-LLaVA tower/temporal attention vs Qwen3.5 visual-encoder save_pretrained keys." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Related v5 breakages, but base_model_prefix loading and custom model initialization are distinct bugs." + "reason": "Both are v5 init/load issues, but they are distinct bugs affecting different initialization behavior." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both involve batching effects, but they hit different models and different attention/mask paths." + "reason": "Missing padding-mask support in MIMI is not the same as Gemma3 mixed-length NaNs." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer filling vs ModelOutput key assignment; unrelated regression types." + "reason": "Non-persistent buffer junk filling is unrelated to ModelOutput key assignment." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Backend packaging issue vs PaddleOCR-VL load failure in vLLM; different problems." - }, - { - "left": "issue:43749", - "right": "issue:44079", - "accept": false, - "reason": "FSDP RAM-efficient loading breakage vs ModelOutput key assignment bug; different subsystems." + "reason": "Backend packaging/install issues are unrelated to PaddleOCR-VL model loading in vLLM." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export / torch_compilable_check failure vs dtype mismatch in bfloat16 inference; not the same code-path problem." - }, - { - "left": "issue:43454", - "right": "issue:44466", - "accept": false, - "reason": "Tied lm_head weights not being tied vs inconsistent lm_head serialization by device; related theme, but distinct bugs." + "reason": "torch.export/torch_compilable_check failure and bf16 dtype mismatch are different problems." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper runtime failure vs meta-device model loading regression; no concrete overlap." + "reason": "Whisper loading failure and meta-device loading regression are not the same bug." }, { - "left": "issue:45325", - "right": "issue:45381", + "left": "issue:43454", + "right": "issue:44466", "accept": false, - "reason": "Both are Qwen2.5-VL positional regressions, but one is still-image temporal position_ids and the other is video-input vision_position_ids; insufficient evidence they are the same bug." - }, - { - "left": "issue:43493", - "right": "issue:43994", - "accept": true, - "reason": "Both describe the same SigLIP2 HF-vs-original implementation mismatch that yields wrong outputs." + "reason": "Both mention lm_head/tied weights, but one is a model-specific tie omission and the other is device-dependent serialization." }, { - "left": "issue:40444", + "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Multiple-images-per-prompt finetuning failure vs video-input position_id bug; same family, different concrete bugs." + "reason": "These are related Qwen2.5-VL position-id regressions, but the packet does not establish that they are the same underlying defect." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Qwen2.5-VL position-id regression vs Kosmos2.5 long-OCR index error; unrelated models and paths." + "reason": "Qwen2.5-VL position-id scaling and Kosmos2.5 OCR long-input index error are unrelated." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3MoE FP8 config failure vs GGUF loading failure for Qwen2.5; different quantization/loading issues." + "reason": "Qwen3MoE FP8 failure and Qwen2.5-GGUF v5 loading failure are unrelated." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Empty-weights init argument mismatch vs strict config validation rejecting granite_speech config; different causes." + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different regressions." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "_is_hf_initialized regression vs dtype mismatch in bfloat16 inference; unrelated." + "reason": "_is_hf_initialized handling and bf16 dtype mismatch are unrelated." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading behavior vs tokenization version change; not the same bug." + "reason": "Quantized offloading behavior and tokenizer changes across versions are different issues." }, { - "left": "issue:42796", - "right": "issue:45381", + "left": "issue:43701", + "right": "issue:45325", "accept": false, - "reason": "Tokenizer BPE-marker handling vs Qwen2.5-VL position IDs; unrelated." + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL position-id scaling." }, { - "left": "issue:43701", - "right": "issue:45325", + "left": "issue:42796", + "right": "issue:45381", "accept": false, - "reason": "resume_from_checkpoint key mismatch vs Qwen2.5-VL rope-index scaling bug; different failure surfaces." + "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video position ids are unrelated." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM vs torch/transformers version incompatibility; different issues." + "reason": "PaddleOCR-VL loading in vLLM and torch/transformers version incompatibility are not the same bug." } ] }, "evaluator_result": { "accept": false, - "feedback": "The cluster summary is mostly grounded, but it overreaches on at least one soft edge: issue:45325 vs issue:45381 are both Qwen2.5-VL position-id regressions, but the packet shows different surfaces (still-image temporal position_ids vs video-input vision_position_ids) and not enough evidence to call them the same bug. The SigLIP2 pair looks plausible, but the accepted Qwen2.5-VL duplicate verdict is too strong for the evidence provided." + "feedback": "The triage summary is mostly grounded, but several soft-edge acceptances are too strong for the evidence in the packet. In particular, 43421\u219444568, 44462\u219445812, and 45325\u219445381 are not established duplicates from titles/metadata alone; they look related at best, and the packet\u2019s soft-pair signal is explicitly non-deterministic/negative. Please keep those verdicts conservative (false or maybe-related, not accepted) unless there is stronger body-level evidence." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -4091,16 +4070,15 @@ "issue:43883", "issue:44186", "issue:44265", - "issue:44387", "issue:44479", "issue:44488", "issue:44534", "issue:44568", "issue:44610", "issue:44898", + "issue:44991", "issue:45072", "issue:45584", - "issue:45704", "issue:45759", "issue:45823", "issue:45920", @@ -4140,216 +4118,216 @@ "issue:41720|issue:43322", "issue:43526|issue:43883", "issue:36331|issue:42617", - "issue:44387|issue:45704" + "issue:43881|issue:44991" ], - "prepared_review_unit_hash": "6fbd3632b0da1ec26a8b4f1ab463c40e10870b988da5b95faea9d5d35efa899a", + "prepared_review_unit_hash": "fa3c2e8c5ac34900f39046796677e2e813f7d5264d13399eff03257ac50692a5", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36857, - "estimated_input_tokens": 9215, - "estimated_eval_tokens": 18686 + "serialized_chars": 36268, + "estimated_input_tokens": 9067, + "estimated_eval_tokens": 18390 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are model-specific regressions, loading/offload failures, tokenizer/processor bugs, or doc/example issues. The soft-similarity edges mostly reflect shared Transformers v5 regression language or broad subsystem overlap, not the same concrete bug.", - "confidence": 0.67, - "canonical_issue_reason": "Issue 43873 is the broadest and most central representative here: it covers offloading + quantization, which is the closest recurring theme among the loading/device-mapping regressions in this set.", + "summary": "The set is a grab-bag of unrelated issues across tokenization, loading, quantization, multimodal processors, training, and offloading. None of the soft pairs look like true duplicates or the same concrete fix; they mostly share only broad subsystem keywords. The broadest and most discussion-heavy issue is 43873, but it is still not a duplicate of the others.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43873 is the broadest active report in the set and has the most follow-on discussion/references around offloading/quantization, so it is the best umbrella candidate, though it does not actually duplicate the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "43873 is the best single issue anchor for this cluster because it is general, current, and closer to the repeated memory/loading/offload failures than the highly model-specific reports.", + "best_issue_reason": "issue:43873 is the most representative/open issue by scope and activity, but only as a broad umbrella; it is not a true duplicate target for the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Different models and failure modes: a video-input regression across Qwen/VL variants versus a Gemma4 PLE device-mismatch forward bug." + "reason": "Different models and failures: a video-input regression in Qwen VL family vs a Gemma4 PLE device-mismatch during forward." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "Bark CPU-offload device mismatch is unrelated to torch.export failures caused by torch_compilable_check." + "reason": "Bark CPU-offload device mismatch and torch.export failure with torch_compilable_check are unrelated code paths." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sample-rate default bug is unrelated to GraniteSpeech PEFT local-checkpoint loading." + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are distinct bugs." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "High perplexity on gpt-oss and Whisper empty-transcription generation are different bugs with different code paths." + "reason": "Perplexity on gpt-oss-20b and Whisper empty-transcription generation failure do not share the same underlying defect." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "inputs_embeds/past_key_values generation bug is not the same as Qwen3 auto device-map cuda assert." + "reason": "Inputs_embeds/past_key_values generation issue is unrelated to Qwen3 auto device-map cudaErrorAssert." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice_preset failure and progressive generation with inputs_embeds are unrelated issues." + "reason": "BarkProcessor voice_preset failure is a different feature path from progressive generation with past_key_values." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "glm-4v-9b loading failure and cjvt/sleng-bert loading failure are separate model-specific load problems." + "reason": "Both are model-loading complaints, but they concern different models and likely different tokenizer/model initialization failures." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision lm_head tying bug and Bamba torch.compile+SDPA bug affect different model code paths." + "reason": "Weight-tying in AyaVision and torch.compile+SDPA failure in Bamba-9B-v2 are unrelated." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "CPU-offload device mismatch and DAC latent/forward mismatch are unrelated." + "reason": "CPU-offload device mismatch in Bark-small is not the same as DAC from_latents vs forward mismatch." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc/example mismatch and Qwen2.5-VL iterable multiple-image finetuning failure are different problems." + "reason": "A documentation/example mismatch in SigLIP2 is unrelated to Qwen2.5-VL fine-tuning with iterable multi-image batches." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer NER/padding crash is unrelated to export failures from torch_compilable_check." + "reason": "LayoutLMv2 tokenizer crashes and torch.export failures are different subsystems and failures." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Qwen3 auto device-map cuda assert is not the same as Gemma4 PLE device mismatch during forward." + "reason": "Both mention device_map=auto, but the concrete failures and models differ; not the same bug." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "MoonshineStreaming off-by-one masking bug and slow convert_tokens_to_ids regression are unrelated." + "reason": "A sliding-window mask off-by-one is unrelated to tokenizer speed regression." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI encoder batching/padding-mask bug is unrelated to OmDet-Turbo processor size mismatch." + "reason": "MIMI padding-mask behavior and OmDet-Turbo image-size mismatch are unrelated." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Qwen3 MoE accelerate+deepspeed weight conversion issue is distinct from missing supports_gradient_checkpointing on Nemotron remote-code variant." + "reason": "Qwen3 MoE weight conversion/loading and a missing gradient-checkpointing flag are distinct issues." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "mdeberta add_special_tokens regression and AutoTokenizer wrong-token-id regression are different tokenizer issues." + "reason": "Both are tokenizer regressions, but one is missing BOS/EOS for a specific tokenizer while the other is wrong token IDs across several models; not the same concrete fix." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "AyaVision weight-tying bug is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel bfloat16 inference." + "reason": "lm_head weight tying and bfloat16 dtype mismatches are unrelated." }, { "left": "issue:42796", "right": "issue:46209", "accept": false, - "reason": "Tokenizer BPE marker handling and FP8 inference on B200 are not the same bug." + "reason": "Tokenizer BPE-marker handling and FP8 inference on B200 are unrelated." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both involve offload/loading, but one is dequantized model loading on gpu+cpu offload while the other is generic quantized offloading; too broad to treat as one bug." + "reason": "Both involve loading/offloading, but one is dequantized model loading with device_map=auto and the other is a broader quantization/offload issue; not clearly the same bug." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "AyaVision lm_head tying bug and BitNet packed-weight unpacking bug are unrelated." + "reason": "AyaVision weight tying and BitNet packed-weight unpacking during accelerate loading are different code paths." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "MIMI encoder padding-mask issue and Perceiver non-default-resolution failure are different model behaviors." + "reason": "MIMI batched-padding-mask bug and Perceiver non-default-resolution failure are unrelated model components." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity regression and gradient checkpointing with PEFT LoRA are unrelated." + "reason": "High perplexity on a language-model eval and ineffective gradient checkpointing with LoRA are different problems." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py run failure and Qwen3Moe FineGrainedFP8Config failure are distinct startup/config problems." + "reason": "3d_parallel.py execution failure and Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk fill are different v5 regressions." + "reason": "FSDP CPU-efficient loading and non-persistent buffer initialization junk are not the same concrete defect." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "StableLmConfig pad_token_idx missing is unrelated to AutoModelForCausalLM.from_config text_config unwrapping." + "reason": "Missing pad_token_idx in a config and failure to unwrap text_config for composite multimodal configs are distinct." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Llava Next loading segfault and sinq save/load breakage are different failures." + "reason": "Llava Next segfault on load and sinq quantized save/load breakage are unrelated." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 cuda assert under auto device map is unrelated to Llava Next load-time segfault." + "reason": "Different models and symptoms: Qwen3 auto device-map cudaErrorAssert vs Llava Next loading segfault." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to Molmo missing all_tied_weights_keys." + "reason": "BeitImageProcessorFast reduce_labels bug and Molmo missing all_tied_weights_keys are unrelated." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "CustomTrainer compute_loss signature regression and 3d_parallel.py execution failure do not share the same bug." + "reason": "Trainer compute_loss signature mismatch and 3d_parallel.py runtime failure are different issues." }, { - "left": "issue:44387", - "right": "issue:45704", + "left": "issue:43881", + "right": "issue:44991", "accept": false, - "reason": "Int4 reserved-memory OOM regression and T5 apex.FusedRMSNorm memory leak are separate memory issues." + "reason": "GLM-4V loading failure and EMBEDDIA/est-roberta tokenizer loading regression are separate model/tokenizer problems." } ] }, @@ -4396,10 +4374,9 @@ "issue:43708", "issue:43720", "issue:43873", - "issue:43881", "issue:44008", "issue:44186", - "issue:44361", + "issue:44387", "issue:44423", "issue:44466", "issue:44610", @@ -4417,13 +4394,14 @@ "issue:45632", "issue:45684", "issue:45701", + "issue:45704", "issue:45823", "issue:46291", "issue:46402", "issue:46665" ], "soft_pairs": [ - "issue:43881|issue:44991", + "issue:44387|issue:45704", "issue:44466|issue:45357", "issue:41720|issue:44977", "issue:36331|issue:42915", @@ -4432,9 +4410,9 @@ "issue:40990|issue:45701", "issue:43065|issue:46402", "issue:44186|issue:45072", - "issue:26413|issue:28530", "issue:43873|issue:44756", "issue:41720|issue:42617", + "issue:26413|issue:28530", "issue:43550|issue:43720", "issue:43577|issue:43708", "issue:43404|issue:45357", @@ -4442,9 +4420,9 @@ "issue:43550|issue:43606", "issue:43646|issue:46665", "issue:44186|issue:44610", - "issue:42915|issue:43317", - "issue:44008|issue:45823", "issue:44610|issue:44898", + "issue:44008|issue:45823", + "issue:42915|issue:43317", "issue:36010|issue:42617", "issue:44423|issue:45406", "issue:45632|issue:45684", @@ -4453,229 +4431,222 @@ "issue:43531|issue:43572", "issue:44671|issue:44743", "issue:42915|issue:46291", - "issue:36331|issue:41720", - "issue:44186|issue:44361" + "issue:36331|issue:41720" ], - "prepared_review_unit_hash": "2b72883cde9334addf7bb66a6b27e7ffddb490bcc1aa21dcf59584a0121219e8", + "prepared_review_unit_hash": "60a5a47097de6ebc1094993e94a11b6688bd34b67cc03a6b694511c4051fca14", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37319, - "estimated_input_tokens": 9330, - "estimated_eval_tokens": 18916 + "soft_pair_count": 31, + "serialized_chars": 37112, + "estimated_input_tokens": 9278, + "estimated_eval_tokens": 18812 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes training/checkpoint regressions, tokenizer/processor bugs, quantization/offloading failures, and several model-specific multimodal issues. The soft pairs are mostly adjacent-sounding but do not describe the same concrete bug path, so they should stay separate.", + "summary": "The cluster is highly heterogeneous: most links are superficial topical overlaps (same model family, same subsystem, or same symptom class) rather than true duplicates. I would not merge any of the soft-edge pairs.", "confidence": 0.93, - "canonical_issue_reason": null, + "canonical_issue_reason": "No clear canonical issue exists for this cluster; the items span unrelated bugs across training, loading, quantization, processors, serving, and model-specific regressions, so the similarity edges are too weak to justify a single duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best umbrella issue in this set: it is active, broadly scoped around quantization/offloading, and already has inbound references, while the other items are narrower and mostly unrelated regressions.", + "best_issue_reason": "issue:45701 is the broadest and most central-looking report in the set (generic version/tokenization regression with multiple references), but it still only fits a small subset of the items and is not a true umbrella for the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:44387", + "right": "issue:45704", "accept": false, - "reason": "Different failure modes: model loading failure vs tokenizer-loading regression." + "reason": "Both are memory-related, but one is an int4 quantization/OOM regression and the other is a T5/apex RMSNorm leak; different models and different code paths." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both involve saving/serialization, but one is about tied lm_head weights and the other about incorrect visual encoder keys." + "reason": "Both involve save/serialization, but one is tied-weight serialization and the other is incorrect visual encoder keys for Qwen3.5; not the same bug." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Same Qwen family, but multi-GPU training is unrelated to flash-attention generation failure." + "reason": "Same Qwen family, but multi-GPU training failure and flash-attention generation failure are distinct issues with different failure modes." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Trainer API kwarg mismatch vs MoE/FP8 model loading failure; not the same bug." + "reason": "A trainer API compatibility error and a FineGrainedFP8 model failure are unrelated." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Config strictness preventing a model load is not the same as tokenizer loading failing on a specific checkpoint." + "reason": "Both are loading regressions, but they affect different objects (config vs tokenizer) and different failure causes." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video batch shape bug vs tokenization changes; different code paths." + "reason": "Batched video processor shape handling is unrelated to a broad tokenization-behavior change." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Perplexity anomaly on a model is not the same as a tokenizer-version regression." + "reason": "High perplexity may be influenced by tokenization, but this is not the same concrete regression as a general tokenization change." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Both are Sam3-related, but pixel-decoder dummy Conv2d and missing text-encoder weights are distinct bugs." + "reason": "Both touch Sam3, but one is a dummy Conv2d in the pixel decoder and the other is missing text encoder weights; different components." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batching vs dtype mismatch in inference; unrelated defects." - }, - { - "left": "issue:26413", - "right": "issue:28530", - "accept": false, - "reason": "Both touch Trainer control flow, but checkpoint resume failure and early-stopping metric detection are separate issues." + "reason": "Tokenizer crashes and dtype mismatches are unrelated." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization/offloading problems, but one is a general offload regression and the other is mmap/OOM behavior on Strix Halo." + "reason": "Both mention memory/offload, but one is quantization offloading behavior and the other is mmap-related OOM on a specific platform." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL is not the same as a 3d_parallel script execution failure." + "reason": "Different features and failures; shared context is only that both are training/runtime problems." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "Checkpoint resumption bug and early-stopping metric lookup bug are separate trainer behaviors." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile/SDPA behavior is unrelated to accelerated loading not unpacking packed BitNet weights." + "reason": "One is a torch.compile/SDPA model bug; the other is an accelerate loading issue with packed BitNet weights." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Loading dtype propagation in Blip2 is separate from checkpoint-resume max_steps miscalculation." + "reason": "Model dtype loading and trainer max_steps recomputation are unrelated." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Tied-weight loading in Mistral3 is a different serialization bug from incorrect visual encoder save keys." + "reason": "Tied lm_head weights in Mistral3 and Qwen3.5 visual encoder key saving are different model-specific serialization problems." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor/tokenizer API breakage is not the same as a general tokenization-version change." + "reason": "Processor/tokenizer API regression is not the same as a general version-caused tokenization change." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Both are runtime/inference issues, but compile+SDPA and CPU offload device mismatch are distinct paths." + "reason": "Both are device/compile related, but one is SDPA under compile and the other is CPU offload device mismatch." }, { "left": "issue:43646", "right": "issue:46665", "accept": false, - "reason": "Related custom-model API area, but one is custom model initialization under v5 and the other is from_pretrained behavior inside Python classes." + "reason": "Both involve model initialization/loading, but they describe different failure surfaces and likely different fixes." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crashes and image-size mismatches are different bugs on different processors/models." + "reason": "Tokenizer padding/truncation crash and image-size mismatch are distinct preprocessing bugs." }, { - "left": "issue:42915", - "right": "issue:43317", + "left": "issue:44610", + "right": "issue:44898", "accept": false, - "reason": "Both are loading/offload-related, but FP8 config failure and dequantized model offload failure are not the same concrete defect." + "reason": "Both relate to image resolution, but one is a processor/model size mismatch and the other is a Perceiver interpolation failure." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma multimodal bugs, but a variable-name collision in forward is unrelated to device_map auto device mismatch." + "reason": "Different Gemma generations and different bugs: variable collision in forward vs PLE device mismatch with auto device mapping." }, { - "left": "issue:44610", - "right": "issue:44898", + "left": "issue:42915", + "right": "issue:43317", "accept": false, - "reason": "Both involve image resolution expectations, but they affect different models and different preprocessing logic." + "reason": "Both involve quantized/dequantized loading, but the concrete failure modes and models differ." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "ImportError for GenerationMixin and a 3d_parallel runtime issue are unrelated." + "reason": "Importing GenerationMixin and running 3d_parallel.py are unrelated." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both hit serve with multimodal models, but the failing attributes and processors differ." + "reason": "Both crash in serving paths, but with different processors, symptoms, and root causes." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision and read-only permission propagation are separate save/load concerns." + "reason": "One is a cache-path collision, the other is permission propagation during save_pretrained; not the same bug." }, { - "left": "issue:41720", - "right": "issue:45161", + "left": "issue:45701", + "right": "issue:45704", "accept": false, - "reason": "Both mention Qwen/MoE-ish scaling, but multi-GPU training and tensor-parallel failure are different problems." + "reason": "A tokenization regression and a T5 RMSNorm memory leak are unrelated." }, { - "left": "issue:28530", - "right": "issue:33666", + "left": "issue:45161", + "right": "issue:41720", "accept": false, - "reason": "Early stopping metric lookup and Qwen2-VL multi-GPU training are unrelated." + "reason": "Tensor-parallel-only failure for GPT-OSS MoE is not the same as Qwen2-VL multi-GPU training." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to missing pad_token_idx in StableLmConfig." + "reason": "Sliding-window behavior in Qwen3-MoE and a missing StableLmConfig pad_token_idx are different issues." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "Masked-LM prediction regression and recurrent-state reset are different model behaviors." + "reason": "CamemBERT masked-LM regression and recurrent-state reset in modular_qwen3_5 are unrelated." }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "Model loading failure vs pipelines import dependency/version requirement; not the same bug." + "reason": "A model/quantization failure and a kernels package import requirement change are not duplicates." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer compute_loss signature breakage is unrelated to Qwen2-VL multi-GPU training." - }, - { - "left": "issue:44186", - "right": "issue:44361", - "accept": false, - "reason": "Both are tokenizer AttributeErrors, but they occur in different tokenizers and different tasks." + "reason": "Trainer loss-signature incompatibility is unrelated to multimodal multi-GPU training." } ] }, @@ -4719,7 +4690,6 @@ "issue:43540", "issue:43550", "issue:43606", - "issue:43644", "issue:43653", "issue:43720", "issue:43746", @@ -4727,6 +4697,8 @@ "issue:43873", "issue:43881", "issue:44077", + "issue:44186", + "issue:44361", "issue:44387", "issue:44451", "issue:44492", @@ -4740,7 +4712,6 @@ "issue:45357", "issue:45406", "issue:45464", - "issue:45479", "issue:45701", "issue:45704", "issue:45759", @@ -4749,19 +4720,20 @@ "issue:46665" ], "soft_pairs": [ + "issue:44186|issue:44361", "issue:43526|issue:43881", - "issue:42915|issue:43322", "issue:43316|issue:44913", + "issue:42915|issue:43322", "issue:28530|issue:30064", "issue:33357|issue:35707", - "issue:43873|issue:45704", "issue:30064|issue:34634", - "issue:43540|issue:43782", - "issue:29159|issue:36032", + "issue:43873|issue:45704", "issue:43550|issue:43653", - "issue:44492|issue:44509", + "issue:29159|issue:36032", + "issue:43540|issue:43782", "issue:44387|issue:45701", "issue:43606|issue:43720", + "issue:44492|issue:44509", "issue:44534|issue:46665", "issue:43531|issue:45759", "issue:41720|issue:43746", @@ -4775,212 +4747,210 @@ "issue:43531|issue:44589", "issue:44451|issue:44991", "issue:45759|issue:45859", - "issue:44077|issue:44625", - "issue:44625|issue:45479", - "issue:43644|issue:46665" + "issue:44077|issue:44625" ], - "prepared_review_unit_hash": "8e7d42dec0ff5598be90f1d063d69aff7b71c83ac1ecf093d8305911efd933df", + "prepared_review_unit_hash": "10bf255cf011f863e35bbcfb6c7b2c36f2e108b724009b69a9e128cd447122d0", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36375, - "estimated_input_tokens": 9094, - "estimated_eval_tokens": 18444 + "soft_pair_count": 28, + "serialized_chars": 36010, + "estimated_input_tokens": 9003, + "estimated_eval_tokens": 18262 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair appears to be the same underlying bug/change; the similarities are mostly broad subsystem overlap or generic regression wording. I would not merge any of these issue pairs.", - "confidence": 0.94, - "canonical_issue_reason": "issue:45479 is the strongest standalone issue: it is concrete, reproducible, and clearly describes a specific failure mode with broad applicability across multiple sequence-classification models.", + "summary": "The cluster is a loose mix of superficially related Transformers issues (tokenizers, configs, loading/offload, multimodal models), but none of the soft pairs look like the same underlying bug or change. I rejected all soft edges.", + "confidence": 0.66, + "canonical_issue_reason": "`issue:44534` is the broadest representative of the cluster: a Transformers v5 runtime regression affecting core model state handling. Even so, the cluster is too heterogeneous to treat as one true duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45479 is the best issue candidate in this set because it is precise, well-scoped, and actionable, making it a good representative issue even though the items are not duplicates.", + "best_issue_reason": "`issue:44534` is the most generally framed and potentially high-impact issue among an otherwise unrelated set, so it is the best single issue to stand in as the cluster anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43526", - "right": "issue:43881", + "left": "issue:44186", + "right": "issue:44361", "accept": false, - "reason": "Different bugs: BeitImageProcessorFast label reduction vs GLM-4V load failure; no shared code path." + "reason": "Both are tokenizer failures, but they hit different tokenizer classes and different paths: NER/batched padding crash vs task-time AttributeError." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:43526", + "right": "issue:43881", "accept": false, - "reason": "One is a Qwen3Moe FP8 config failure, the other a Llava Next segmentation fault; unrelated model-loading problems." + "reason": "`reduce_labels` in BeitImageProcessorFast is unrelated to a GLM-4V model loading failure; different component and failure mode." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API mismatch vs GPTNeoX rotary_pct reload behavior; different config fields and failure modes." + "reason": "Both are config-related, but one is a Gemma3TextConfig API mismatch and the other is GPTNeoX rotary_pct not persisting on reload\u2014different fields, different bug." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated model-loading/runtime problems." }, { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early stopping metric selection and void segmentation-map processing are unrelated." + "reason": "Early stopping metric selection and void segmentation map processing are completely separate features." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error and generation with inputs_embeds/past_key_values are different runtime issues." - }, - { - "left": "issue:43873", - "right": "issue:45704", - "accept": false, - "reason": "Quantization/offloading behavior vs T5 RMSNorm memory leak; different code paths and symptoms." + "reason": "MacOS bus error with a CLIP model vs generation with inputs_embeds/past_key_values; no shared code path." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "ImageProcessor segmentation-map handling is not the same as BarkProcessor voice_preset parsing." + "reason": "Image processor void segmentation maps and BarkProcessor voice_preset are unrelated processor bugs." }, { - "left": "issue:43540", - "right": "issue:43782", + "left": "issue:43873", + "right": "issue:45704", "accept": false, - "reason": "Video-input ValueError in Qwen3OmniMoe is distinct from Qwen3VL weight_only loading failure." + "reason": "Quantization/offloading behavior and an apex RMSNorm memory leak are different runtime issues with different root causes." }, { - "left": "issue:29159", - "right": "issue:36032", + "left": "issue:43550", + "right": "issue:43653", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is not the same as T5 add_special_tokens method name conflict." + "reason": "torch.compile + SDPA on Bamba-9B-v2 is unrelated to BigBirdTokenizer special-token registration/empty decode behavior." }, { - "left": "issue:43550", - "right": "issue:43653", + "left": "issue:29159", + "right": "issue:36032", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and BigBirdTokenizer special-token registration are unrelated." + "reason": "Slow/fast tokenizer inconsistency is a different bug from T5Tokenizer's add_special_tokens method conflict during load." }, { - "left": "issue:44492", - "right": "issue:44509", + "left": "issue:43540", + "right": "issue:43782", "accept": false, - "reason": "A typo in cache strategies is not the same as stale documentation for removed pipeline tasks." + "reason": "One is a video-input processing error in Qwen3OmniMoe; the other is a Qwen3VL from_pretrained weight_only loading error." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization memory growth and tokenization changes are different regressions." + "reason": "CUDA reserved-memory regression under int4 quantization is not the same as a tokenization change across versions." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both involve loading/offload, but the affected models and concrete failures differ; too broad to treat as one bug." + "reason": "Both involve loading/offload, but one is a CPU-offload device mismatch for Bark and the other is missing unpacking of packed BitNet weights during accelerate loading." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategy typo and stale docs mentioning removed pipeline tasks are both doc-level issues, but not the same defect." }, { "left": "issue:44534", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer initialization junk and from_pretrained behavior inside Python classes are not the same defect." + "reason": "Non-persistent buffer initialization junk in v5 and `from_pretrained` behavior inside Python classes are separate problems with different mechanics." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and multimodal config unwrapping are separate issues." + "reason": "Qwen3-MoE sliding_window behavior and unwrapping `text_config` in composite Qwen configs are different config/code-path bugs." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert is unrelated to GraniteSpeech PEFT local-checkpoint loading." + "reason": "Auto device mapping CUDA assert on A800 is unrelated to GraniteSpeech PEFT local-checkpoint loading failures." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash import warning and undefined variables in multimodal token counting are different code paths." + "reason": "A trailing-slash import warning is not the same as an undefined-function bug in the video branch of multimodal token counting." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "gather_for_metrics batch truncation is unrelated to SAM3 text/bbox behavior." + "reason": "Dropping tuple-label elements in `gather_for_metrics` is unrelated to SAM3 text/bounding-box behavior." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "CPU offload device mismatch is not the same as BigBirdTokenizer decode output failure." + "reason": "Bark CPU-offload device mismatch and BigBirdTokenizer special-token registration are different bugs in different subsystems." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve and Qwen3.5 streaming inference failure are different service bugs." + "reason": "Gemma4Processor missing `_tokenizer` and Qwen3.5 streaming chat/completions failure are distinct API/runtime issues." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token-count tracking and progressive generation with cached keys are unrelated." + "reason": "`num_input_tokens_seen` not updating is a TrainerState bookkeeping bug, not a generation bug involving embeds and cache state." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset parsing is not the same as Qwen3 A800 device-mapping failure." + "reason": "BarkProcessor voice_preset failure and Qwen3 auto device mapping cudaErrorAssert are unrelated." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Incorrect visual encoder keys on save_pretrained and text encoder weights not loading are different multimodal persistence bugs." + "reason": "Incorrect Qwen3.5 visual-encoder keys on save and Sam3Model text-encoder weights not loading are both multimodal, but they affect different models and directions (save vs load)." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "A policy change around optional `post_init` and Qwen3.5 `num_labels` propagation are separate config-validation issues." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Sliding_window behavior in Qwen3-MoE is unrelated to the Float8 storage lookup error." + "reason": "Qwen3-MoE sliding_window behavior and missing Float8 storage objects are unrelated." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Two tokenizer-loading regressions, but for different models and failure mechanisms; not the same underlying bug." + "reason": "Loading `vesteinn/ScandiBERT` and tokenizer loading for EMBEDDIA/est-roberta are both model-loading reports, but they do not share the same failure mode or code path." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Config unwrapping for composite Qwen multimodal models is different from CPU-offload compatibility in Qwen3_5MoeTextRotaryEmbedding." + "reason": "Both are Qwen3.5/3.6-related config bugs, but one is about `from_config` unwrapping and the other is CPU-offload compatibility in rotary embedding forward; different fixes." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Optional post_init validation and num_labels propagation are separate configuration problems." - }, - { - "left": "issue:44625", - "right": "issue:45479", - "accept": false, - "reason": "Both touch num_labels, but one is Qwen3.5 config propagation while the other is a classification-loss edge case across models." - }, - { - "left": "issue:43644", - "right": "issue:46665", - "accept": false, - "reason": "Non-persistent buffer junk in v5 and class-wrapped from_pretrained inconsistency are not the same defect." + "reason": "`post_init` allowance and `num_labels` propagation are different configuration semantics issues." } ] }, @@ -5009,16 +4979,14 @@ "issue:39692", "issue:41720", "issue:42886", - "issue:43295", "issue:43381", "issue:43421", "issue:43531", + "issue:43644", "issue:43645", "issue:43653", - "issue:43720", "issue:43746", "issue:43782", - "issue:43819", "issue:43844", "issue:43873", "issue:43927", @@ -5030,6 +4998,7 @@ "issue:44514", "issue:44545", "issue:44610", + "issue:44625", "issue:44671", "issue:44843", "issue:44898", @@ -5042,18 +5011,21 @@ "issue:45356", "issue:45357", "issue:45381", + "issue:45479", "issue:45704", "issue:45759", - "issue:45834", "issue:45907", "issue:45923", "issue:46032", "issue:46050", "issue:46463", "issue:46655", + "issue:46665", "issue:46693" ], "soft_pairs": [ + "issue:44625|issue:45479", + "issue:43644|issue:46665", "issue:44062|issue:44451", "issue:44898|issue:45072", "issue:44987|issue:45042", @@ -5061,244 +5033,235 @@ "issue:43381|issue:43844", "issue:39692|issue:44610", "issue:41720|issue:45704", + "issue:43421|issue:43927", "issue:46655|issue:46693", "issue:45704|issue:45923", - "issue:43421|issue:43927", "issue:43381|issue:45704", "issue:43531|issue:43645", - "issue:44918|issue:44936", "issue:43746|issue:46050", + "issue:44918|issue:44936", "issue:42886|issue:44843", "issue:45759|issue:45923", "issue:44514|issue:44545", - "issue:39692|issue:43873", "issue:46032|issue:46463", - "issue:43994|issue:44671", + "issue:39692|issue:43873", "issue:43746|issue:43873", + "issue:43994|issue:44671", "issue:43421|issue:43653", "issue:30064|issue:36010", - "issue:43782|issue:45381", "issue:43421|issue:44186", + "issue:43782|issue:45381", "issue:45081|issue:45356", - "issue:45356|issue:45907", - "issue:43720|issue:43819", - "issue:43653|issue:43720", - "issue:43295|issue:45834" + "issue:45356|issue:45907" ], - "prepared_review_unit_hash": "c7d9bfdedc0e9776dade821f3013e8492cbfed3b7fbb034badd058754778b4d9", + "prepared_review_unit_hash": "104aa2775cad7eae59b90ace02052b55298a590ff35f5846e5ab2d32a1544215", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 37066, - "estimated_input_tokens": 9267, - "estimated_eval_tokens": 18790 + "soft_pair_count": 29, + "serialized_chars": 36804, + "estimated_input_tokens": 9201, + "estimated_eval_tokens": 18658 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of tokenizer/processor/model regression issues rather than one duplicate set. The only clear duplicate pair is the two Qwen2_5_VLProcessor.apply_chat_template batched padding=False reports.", + "summary": "Most items are unrelated and should not be clustered together. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False (44514/44545).", "confidence": 0.94, - "canonical_issue_reason": "issue:44514 is the best canonical issue because it is the older of the two clearly duplicate Qwen2_5_VLProcessor.apply_chat_template reports and has the broader, cleaner title for the same crash.", + "canonical_issue_reason": "issue:44514 is the best canonical issue because it is the older, clearer report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash; 44545 is a near-verbatim duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the strongest issue to anchor the cluster since it is the only unmistakable duplicate target; the rest of the items are related only at a broad subsystem level and should not be merged as one bug.", + "best_issue_reason": "issue:44514 is the strongest representative issue in the set since it matches the one true duplicate pair and has the cleaner original description.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both are classification-label bugs, but they concern different model families and different failure modes." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Unrelated bugs: non-persistent buffer initialization vs from_pretrained behavior inside Python classes." + }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Both involve tokenizers, but one is an AddedToken TypeError and the other is a model loading failure; different underlying bugs." + "reason": "Tokenizer AddedToken argument handling and loading a specific model are different issues." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Unrelated failures: Perceiver image resizing vs dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different subsystems and symptoms: Perceiver image resolution handling vs bfloat16 dtype mismatches." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different code paths and symptoms; one is a model-loading regression, the other is PIL image processors wrongly requiring torchvision." + "reason": "Model loading regression and PIL/torchvision dependency behavior are not the same bug." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Both mention warnings/config, but one is about tie_word_embeddings and the other about incorrect visual encoder keys in save_pretrained." + "reason": "Tie_word_embeddings warning is unrelated to incorrect visual encoder keys on save_pretrained." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Both are training/gradient-related, but eval-mode gradient checkpointing and ZeRO-3 gradient inflation are distinct bugs." + "reason": "Gradient checkpointing eval-mode restriction is distinct from ZeRO-3 gradient inflation." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both are processor/model mismatches, but for different models and different failure modes; not the same bug." + "reason": "Docs/example mismatch and quantization failures are different from OmDet-Turbo processor image-size mismatch." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Completely different issues: CUDA assert during Qwen3 auto device mapping vs a T5 RMSNorm memory leak." + "reason": "Auto device mapping CUDA assert and T5 Apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both involve tokenizers, but one is runtime post-processor updates and the other is save/load of custom token IDs." }, { "left": "issue:46655", "right": "issue:46693", "accept": false, - "reason": "Video-processor padding bug vs FA2 long-context performance regression; unrelated root causes." + "reason": "Video padding calculation and long-context FA2 performance regression are different problems." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "A T5 memory leak and a missing supports_gradient_checkpointing flag are unrelated." - }, - { - "left": "issue:43421", - "right": "issue:43927", - "accept": false, - "reason": "Both touch token/special-token handling, but one is runtime post-processor updates and the other is save/load config corruption." + "reason": "Memory-leak behavior in T5 is unrelated to a missing supports_gradient_checkpointing flag." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "Different problems in different subsystems; eval-mode checkpointing is not the same as a T5 RMSNorm leak." + "reason": "These are different training/runtime bugs with no shared concrete code path." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and Jupyter custom-model initialization are unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to Jupyter custom model initialization." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:43746", + "right": "issue:46050", "accept": false, - "reason": "Both involve trainer workflows, but one is SFT embedding unpacking and the other is evaluate-after-train failure; different bugs." + "reason": "Local checkpoint loading with PEFT adapters is unrelated to sinq quantization save/load breakage." }, { - "left": "issue:43746", - "right": "issue:46050", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "Both are save/load issues, but PEFT adapter loading and sinq quantized persistence are separate code paths." + "reason": "TRL SFT embedding unpacking and trainer.evaluate after trainer.train are different trainer issues." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both break HF_HUB_OFFLINE tokenization flows, but the reported causes differ: cache loading vs an unconditional model_info call in _patch_mistral_regex." + "reason": "Both mention HF_HUB_OFFLINE, but one is general tokenizer cache loading and the other is a specific Mistral regex patch calling model_info(); not clearly the same bug." }, { "left": "issue:45759", "right": "issue:45923", "accept": false, - "reason": "A composite-config unwrap bug and a missing gradient-checkpointing flag are not the same issue." + "reason": "Composite Qwen config handling and a missing capability flag are unrelated." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "These are the same Qwen2_5_VLProcessor.apply_chat_template batched padding=False crash, just phrased differently." - }, - { - "left": "issue:39692", - "right": "issue:43873", - "accept": false, - "reason": "SigLIP2 doc/example problems are unrelated to quantization offloading behavior." + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash; 44545 is effectively a duplicate of 44514." }, { "left": "issue:46032", "right": "issue:46463", "accept": false, - "reason": "Mamba2Mixer use_cache correctness and T5 decoder validation error messaging are unrelated." + "reason": "Mamba2Mixer cache correctness and T5 decoder input validation messaging are unrelated." }, { - "left": "issue:43994", - "right": "issue:44671", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "Both are incorrect model outputs, but for different models and unrelated causes." + "reason": "SigLIP2 documentation errors and offloading with quantization are different issues." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Adapter loading vs quantization offloading are separate loading/runtime issues." + "reason": "PEFT local checkpoint loading and quantization/offloading behavior are not the same underlying bug." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different model families and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked LM prediction regression." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Special-token post-processor updates and BigBirdTokenizer special-token registration are related in theme only, not the same bug." + "reason": "Both touch special tokens, but one is runtime post-processor syncing and the other is BigBirdTokenizer decode output registration." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Image segmentation processing and a GenerationMixin import error are entirely unrelated." + "reason": "Image processor void segmentation maps and a GenerationMixin import error are unrelated." }, { - "left": "issue:43782", - "right": "issue:45381", + "left": "issue:43421", + "right": "issue:44186", "accept": false, - "reason": "Both involve Qwen VL variants, but weight_only loading and video vision_position_ids are different failures." + "reason": "Tokenizer post-processor updates and LayoutLMv2Tokenizer NER/padding crashes are different code paths." }, { - "left": "issue:43421", - "right": "issue:44186", + "left": "issue:43782", + "right": "issue:45381", "accept": false, - "reason": "Both are tokenizer-related, but one is runtime special-token updates and the other is NER padding/truncation crashes." + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL video vision_position_ids are separate multimodal bugs." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both mention fix_mistral_regex/tokenizer regression, but the reported root causes and affected models differ." + "reason": "Both involve tokenizer regressions, but one is an AttributeError in fix_mistral_regex and the other is a Kimi-K2.5 codec/warning regression." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regression in Kimi-K2.5 and list_repo_files revision forwarding are unrelated." - }, - { - "left": "issue:43720", - "right": "issue:43819", - "accept": false, - "reason": "BitNet accelerate-loading unpacking and DAC from_latents/STE mismatch are different model bugs." - }, - { - "left": "issue:43653", - "right": "issue:43720", - "accept": false, - "reason": "BigBird special-token registration and BitNet packed-weight loading are unrelated." - }, - { - "left": "issue:43295", - "right": "issue:45834", - "accept": false, - "reason": "Both involve processor/model regressions, but custom processor.tokenizer access and Kosmos2.5 long OCR indexing are different bugs." + "reason": "Tokenizer regression and list_repo_files revision forwarding are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet: the only clear duplicate pair is issue:44514/issue:44545, and the remaining soft-edge rejections are conservative and consistent with the titles. I don\u2019t see evidence of overstatement or an unsupported merge suggestion." + "feedback": "Grounded overall. The summary matches the packet\u2019s lone deterministic duplicate pair (44514/44545), and the other soft-edge rejections stay conservative by distinguishing different titles, model families, or failure modes without overclaiming. The canonical choice of 44514 is also supported by the earlier timestamp and clearer wording." }, "evaluator_used": true, "retried": false, @@ -5325,7 +5288,7 @@ "issue:36331", "issue:39692", "issue:41720", - "issue:43322", + "issue:43295", "issue:43335", "issue:43531", "issue:43540", @@ -5337,6 +5300,7 @@ "issue:43644", "issue:43653", "issue:43697", + "issue:43720", "issue:43749", "issue:43756", "issue:43819", @@ -5362,24 +5326,26 @@ "issue:45084", "issue:45381", "issue:45584", - "issue:45685", "issue:45834", "issue:46294", "issue:46665", "issue:46693" ], "soft_pairs": [ - "issue:43550|issue:45072", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:43295|issue:45834", "issue:43596|issue:44220", - "issue:44062|issue:44488", + "issue:43550|issue:45072", "issue:43550|issue:43927", + "issue:44062|issue:44488", "issue:43540|issue:46294", "issue:45381|issue:46693", "issue:43618|issue:43697", "issue:43531|issue:43643", "issue:43611|issue:46665", - "issue:43653|issue:43819", "issue:44220|issue:44661", + "issue:43653|issue:43819", "issue:44186|issue:44898", "issue:30064|issue:36331", "issue:43643|issue:44877", @@ -5391,209 +5357,220 @@ "issue:43540|issue:44008", "issue:39692|issue:45834", "issue:44186|issue:44442", - "issue:44448|issue:44671", "issue:44485|issue:44509", + "issue:44448|issue:44671", "issue:43644|issue:43749", "issue:45584|issue:45834", - "issue:43322|issue:45685", "issue:34634|issue:36010" ], - "prepared_review_unit_hash": "5d67dc6bbf6757d4aa5156a9d613a86280077aef5d20b8ccca5abfeb1c58e0c0", + "prepared_review_unit_hash": "d4254d125a320477cf7f806e4ce2dce4633906b055930bea508130186bc2df80", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36077, - "estimated_input_tokens": 9020, - "estimated_eval_tokens": 18296 + "soft_pair_count": 30, + "serialized_chars": 36749, + "estimated_input_tokens": 9188, + "estimated_eval_tokens": 18632 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items cover unrelated bugs across tokenizers, vision/audio models, configs, runtime regressions, and docs. None of the soft-edge pairs look like the same underlying defect, so no merges are justified.", - "confidence": 0.97, - "canonical_issue_reason": "No true canonical issue: the cluster mixes unrelated bug reports with different models, code paths, and failure modes.", + "summary": "The set is highly heterogeneous, with most soft pairs describing distinct model-, tokenizer-, or loading-specific bugs rather than true duplicates. The closest common thread is Transformers v5 regression/loading behavior, so 43611 is the best broad anchor, but no pair is strong enough to merge.", + "confidence": 0.42, + "canonical_issue_reason": "issue:43611 is the broadest, most representative v5 loading-regression report in the set; it best matches the small subset of config/load-related issues, though the cluster overall is mixed.", "canonical_pr_reason": null, - "best_issue_reason": "If a representative issue must be chosen, 43611 is the cleanest standalone regression report and the broadest of the set, but it is not a duplicate of the others.", + "best_issue_reason": "issue:43611 is the best triage anchor because it is user-facing and broadly about model loading regressions, making it more representative than the narrower model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43550", - "right": "issue:45072", + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different bugs: packed-weight loading in BitNet vs DAC latent decoding/STE mismatch." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug vs accelerate loading of packed weights; unrelated code paths." + }, + { + "left": "issue:43295", + "right": "issue:45834", "accept": false, - "reason": "Both are inference/runtime regressions, but they affect different models and different failure modes (SDPA compile vs bfloat16 dtype mismatch)." + "reason": "Both are multimodal, but one is a processor/tokenizer regression and the other is a long OCR input index error in a different model." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Different subsystems and symptoms: ZeRO-3/BertModel indexing versus fbank feature extraction." + "reason": "Deepspeed ZeRO3/BertModel init failure is unrelated to audio fbank extraction." }, { - "left": "issue:44062", - "right": "issue:44488", + "left": "issue:43550", + "right": "issue:45072", "accept": false, - "reason": "Tokenizer/additional-token error versus a model loading failure for a specific checkpoint; not the same bug." + "reason": "Different failure modes: torch.compile/SDPA in Bamba vs bfloat16 dtype mismatch in different models." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Unrelated failures: torch.compile/SDPA on Bamba versus DiaConfig save/load token ID loss." + "reason": "Bamba compile/SDPA bug is unrelated to DiaConfig token ID persistence and generation index errors." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken keyword duplication is unrelated to tokenizer loading for a specific model." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both involve video, but one is a Qwen3OmniMoe processing error and the other is a chat-template feature/token mismatch in a different model." + "reason": "Both concern video input, but they describe different Qwen model families and different mismatches." }, { "left": "issue:45381", "right": "issue:46693", "accept": false, - "reason": "Different problem classes: wrong video position IDs versus a long-context FA2 performance regression." + "reason": "Qwen2.5-VL vision_position_ids bug is unrelated to a performance regression after 5.2.0." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Both are v5 output/behavior changes, but they concern different models and distinct code paths." + "reason": "Both are v5 regressions, but one is missing CLIP attentions and the other is RT-DETR output drift; not the same bug." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window Qwen3-MoE bug is unrelated to AutoConfig trust_remote_code field loss." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to AutoConfig missing fields with trust_remote_code." }, { "left": "issue:43611", "right": "issue:46665", "accept": false, - "reason": "Both touch loading, but one is a base_model_prefix regression and the other is inconsistent from_pretrained behavior inside Python classes." + "reason": "Both mention from_pretrained/loading, but one is a base_model_prefix regression and the other is inconsistent class-wrapper behavior." }, { - "left": "issue:43653", - "right": "issue:43819", + "left": "issue:44220", + "right": "issue:44661", "accept": false, - "reason": "Tokenizer special-token registration bug is unrelated to DAC.from_latents/STE mismatch." + "reason": "Audio feature extraction bug vs tokenizer mapping failure when adding a new model-like entry." }, { - "left": "issue:44220", - "right": "issue:44661", + "left": "issue:43653", + "right": "issue:43819", "accept": false, - "reason": "Audio feature extraction failure and add-new-model-like mapping failure are different bugs." + "reason": "BigBirdTokenizer special-token registration and DAC latent decoding are unrelated." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer padding/NER crash is not the same as Perceiver non-default-resolution image classification failure." + "reason": "Tokenizer crash on NER padding/truncation is unrelated to Perceiver image classification with non-default resolution." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor void segmentation-map failure and CustomTrainer compute_loss signature mismatch are unrelated." + "reason": "Image processor void-segmentation failure is unrelated to CustomTrainer.compute_loss signature changes." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both are config-loading issues, but one drops fields from trust_remote_code and the other rejects granite_speech under strict config." + "reason": "Missing fields from trust_remote_code configs is different from strict config blocking granite_speech loading." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Different model-specific bugs: sparse layer creation in SwitchTransformers versus RoPE layer dropping in Smollm3." + "reason": "SwitchTransformers sparse-layer construction bug and Smollm3 RoPE-layer dropping are different model/config issues." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Qwen3 CUDA assert on A800 is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Different models and code paths; torch.compile/SDPA is not the same as DAC.from_latents forward mismatch." + "reason": "Bamba torch.compile/SDPA failure is unrelated to DAC forward/from_latents mismatch." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Same helper function, but two distinct bugs: an unconditional hub call offline and an AttributeError on a different tokenizer object." + "reason": "Same helper function, but different bugs: offline model_info call vs backend_tokenizer attribute error." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template compilation errors are unrelated." + "reason": "AutoImageProcessor URL loading failure is unrelated to template compilation TypeError." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Video-input processing issue and Gemma3n variable-name collision are different multimodal bugs." + "reason": "Different multimodal code paths: Qwen3OmniMoe video processing vs Gemma3n audio_mel_mask variable collision." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Documentation-example errors and Kosmos2.5 long-ocr index errors are not the same underlying issue." + "reason": "SigLIP2 doc example errors are unrelated to Kosmos2.5 long OCR index errors." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer crash and FastSpeech2ConformerTokenizer loading failure are unrelated tokenizer bugs." + "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to AutoTokenizer failing on FastSpeech2ConformerTokenizer." }, { - "left": "issue:44448", - "right": "issue:44671", + "left": "issue:44485", + "right": "issue:44509", "accept": false, - "reason": "Both are v5 regressions in model outputs, but they affect different models and distinct implementations." + "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipeline tasks." }, { - "left": "issue:44485", - "right": "issue:44509", + "left": "issue:44448", + "right": "issue:44671", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to docs mentioning removed pipeline tasks." + "reason": "Different model regressions: Pegasus output drift vs CamemBERT masked LM prediction errors." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer initialization bug and FSDP CPU RAM efficient loading failure are different regressions." + "reason": "Non-persistent buffer junk in v5 is not the same as FSDP CPU RAM efficient loading being broken." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation failure and Kosmos2.5 long OCR index error are unrelated." - }, - { - "left": "issue:43322", - "right": "issue:45685", - "accept": false, - "reason": "Llava Next load segfault and MPS histogram support error are different code paths and symptoms." + "reason": "Whisper empty-transcription failure after token alignment is unrelated to Kosmos2.5 OCR length handling." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset issue and GenerationMixin import error are unrelated." + "reason": "BarkProcessor voice preset issue is unrelated to the GenerationMixin import error." } ] }, @@ -5619,9 +5596,8 @@ "nodes": [ "issue:28218", "issue:29159", - "issue:34634", - "issue:36331", "issue:42175", + "issue:43322", "issue:43388", "issue:43421", "issue:43540", @@ -5655,6 +5631,7 @@ "issue:45183", "issue:45381", "issue:45479", + "issue:45685", "issue:45701", "issue:45834", "issue:45987", @@ -5667,8 +5644,9 @@ "issue:46665" ], "soft_pairs": [ - "issue:45183|issue:45479", + "issue:43322|issue:45685", "issue:43606|issue:45072", + "issue:45183|issue:45479", "issue:43716|issue:44008", "issue:43388|issue:46032", "issue:44479|issue:45834", @@ -5677,224 +5655,220 @@ "issue:44991|issue:46291", "issue:43582|issue:46512", "issue:43577|issue:43818", - "issue:43782|issue:44977", "issue:46294|issue:46463", + "issue:43782|issue:44977", "issue:44561|issue:45020", "issue:43421|issue:44811", "issue:44805|issue:44936", "issue:43540|issue:45381", "issue:44610|issue:45072", - "issue:43720|issue:45072", "issue:43825|issue:43827", + "issue:43720|issue:45072", "issue:28218|issue:29159", "issue:43749|issue:44291", "issue:43645|issue:46665", "issue:45987|issue:46192", "issue:44671|issue:45701", "issue:43577|issue:43582", - "issue:42175|issue:45081", - "issue:34634|issue:36331" + "issue:42175|issue:45081" ], - "prepared_review_unit_hash": "30ee9f823a104605b873dcbcba891dcc94a67c47a58364313e2c5e1a66d9289b", + "prepared_review_unit_hash": "81e26f8591ecab93bddb46e33f003f90182a5341eee6776de92f91a138e20ebd", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36336, - "estimated_input_tokens": 9084, - "estimated_eval_tokens": 18424 + "serialized_chars": 36320, + "estimated_input_tokens": 9080, + "estimated_eval_tokens": 18416 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one soft duplicate pair looks like the same underlying bug: the tokenizer added-token spacing regression (#28218 / #29159). The rest are distinct issues across different models, loading paths, or error conditions and should stay separate.", - "confidence": 0.86, - "canonical_issue_reason": "#28218 is the clearest, most concrete statement of the tokenizer spacing bug and the best representative of the only credible duplicate pair in this set.", + "summary": "The set is mostly a thematically related but heterogeneous mix of Transformers issues: tokenizer regressions, model loading/dtype problems, multimodal/video bugs, and a few v5 breakages. I do not see any soft pair that clearly describes the same concrete bug or change, so I would keep them separate.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set for version-related remote_code breakages, and is the closest thing here to a cluster anchor; however, the overall set is still too mixed for a strong duplicate canonical.", "canonical_pr_reason": null, - "best_issue_reason": "#28218 is the best issue to keep as the representative tokenizer bug; it is specific, concrete, and matches the accepted duplicate closely.", + "best_issue_reason": "issue:45020 is the best representative of the main recurring theme (recent-version regressions affecting model loading/execution), even though many other items are only loosely related and not duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45183", - "right": "issue:45479", + "left": "issue:43322", + "right": "issue:45685", "accept": false, - "reason": "Validation/XOR input checking across models is a different bug from the single-label classification zero-loss regression." + "reason": "Different failures in different subsystems: Llava Next loading segfault vs MPS histogram on Int in moe." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "CPU offload device mismatch and bf16 dtype mismatch are different failure modes in different code paths." + "reason": "Both are runtime/backend mismatches, but they involve different models and distinct code paths (CPU offload vs bfloat16 dtype handling)." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "One is about input-validation error messaging; the other is a training loss bug for single-label classification." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Image preprocessor/model dtype mismatch is unrelated to Gemma3n's audio_mel_mask variable collision." + "reason": "Distinct multimodal bugs: image preprocessor/model dtype mismatch vs an AttributeError from a variable-name collision." }, { "left": "issue:43388", "right": "issue:46032", "accept": false, - "reason": "Dropping label elements in gather_for_metrics is unrelated to Mamba2 use_cache corruption for seq_len > 1." + "reason": "Different core problems: metric gathering drops labels in the last batch vs Mamba2 cache generation producing wrong results." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "These hit different models and symptoms: a Qwen VL video regression vs a Kosmos2.5 long-ocr index error." + "reason": "Both are vision/multimodal issues, but the concrete failures are unrelated (Qwen video-input regression vs Kosmos2.5 OCR index error)." }, { "left": "issue:43950", "right": "issue:45987", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained is not the same as StaticCache returning the wrong type." + "reason": "From-pretrained non-persistent buffer corruption is a loading regression; StaticCache.get_seq_length is a return-type/contract bug." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline deprecation/error messaging is distinct from SigLIP2 producing wrong outputs under pipeline/AutoModel." + "reason": "One is an incorrect pipeline error message; the other is a model producing nonsensical outputs via AutoModel/pipeline." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "Tokenizer loading failure and pipelines import failing on a dependency requirement are unrelated issues." + "reason": "Tokenizer loading failure for one model vs importing transformers.pipelines failing due to kernels version requirements." }, { "left": "issue:43582", "right": "issue:46512", "accept": false, - "reason": "Apple Silicon allocator TypeError is unrelated to stale config.dtype after from_config(dtype=...)." + "reason": "Both mention dtype-ish behavior, but one is an Apple Silicon warmup TypeError and the other is stale config dtype after from_config()." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype loading issue is different from Video-LLaVA missing temporal attention and weight-sharing behavior." + "reason": "BLIP2 dtype persistence issue vs Video-LLaVA architecture/weight-sharing problem; not the same bug." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:46294", + "right": "issue:46463", "accept": false, - "reason": "weight_only loading errors are a different path from flash-attention generation failures." + "reason": "Qwen3.5-9B video token/feature mismatch is unrelated to T5 decoder-input validation errors." }, { - "left": "issue:46294", - "right": "issue:46463", + "left": "issue:43782", + "right": "issue:44977", "accept": false, - "reason": "Qwen3.5 video chat-template token mismatch and T5 decoder-input validation are different model/input-validation bugs." + "reason": "Both are Qwen-related, but one is a weight_only loading error and the other is a flash-attention generation bug; different code paths." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Removal of one helper function is not the same as the broader remote_code breakage report." + "reason": "44561 is one specific remote_code breakage caused by removing is_torch_fx_available; 45020 is a broader report of recent-version remote_code failures, not the same concrete bug." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Special-token post-processor updates and batch_decode skip_special_tokens handling are different tokenizer features." + "reason": "Special-token post-processor updates and Whisper batch_decode skip_special_tokens are separate tokenizer/decoder behaviors." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask/index shape mismatch is not the same as trainer.evaluate failing after trainer.train()." + "reason": "A tensor shape IndexError is not the same issue as trainer.evaluate() failing after trainer.train()." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "These are different video-input regressions in different Qwen-family models and likely different code paths." + "reason": "Both involve video inputs, but the reported mismatches differ (processing ValueError vs wrong vision_position_ids)." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor image-size mismatch is unrelated to bfloat16 dtype mismatch in inference." + "reason": "Processor image-size mismatch for OmDet-Turbo vs dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." }, { - "left": "issue:43720", - "right": "issue:45072", + "left": "issue:43825", + "right": "issue:43827", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is a different bug than a dtype mismatch." + "reason": "Runtime error-message regression vs stale docs after pipeline removals; related topic, but not the same underlying bug/change." }, { - "left": "issue:43825", - "right": "issue:43827", + "left": "issue:43720", + "right": "issue:45072", "accept": false, - "reason": "Both mention pipeline removals, but one is an error-message issue and the other is stale documentation." + "reason": "Packed-weight loading bug in BitNet is unrelated to generic bfloat16 dtype mismatch issues." }, { "left": "issue:28218", "right": "issue:29159", - "accept": true, - "reason": "These describe the same underlying tokenizer spacing regression around added tokens; one is the specific symptom, the other the slow/fast inconsistency framing." + "accept": false, + "reason": "The first is specifically about an added-token spacing bug; the second is a broader slow-vs-fast tokenizer inconsistency report." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM efficient loading and unexpected _is_hf_initialized argument handling are separate loading bugs." + "reason": "FSDP CPU RAM-efficient loading breakage and init_empty_weights rejecting _is_hf_initialized are different loading-path regressions." }, { "left": "issue:43645", "right": "issue:46665", "accept": false, - "reason": "Custom-model initialization in notebooks and from_pretrained inside Python classes are related only at a very broad level, not the same bug." + "reason": "Jupyter custom-model initialization in v5 vs from_pretrained inside Python classes are different usage patterns and failures." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "Cache length return-type contract and return_tensors being ignored are unrelated API bugs." + "reason": "StaticCache length contract bug and return_tensors being ignored when text_kwargs is passed are unrelated API issues." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT MLM prediction regression is not the same as a general tokenization-change complaint." + "reason": "CamemBERT masked-LM regression and generic tokenization changes in new versions are too broad to be the same bug." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "BLIP2 dtype state and Apple Silicon warmup TypeError are different problems." + "reason": "Blip2 dtype mismatch and AppleSilicon caching_allocator_warmup TypeError are unrelated." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "A missing TensorFlow backend in installation extras is unrelated to Mistral tokenizer backend_tokenizer access failure." - }, - { - "left": "issue:34634", - "right": "issue:36331", - "accept": false, - "reason": "BarkProcessor voice_preset failure and CustomTrainer.compute_loss signature breakage are entirely different issues." + "reason": "Packaging/backend dependency issue vs tokenizer regex patch crash; no shared code-path bug." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The conclusion is well grounded in the packet: #28218 and #29159 are plausibly the same tokenizer spacing/added-token regression, and the remaining pairs are conservatively rejected as different issues. No overstatement or mergeability problem is evident." - }, - "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 }, @@ -5911,9 +5885,9 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ + "issue:34634", "issue:36010", "issue:36331", - "issue:39692", "issue:42175", "issue:42490", "issue:43257", @@ -5961,6 +5935,7 @@ "issue:46728" ], "soft_pairs": [ + "issue:34634|issue:36331", "issue:45003|issue:46285", "issue:46697|issue:46728", "issue:43653|issue:44912", @@ -5970,222 +5945,221 @@ "issue:43742|issue:45020", "issue:44589|issue:45042", "issue:43782|issue:45325", - "issue:44060|issue:44368", "issue:45372|issue:46326", - "issue:43329|issue:45600", - "issue:43550|issue:44912", + "issue:44060|issue:44368", "issue:45084|issue:45538", + "issue:43550|issue:44912", + "issue:43329|issue:45600", "issue:43526|issue:43697", "issue:43452|issue:44991", "issue:43526|issue:44554", "issue:43257|issue:43329", "issue:36010|issue:36331", + "issue:43906|issue:46209", "issue:36010|issue:43645", "issue:43819|issue:43927", - "issue:43906|issue:46209", - "issue:43638|issue:44936", "issue:43531|issue:45042", + "issue:43638|issue:44936", "issue:42175|issue:46291", "issue:43720|issue:44610", - "issue:39692|issue:44912", "issue:42490|issue:43257" ], - "prepared_review_unit_hash": "c48078d9570b9e82c367d4fe3bab1a3e206dd9d8f4e453bbf6c6f3b22bd34134", + "prepared_review_unit_hash": "f1b8bdf905d98ec4de92b88ae407f0c6e2c4763e614a8991b51000f36d0aa2ba", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36586, - "estimated_input_tokens": 9147, - "estimated_eval_tokens": 18550 + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is not a true duplicate cluster: the items span unrelated bugs in loading, tokenizers, processors, training, quantization, and dependency handling. I would not merge any of the soft-edge pairs; they are only superficially similar.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue cleanly fits the whole set; the cluster is too heterogeneous. If forced, issue:45020 is the broadest loading/compatibility regression report, but it does not unify the rest.", + "summary": "This cluster is highly heterogeneous: the soft pairs mostly share only broad subsystem or keyword similarity, not the same concrete bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated bugs in tokenizers, multimodal processors, training, quantization, and backend/package compatibility.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the broadest and most central report among these, but only as a loose representative of a loading-regression subset; it is not a true umbrella for the cluster.", + "best_issue_reason": "No issue is clearly better than the rest as a cluster representative because the set is not a true duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss are unrelated failures in different code paths." + }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Different bugs: unsafe sys.modules access vs a CLIP text-model structure regression." + "reason": "sys.modules access in modeling_utils and CLIP text-model structure inconsistency are different regressions." }, { "left": "issue:46697", "right": "issue:46728", "accept": false, - "reason": "Different vision preprocessing bugs in different models: over-padding vs wrong patch counting." + "reason": "Both are image-processor bugs, but one is Swin2SR over-padding and the other is Aria patch counting; different logic." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Unrelated: BigBirdTokenizer special-token registration vs MXFP4 quantization fallback." + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are unrelated." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Both audio-related, but different code paths: chat-template sampling-rate default vs DAC latent reconstruction mismatch." + "reason": "Audio chat-template sampling-rate handling and DAC latent reconstruction are separate bugs." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Different runtime failures: Apple Silicon allocator warmup TypeError vs CUDA AMP loss crash." + "reason": "Apple Silicon caching allocator TypeError and CUDA float16 AMP crash are different backend-specific issues." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Different tokenizer/processor failures: Mistral regex patching vs Gemma4Processor serve initialization." + "reason": "Both involve processor/tokenizer loading, but one is Mistral regex patching and the other is Gemma4 missing _tokenizer; not the same defect." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "A specific model loading key error is not the same bug as a broad remote_code compatibility regression." + "reason": "A key error loading one model is not the same as the broad remote_code breakage described in the other issue." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Different subsystems: float8 storage lookup vs PIL image processor dependency handling." + "reason": "Float8 storage lookup failure and PIL/torchvision dependency regression are unrelated." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Both Qwen-VL, but one is a from_pretrained weight_only load error and the other is a RoPE index calculation bug." + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL rope index scaling are different model-specific problems." }, { - "left": "issue:44060", - "right": "issue:44368", + "left": "issue:45372", + "right": "issue:46326", "accept": false, - "reason": "Similar warning theme, but different model families and contexts; not enough to establish the same underlying bug." + "reason": "Both mention mistral_common, but one is an import error and the other is continue_final_message behavior; different root causes." }, { - "left": "issue:45372", - "right": "issue:46326", + "left": "issue:44060", + "right": "issue:44368", "accept": false, - "reason": "Both mention mistral_common, but one is an import error and the other is continue_final_message behavior." + "reason": "They both report tied-weights warnings, but the models and warning cause differ; not the same underlying bug." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:45084", + "right": "issue:45538", "accept": false, - "reason": "Different areas: multimodal token counting video branch vs auto_mappings referencing removed configs." + "reason": "Template compilation error and CLIPTokenizer max_length issue are unrelated." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "Different failures: torch.compile/SDPA on Bamba vs MXFP4 quantization loading on git-oss-20b." + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to MXFP4 loading fallback in git-oss-20b." }, { - "left": "issue:45084", - "right": "issue:45538", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "Different bugs: template-node compilation error vs CLIPTokenizer max-length default." + "reason": "Undefined video token helper variables and stale auto_mappings config references are different code paths." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Both vision-related, but one is a BEiT label reduction bug and the other is an RTDetrV2 output regression." + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output drift are not the same bug." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both loading-related, but one is gguf_file argument handling and the other is a tokenizer-loading regression for a specific model." + "reason": "gguf_file handling and tokenizer loading for EMBEDDIA/est-roberta are separate from-pretrained failures." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Unrelated: BEiT processor label reduction vs an MPS attention correctness issue." + "reason": "A label-processing bug in BeitImageProcessorFast is unrelated to the MPS attention correctness issue." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Different code paths: Qwen3 MoE weight conversion/loading vs multimodal token counting." + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed and multimodal token counting are unrelated." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Different incompatibilities: GenerationMixin import path vs CustomTrainer.compute_loss signature change." + "reason": "GenerationMixin import failure and CustomTrainer compute_loss signature mismatch are different API breaks." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of an external issue and FP8 inference failure on B200 are not the same defect." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "Both involve custom/model-user code, but the actual bugs differ: import path breakage vs Jupyter custom model initialization." + "reason": "GenerationMixin import error and Jupyter custom-model initialization breakage are different compatibility issues." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "Different domains: DAC latent inference mismatch vs DiaConfig save/load losing custom token IDs." + "reason": "DAC latent/forward mismatch and DiaConfig token-ID persistence loss are unrelated model/config bugs." }, { - "left": "issue:43906", - "right": "issue:46209", + "left": "issue:43531", + "right": "issue:45042", "accept": false, - "reason": "No shared bug: an isolated reproduction of a prior issue is not the same as an FP8 inference failure on B200." + "reason": "Qwen3-MoE sliding window logic and PIL backend torchvision dependency regression are unrelated." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Both training-related, but one is a ZeRO-3 index error and the other is evaluate() failing after train(); different code paths." - }, - { - "left": "issue:43531", - "right": "issue:45042", - "accept": false, - "reason": "Unrelated: Qwen3-MoE sliding-window behavior vs PIL backend torchvision dependency regression." + "reason": "Zero3 indexing failure with a non-pretrained Bert model and evaluate-after-train behavior are distinct trainer issues." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Different packaging regressions: torch-only install missing TensorFlow vs kernels v0.15 import requirements." + "reason": "Torch-only install missing TensorFlow backend and kernels v0.15 import failure are different packaging/import problems." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Different bugs: packed-weight unpacking during accelerate loading vs processor output size mismatch." - }, - { - "left": "issue:39692", - "right": "issue:44912", - "accept": false, - "reason": "Documentation example errors are not the same as a quantization-loading regression." + "reason": "Packed-weight unpacking during accelerate loading and OmDet-Turbo input-size mismatch are unrelated." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Different issues: trailing-slash local-model import warning vs Qwen3 MoE weight conversion/loading." + "reason": "A trailing-slash local-path warning and Qwen3 MoE conversion failure do not share the same underlying bug." } ] }, @@ -6212,6 +6186,7 @@ "issue:29159", "issue:34567", "issue:38175", + "issue:39692", "issue:40990", "issue:42915", "issue:43257", @@ -6258,6 +6233,7 @@ "issue:46701" ], "soft_pairs": [ + "issue:39692|issue:44912", "issue:43746|issue:44811", "issue:38175|issue:44610", "issue:43643|issue:45759", @@ -6274,205 +6250,211 @@ "issue:43749|issue:44756", "issue:44291|issue:45042", "issue:42915|issue:44821", - "issue:43606|issue:45859", "issue:34567|issue:40990", + "issue:43606|issue:45859", "issue:43650|issue:43906", "issue:43540|issue:46701", "issue:43927|issue:44912", "issue:29159|issue:34567", - "issue:43454|issue:43577", "issue:43525|issue:46291", - "issue:44568|issue:45356", - "issue:44987|issue:45538" + "issue:43454|issue:43577", + "issue:44987|issue:45538", + "issue:44568|issue:45356" ], - "prepared_review_unit_hash": "c42eb0c674f28fbeb10b58d2fdcdef3b14838816e3fbc27d2a65b8a769c35302", + "prepared_review_unit_hash": "f1a65433d519d521c25ec53332740f6db5151d39b529dc37267f0abad4e657f5", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34990, - "estimated_input_tokens": 8748, - "estimated_eval_tokens": 17752 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35889, + "estimated_input_tokens": 8973, + "estimated_eval_tokens": 18202 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft pairs mostly share only broad subsystem terms (tokenizers, video, loading, cache, quantization) but not the same concrete bug. I would not merge any of the suggested pairs.", + "summary": "The cluster is highly heterogeneous: it mixes tokenizer, model-loading, processor, cache, quantization, and offload bugs that do not describe the same underlying defect. No soft pair looks like a true duplicate/merge candidate.", "confidence": 0.96, - "canonical_issue_reason": "No single canonical duplicate stands out across this mixed set. If forced to pick one representative issue, issue:43749 is the strongest standalone tracker because it has the most discussion and inbound references, but it does not represent the other topics.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue:43749 is the best global representative by activity and external references, making it the most visible standalone bug report in the set.", + "best_issue_reason": "issue:43749 is the most prominent and broadly impactful item here (highest discussion/inbound reference signal), so it is the closest thing to a representative issue, but the set is too diverse for a real canonical duplicate anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Different problems: SigLIP2 doc example errors vs MXFP4 quantization loading/fallback behavior." + }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "Different problems: PEFT/local checkpoint loading vs Whisper batch_decode ignoring skip_special_tokens." + "reason": "PEFT local-checkpoint loading bug vs Whisper batch_decode skip_special_tokens handling; unrelated code paths." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Both involve vision models, but one is zero probabilities and the other is a processor/model input-size mismatch." + "reason": "Unexpected zero probabilities in SigLIP2 vs OmDet-Turbo processor/model size mismatch." }, { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "Different config-loading bugs: missing fields from trust_remote_code vs not unwrapping text_config for composite multimodal configs." + "reason": "Both are config-loading issues, but one is missing fields from trust_remote_code and the other is composite Qwen unwrapping; not the same bug." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Separate model-specific failures: GLM-4.7-Flash load/test issue vs Qwen3.5 generation with flash-attention." + "reason": "GLM-4.7 unit-test load failure vs Qwen3.5 flash-attention generation regression; different failures." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "One is int4 quantization reserved-memory growth; the other is Sam3Video OOM. Same symptom class, different bugs." + "reason": "Both mention OOM, but one is int4 reserved-memory regression and the other is a model-specific video OOM." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MOE weight-conversion failure is not the same as GPT-oss-20b CUDA OOM." + "reason": "Qwen3 MOE weight conversion/loading issue vs a GPT-oss CUDA OOM; not the same underlying bug." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate default is unrelated to Whisper batch_decode special-token handling." + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode option handling; different processor behaviors." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen-related loading problems, but one is weight conversion with accelerate/deepspeed and the other is a shape mismatch for Qwen3-VL." + "reason": "Qwen3 MOE conversion bug vs Qwen3-VL shape mismatch on load; separate model/load paths." }, { "left": "issue:46421", "right": "issue:46424", "accept": false, - "reason": "Both touch cache behavior, but one is spurious recompilation from lazy init while the other is max_cache_length handling for quantized cache." + "reason": "Both cache-related, but one is lazy-init recompilation and the other is max_cache_length scope; related area, different bugs." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Both mention exception handling, but they are different code paths and different failure modes." + "reason": "Both are error-swallowing patterns, but they affect different tensor/conversion paths and are not the same defect." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different load-time failures: FineGrainedFP8Config on Qwen3Moe vs a key error for MobileLLM-125M." + "reason": "FineGrainedFP8Config failure for Qwen3Moe vs key error loading MobileLLM-125M; unrelated model-load issues." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "Both involve video, but 43329 is an undefined-variable bug in token counting while 44479 is a v5.3.0 video-input regression across several Qwen VL models." + "reason": "Undefined video token-counting helpers vs a broad v5.3.0 video input regression across models; not one concrete bug." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "One is a multimodal token-counting bug; the other is incorrect batched video processor output shape." + "reason": "Video token counting code bug vs batched video processor output shape bug; separate failure modes." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Different memory/OOM contexts: FSDP CPU RAM efficient loading vs mmap on Strix Halo." + "reason": "FSDP CPU RAM efficient loading vs disabling mmap on Strix Halo; both memory-related but different fixes." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is unrelated to torchvision being required by PIL image processors." + "reason": "init_empty_weights argument incompatibility vs PIL image processors incorrectly requiring torchvision; unrelated subsystems." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "One is a model/config loading issue; the other is AutoImageProcessor URL loading." + "reason": "Qwen3Moe FP8 loading problem vs AutoImageProcessor URL loading failure; no shared code-path bug." }, { - "left": "issue:43606", - "right": "issue:45859", + "left": "issue:34567", + "right": "issue:40990", "accept": false, - "reason": "Different CPU-offload bugs in different models; no shared concrete code path." + "reason": "TrainerState token counter not updating vs high perplexity on GPT-oss; unrelated training/eval issues." }, { - "left": "issue:34567", - "right": "issue:40990", + "left": "issue:43606", + "right": "issue:45859", "accept": false, - "reason": "TrainerState token-count tracking is unrelated to GPT-oss-20b perplexity on WikiText-2." + "reason": "CPU offload device mismatch for bark-small vs Qwen3_5Moe rotary embedding offload incompatibility; different models and code paths." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "'ADD THE DATA' is not a technical duplicate of an isolated reproduction for issue 38071." + "reason": "'ADD THE DATA' is not a technical bug report, and the other is an isolated reproduction of a different issue." }, { "left": "issue:43540", "right": "issue:46701", "accept": false, - "reason": "Different video-processing failures: Qwen3OmniMoe video ValueError vs DiffusionGemma image-text-to-text pipeline failure." + "reason": "Qwen3OmniMoe video-processing ValueError vs DiffusionGemma image-text pipeline failure; different models and surfaces." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig token-ID persistence bug is unrelated to MXFP4 quantization fallback to bf16." + "reason": "DiaConfig save/load token-ID loss vs MXFP4 quantization fallback; unrelated configuration vs quantization behavior." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState num_input_tokens_seen not updating." - }, - { - "left": "issue:43454", - "right": "issue:43577", - "accept": false, - "reason": "AyaVision lm_head tying bug is unrelated to Blip2 dtype propagation failing to set model/model.qformer to half precision." + "reason": "Tokenizer slow/fast inconsistency vs TrainerState input-token accounting; no shared bug." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Llama4Config missing pad_token_id is unrelated to pipelines import failing due to kernels version/revision requirements." + "reason": "Llama4Config missing pad_token_id vs kernels v0.15 import/version requirement; unrelated failures." }, { - "left": "issue:44568", - "right": "issue:45356", + "left": "issue:43454", + "right": "issue:43577", "accept": false, - "reason": "Different tokenizer regressions: missing BOS/EOS addition vs codec handling and warning behavior in Kimi-K2.5." + "reason": "AyaVision lm_head tying bug vs BLIP2 dtype propagation issue; different model internals." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Loading physical-intelligence/fast with transformers>=5.1.0 is unrelated to CLIPTokenizer's excessive model_max_length." - } - ] - }, - "evaluator_result": null, + "reason": "Physical-intelligence/fast loading regression vs CLIPTokenizer max_length bug; different tokenizer/load issues." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "mdeberta add_special_tokens regression vs Kimi-K2.5 codec/regex regression; separate tokenizer bugs." + } + ] + }, + "evaluator_result": null, "evaluator_used": false, "retried": false, "accepted_nontrivial_soft_edge": false, @@ -6552,8 +6534,8 @@ "issue:43742|issue:44291", "issue:28282|issue:29127", "issue:44479|issue:46693", - "issue:44062|issue:44991", "issue:43317|issue:45538", + "issue:44062|issue:44991", "issue:44849|issue:45701", "issue:42915|issue:43749", "issue:43577|issue:46512", @@ -6562,217 +6544,217 @@ "issue:27957|issue:28282", "issue:24643|issue:29159", "issue:43540|issue:44560", - "issue:44568|issue:45488", "issue:42175|issue:44488", - "issue:44987|issue:45907", "issue:44291|issue:44821", + "issue:44987|issue:45907", "issue:43540|issue:44479", "issue:44479|issue:45381", + "issue:44568|issue:45488", "issue:38175|issue:44898", "issue:34567|issue:36331" ], - "prepared_review_unit_hash": "7eac286cc87f63160e5a3088a8ca9c28f95ac21765785b8749c8a6084574f851", + "prepared_review_unit_hash": "2ca7c5b4e7c785b01b128214d802d8af96eb4d2223dc90db3c5fae9caa10badc", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35841, - "estimated_input_tokens": 8961, - "estimated_eval_tokens": 18178 + "serialized_chars": 35839, + "estimated_input_tokens": 8960, + "estimated_eval_tokens": 18176 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous and does not form a single duplicate set: it mixes tokenizer regressions, multimodal/video bugs, model-loading/import errors, training/runtime failures, and packaging/version issues. None of the soft pairs look like the same underlying bug or change.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 45488 is the best available anchor because it is open, concrete, and in the strongest tokenizer-regression thread, but the cluster is too mixed for a confident true canonical.", + "summary": "This cluster is mostly unrelated transformer issues with only broad subsystem overlap (tokenizers, Qwen, model loading, dtype, training). No soft pair looks like the same concrete bug/change, so I would keep them separate. No PRs are present.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 45488 is the most representative cluster anchor: it is detailed, currently open, and describes a broadly impactful tokenizer regression with inbound references, making it the best umbrella issue in this set.", "canonical_pr_reason": null, - "best_issue_reason": "45488 is the most reusable issue target here; it has an explicit, actionable regression description and active discussion, unlike the other items which are mostly unrelated one-off bugs.", + "best_issue_reason": "Issue 45488 best fits as the cluster\u2019s canonical issue because it is specific enough to be actionable while still broad in impact, and its discussion/inbound references suggest it is the most central/visible report here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Different failure modes and subsystems: generic PyTorch-missing import versus a specific Llama 3.2 Vision loading regression." + "reason": "Both are model-loading failures, but one is a PyTorch-missing import error and the other is a Llama 3.2 Vision loading regression; different causes and code paths." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is Mistral3 lm_head tying and the other is LoRA merge with extended vocabulary; different code paths." + "reason": "Both mention tied embeddings, but one is about Mistral3 lm_head weight tying and the other is LoRA merge collapse with extended vocab; not the same bug." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different models and bugs: RTDetrV2 output mismatch versus Qwen3.5 save_pretrained key mapping." + "reason": "RTDetrV2 output mismatch and Qwen3.5 save_pretrained visual-key regression are separate model-specific failures." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Unrelated dependency problems: torch compatibility versus kernels requiring a revision/version argument." + "reason": "Torch version incompatibility and kernels import requiring revision/version are unrelated issues." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Unrelated models and symptoms: FineGrainedFP8Config failure versus CLIPTokenizer model_max_length behavior." + "reason": "Qwen3Moe FP8 config failure and CLIPTokenizer max_length bug are unrelated." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and PEFT version pinning are unrelated." + "reason": "DeepSpeed ZeRO3 deque error and a PEFT minimum-version bump are not the same underlying problem." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error in model loading is unrelated to TrainerState token-count bookkeeping." + "reason": "MacOS bus error in CLIP loading and TrainerState token counter not updating are unrelated." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging/backend inclusion issue versus missing Qwen class import; different problem areas." + "reason": "Backend packaging omission and missing Qwen2.5-VL class import are different failures." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "KeyError loading MobileLLM is not the same bug as init_empty_weights receiving an unexpected argument." + "reason": "MobileLLM key-loading error and init_empty_weights unexpected argument error are distinct load-path bugs." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "Generic missing-PyTorch import error versus LayoutLMv3 box-validation messaging; not the same bug." + "reason": "PyTorch import failure and LayoutLMv3 box-validation clarity issue are unrelated." }, { "left": "issue:44479", "right": "issue:46693", "accept": false, - "reason": "Different regression domains: multimodal video-input behavior versus FA2 long-context generation performance." + "reason": "Both involve vision/FA2-adjacent transformer regressions, but one is a video-input functional regression and the other is a long-context performance regression; not the same bug." }, { - "left": "issue:44062", - "right": "issue:44991", + "left": "issue:43317", + "right": "issue:45538", "accept": false, - "reason": "Both are tokenizer-related, but one is an AddedToken keyword collision and the other is a model-specific tokenizer load failure; not enough evidence of one root cause." + "reason": "device_map/offload loading failure and CLIPTokenizer max_length bug are unrelated." }, { - "left": "issue:43317", - "right": "issue:45538", + "left": "issue:44062", + "right": "issue:44991", "accept": false, - "reason": "Device-map/offload loading failure is unrelated to CLIPTokenizer max-length handling." + "reason": "Both are tokenizer-related, but one is an AddedToken argument duplication bug and the other is a model-specific tokenizer load failure; insufficient evidence of same root cause." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 hidden-states bug versus a broad tokenizer-version regression; different concrete failures." + "reason": "Qwen3.5 hidden-state output bug and a general tokenization version-change report are not the same concrete issue." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "FP8 configuration failure and FSDP CPU RAM efficient loading are separate training-stack issues." + "reason": "FP8 config failure and FSDP CPU RAM efficient loading breakage are separate training/inference paths." }, { "left": "issue:43577", "right": "issue:46512", "accept": false, - "reason": "Both involve dtype, but one is BLIP2 loading behavior and the other is config dtype staying stale; different bugs." + "reason": "Both concern dtype handling, but one is Blip2/qformer weights staying float32 and the other is config.dtype being stale after from_config; different bugs." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf model/tokenizer loading and AppleSilicon caching_allocator_warmup TypeError are unrelated." + "reason": "gguf_file loading breakage and Apple Silicon caching_allocator_warmup TypeError are unrelated." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "CUDA assert under auto device mapping and Qwen3OmniMoe video-input ValueError are different model-path failures." + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3OmniMoe video-input ValueError are different failure modes." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta is unrelated to a missing-PyTorch import error." + "reason": "Flash Attention 2 with XLMRoberta is unrelated to a missing PyTorch import error." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed training RuntimeError and tokenizer slow/fast inconsistency are unrelated." + "reason": "DeepSpeed training runtime error and tokenizer slow/fast inconsistency are unrelated." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are video-input issues in Qwen variants, but they involve different models and failure signatures; not the same concrete bug." - }, - { - "left": "issue:44568", - "right": "issue:45488", - "accept": false, - "reason": "Both are tokenizer regressions, but one is special-token insertion and the other is a pre-tokenizer override; different mechanisms." + "reason": "Both are Qwen video-input problems, but they involve different models and different errors (ValueError vs StopIteration), so they do not look like the same bug." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Packaging/backend issue versus a specific model-loading failure for cjvt/sleng-bert." + "reason": "Packaging/backend omission and a specific model-loading failure for sleng-bert are not the same issue." }, { - "left": "issue:44987", - "right": "issue:45907", + "left": "issue:44291", + "right": "issue:44821", "accept": false, - "reason": "Model-loading regression and CI failure in list_repo_files are unrelated." + "reason": "Unexpected _is_hf_initialized argument during empty-weight loading and URL image-processor loading are separate code paths." }, { - "left": "issue:44291", - "right": "issue:44821", + "left": "issue:44987", + "right": "issue:45907", "accept": false, - "reason": "init_empty_weights TypeError and AutoImageProcessor URL loading are different code paths." + "reason": "Model loading failure under transformers>=5.1.0 and a missing revision-forwarding bug in list_repo_files are unrelated." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Same broad Qwen video area, but the models and error modes differ enough that they do not look like the same underlying bug." + "reason": "Same general Qwen video area, but different models and different regressions; too broad to treat as one duplicate." }, { "left": "issue:44479", "right": "issue:45381", "accept": false, - "reason": "The second is a specific qwen2.5-vl vision_position_ids complaint, while the first is a broader multi-model video regression; related theme, but not clearly one bug." + "reason": "Both affect Qwen video inputs, but one is a broader regression across models and the other is specifically wrong vision_position_ids in qwen2.5-vl; not the same concrete bug." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but one is mdeberta special-token insertion and the other is LlamaTokenizer pre-tokenizer override; different root causes." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Siglip2 zero-probability behavior and Perceiver non-default-resolution failure are unrelated vision-model bugs." + "reason": "Unexpected zero probabilities in SigLIP2 and Perceiver interpolation failure are unrelated vision-model bugs." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token-count updates and CustomTrainer compute_loss signature mismatch are separate trainer API issues." + "reason": "TrainerState token counting and custom compute_loss signature mismatch are separate Trainer API issues." } ] }, @@ -6857,12 +6839,12 @@ "issue:44568|issue:46491", "issue:43531|issue:44821", "issue:45405|issue:45600", - "issue:43577|issue:45072", "issue:43906|issue:46468", + "issue:43577|issue:45072", "issue:45823|issue:46463", "issue:43452|issue:44488", - "issue:43404|issue:45987", "issue:43906|issue:46693", + "issue:43404|issue:45987", "issue:44737|issue:45588", "issue:44625|issue:44877", "issue:43526|issue:45479", @@ -6870,23 +6852,23 @@ "issue:44913|issue:46121", "issue:43540|issue:46693", "issue:28282|issue:32090", - "issue:43697|issue:45216", - "issue:24643|issue:29127" + "issue:24643|issue:29127", + "issue:43697|issue:45216" ], - "prepared_review_unit_hash": "4263b45a3e1f442d921f1ac6f4f245f41e662bc8cfc40da824b9c2eec840b304", + "prepared_review_unit_hash": "3912f437c7cc66de466c2ff82a5a357fdd76e164634b5d808510c69cdf948f85", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35593, + "serialized_chars": 35594, "estimated_input_tokens": 8899, "estimated_eval_tokens": 18054 }, @@ -6895,174 +6877,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set: the issues span unrelated bugs across model loading, tokenizer behavior, config serialization, dtype/device mismatches, vision/multimodal paths, and performance regressions. None of the soft pairs look like the same underlying bug.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This cluster is mostly a grab bag of unrelated bug reports across different models, tokenizers, configs, and runtime paths. The soft-similarity links do not look like true duplicates, so I reject all of them and see no strong canonical duplicate target.", + "confidence": 0.97, + "canonical_issue_reason": "There is no single true canonical issue here because the items span unrelated subsystems and model families; if one representative must be chosen, issue:45920 is the broadest current regression report.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue:45920 is the most representative as a broad, active regression affecting multiple models/tokenizers; however, the set is not a coherent duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Both mention Qwen/weights-like behavior, but one is DeepSpeed loading conversion for MOE weights and the other is RoPE layer count mismatch in Smollm3; different models and code paths." + "reason": "Different models and different failure modes: Qwen3 MoE weight conversion vs Smollm3 RoPE layer dropping." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both are multimodal/model-loading failures, but one is a janus test failure and the other is a Phi-4 multimodal load regression; no shared underlying bug." + "reason": "A Janus test failure and a Phi-4 multimodal load regression are not the same underlying bug." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Tokenizer-related errors, but one is AddedToken argument duplication and the other is a Mistral regex patch crashing on a tokenizer attribute; distinct failures." + "reason": "Both are tokenizer-related, but one is an AddedToken argument conflict and the other is a Mistral regex/backend_tokenizer crash." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "Both are loading errors, but one is a missing key for MobileLLM and the other is a Float8 storage deserialization problem; unrelated causes." + "reason": "Unrelated issues: MobileLLM key error versus missing Float8 storage object." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Both involve Qwen3.5, but one is missing/dropped RoPE layers while the other is kernel registration failing for apply_rotary_pos_emb; not the same bug." + "reason": "Different Qwen/Smollm model-specific rotary/kernel problems; not the same code-path bug." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "One is a torchvision dependency issue in PIL image processors; the other is a tokenizer regex crash. Different subsystems." + "reason": "Image-processor torchvision dependency regression is unrelated to Mistral tokenizer regex patching." }, { "left": "issue:43716", "right": "issue:46701", "accept": false, - "reason": "Both touch multimodal models, but one is a dtype mismatch in Mistral-3 preprocessing and the other is a DiffusionGemma generation/pipeline failure; no common code-path." + "reason": "Different multimodal models and different symptoms: dtype mismatch versus pipeline failure." }, { "left": "issue:44737", "right": "issue:46463", "accept": false, - "reason": "An XLNet CPU-device bug and a T5 error-message issue are unrelated." + "reason": "CPU device placement bug in XLNet is unrelated to T5 decoder-input validation messaging." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen-related, but one is RoPE layer dropping in Smollm3 and the other is still-image temporal position scaling in Qwen2.5-VL; different logic." + "reason": "Both involve RoPE-adjacent topics, but they concern different Qwen variants and different logic errors." }, { "left": "issue:44568", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but one misses BOS/EOS insertion and the other strips a leading space on round-trip; separate behavior." + "reason": "Tokenizer regressions, but different models and different behaviors: BOS/EOS omission versus leading-space loss." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to an AutoImageProcessor URL loading failure." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to loading AutoImageProcessor from URL." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "One is a dependency version bump problem; the other is stale auto_mappings references causing CI breakage. Different issues." + "reason": "Both affect CI/runtime setup, but one is a dependency version bump and the other is stale auto-mapping references." }, { - "left": "issue:43577", - "right": "issue:45072", + "left": "issue:43906", + "right": "issue:46468", "accept": false, - "reason": "Both mention dtype mismatches, but they affect different models and forward paths (Blip2 vs SwitchTransformers/TimmWrapperModel)." + "reason": "A reproduction of an older issue and a Nemotron example breakage are not the same bug." }, { - "left": "issue:43906", - "right": "issue:46468", + "left": "issue:43577", + "right": "issue:45072", "accept": false, - "reason": "An isolated reproduction of an older issue and a broken Nemotron example are not the same bug." + "reason": "Both mention dtype, but they affect different models and different forward/inference paths." }, { "left": "issue:45823", "right": "issue:46463", "accept": false, - "reason": "Gemma4 device-map mismatch is unrelated to a T5 decoder-input validation error message." + "reason": "A device-map mismatch in Gemma4 PLE is unrelated to T5 decoder input validation." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both concern loading models/tokenizers, but one is a gguf_file-from_pretrained regression and the other is a specific model failing to load; no shared concrete fix." + "reason": "Both are loading-related, but the evidence is too weak to treat them as the same gguf/model-loading bug." }, { - "left": "issue:43404", - "right": "issue:45987", + "left": "issue:43906", + "right": "issue:46693", "accept": false, - "reason": "Mistral3 lm_head tying and StaticCache return type are unrelated bugs." + "reason": "An isolated reproduction of one issue is unrelated to a long-context FA2 performance regression." }, { - "left": "issue:43906", - "right": "issue:46693", + "left": "issue:43404", + "right": "issue:45987", "accept": false, - "reason": "An isolated reproduction issue and a long-context FA2 performance regression do not share the same underlying defect." + "reason": "Weight tying in Mistral3 is unrelated to StaticCache sequence-length return type." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet device placement bug and flash_attention s_aux=None crash are different code paths." + "reason": "XLNet CPU tensor creation and flash_attention s_aux=None crashes are different code paths." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "One is num_labels propagation in Qwen3.5 config; the other is strict config blocking granite_speech loading. Different config problems." + "reason": "A num_labels propagation bug and a strict-config loading failure are not the same issue." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "A BeitImageProcessorFast reduce_labels bug and a sequence-classification zero-loss issue are not the same problem." + "reason": "ImageProcessor label reduction and sequence-classification loss degeneracy are different bugs." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Qwen3.5 save_pretrained visual-key regression and AutoTokenizer wrong token IDs are separate regressions." + "reason": "save_pretrained visual-encoder key regression is unrelated to tokenizer ID corruption." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "Both involve RoPE/config handling, but one silently resets rotary_pct on reload while the other errors on JSON-loaded ignore_keys lists; distinct failure modes." + "reason": "Both mention rope/config handling, but one is silent default reversion on reload and the other is a TypeError in validation." }, { "left": "issue:43540", "right": "issue:46693", "accept": false, - "reason": "Qwen3OmniMoe video input processing and FA2 long-context performance regression are unrelated." + "reason": "Video input processing failure in Qwen3OmniMoe is unrelated to FA2 long-context performance regression." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "A missing PyTorch import and a NoneType broadcast error under Trainer are unrelated." + "reason": "Missing PyTorch import and _gpu_broadcast_one NoneType handling are completely different failures." }, { - "left": "issue:43697", - "right": "issue:45216", + "left": "issue:24643", + "right": "issue:29127", "accept": false, - "reason": "RTDetrV2 output differences and Qwen3.5 save_pretrained checkpoint corruption are different regressions." + "reason": "DeepSpeed training weight-shape runtime error is unrelated to a LayoutLMv3 error-message clarity issue." }, { - "left": "issue:24643", - "right": "issue:29127", + "left": "issue:43697", + "right": "issue:45216", "accept": false, - "reason": "DeepSpeed training weight-shape runtime error and a LayoutLMv3 box-validation message issue are unrelated." + "reason": "RTDetrV2 output changes and Qwen3.5 save_pretrained regression are different symptoms and code paths." } ] }, @@ -7119,7 +7101,6 @@ "issue:44843", "issue:44898", "issue:44936", - "issue:44991", "issue:45005", "issue:45200", "issue:45356", @@ -7142,232 +7123,225 @@ "issue:44466|issue:45005", "issue:45200|issue:46463", "issue:41762|issue:44936", - "issue:43688|issue:43749", - "issue:44493|issue:46050", "issue:44361|issue:44442", + "issue:43688|issue:43749", "issue:43452|issue:44843", + "issue:44493|issue:46050", "issue:43653|issue:44361", - "issue:43525|issue:46701", "issue:43819|issue:44811", + "issue:43525|issue:46701", "issue:43994|issue:44898", "issue:38175|issue:43295", "issue:45372|issue:46170", "issue:44936|issue:45823", - "issue:45542|issue:46178", "issue:43531|issue:43742", + "issue:45542|issue:46178", "issue:42898|issue:46491", "issue:43646|issue:43749", - "issue:43295|issue:45356", "issue:43653|issue:44568", + "issue:43295|issue:45356", "issue:43577|issue:43697", "issue:29159|issue:33357", - "issue:43452|issue:43577", - "issue:44843|issue:44991", "issue:44464|issue:44743", + "issue:43452|issue:43577", "issue:43688|issue:45704" ], - "prepared_review_unit_hash": "22b56d0f0b5dc0b96bbddc039eb57a4f70ed4e6457ecb93a80079f2490b54d77", + "prepared_review_unit_hash": "1343194a1e5e3b428bdf67fd01ae3da37889f62eb8747b2090cbd8811c236608", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36756, - "estimated_input_tokens": 9189, - "estimated_eval_tokens": 18634 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35905, + "estimated_input_tokens": 8977, + "estimated_eval_tokens": 18210 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are mostly unrelated, spanning distinct tokenizer, loading, config, generation, and backend regressions. I found no soft-edge pair that clearly describes the same concrete bug or change; the closest candidates are still too different in model, code path, or symptom.", - "confidence": 0.62, - "canonical_issue_reason": "issue:43646 is the broadest umbrella regression here: Transformers 5.0.0 breaking custom model initialization. The other reports are narrower, model- or tokenizer-specific breakages rather than the same underlying issue.", + "summary": "This cluster is very heterogeneous: most items are distinct tokenizer, model-loading, or regression reports, with only a weak tied-weights subtheme standing out. No pull requests are present in the cluster.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45005 is the broadest and most detailed issue around the only clearly recurring theme here: v5 tied-weights behavior. It can plausibly serve as the umbrella report for the related lm_head.weight serialization symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best representative of the cluster because it captures the general v5 breaking-change theme and is the most broadly applicable, actionable report in the set.", + "best_issue_reason": "issue:45005 has the richest scope and discussion among the partially related items, making it the strongest single issue to anchor triage for the tied-weights regression thread.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Different failures: BitNet packed-weight handling during accelerate loading vs Perceiver image-resolution forward behavior." + "reason": "Different models and unrelated failure modes; no shared concrete bug." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "ImportError due to missing PyTorch is unrelated to the post-init tied-embedding reinitialization bug." + "reason": "ImportError for missing PyTorch vs embedding resize/reinit regression are unrelated." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Both involve missing config fields, but they are different models and different missing attributes." + "reason": "Both involve config fields, but the underlying causes and affected paths differ." }, { "left": "issue:44466", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but one is device-dependent lm_head serialization and the other is a broader translation-model complaint; not enough to treat as the same bug." + "reason": "Both mention tied weights, but one is about lm_head.weight serialization by device and the other is a broader translation-model regression; not clearly the same bug." }, { "left": "issue:45200", "right": "issue:46463", "accept": false, - "reason": "Gemma 4 mm_token_type_ids defaulting and T5 decoder input validation are distinct code paths and symptoms." + "reason": "Gemma 4 multimodal input defaults vs T5 decoder-input validation are different issues." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed ZeRO-3 loading IndexError is not the same as trainer.evaluate() failing after train." + "reason": "DeepSpeed ZeRO-3 loading crash vs trainer.evaluate-after-train failure are unrelated." }, { - "left": "issue:43688", - "right": "issue:43749", + "left": "issue:44361", + "right": "issue:44442", "accept": false, - "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated problems." + "reason": "Different tokenizers and different errors; no evidence of the same code path." }, { - "left": "issue:44493", - "right": "issue:46050", + "left": "issue:43688", + "right": "issue:43749", "accept": false, - "reason": "Position-id key warnings and sinq quantization save/load breakage are different issues." + "reason": "Aux-loss normalization bug and FSDP loading breakage are unrelated." }, { - "left": "issue:44361", - "right": "issue:44442", + "left": "issue:43452", + "right": "issue:44843", "accept": false, - "reason": "Different tokenizer classes and different errors; no shared concrete bug path is evident." + "reason": "Both hit from_pretrained flows, but gguf_file handling and offline model_info patching are different bugs." }, { - "left": "issue:43452", - "right": "issue:44843", + "left": "issue:44493", + "right": "issue:46050", "accept": false, - "reason": "gguf_file loading support and offline hub model_info() calls are distinct AutoTokenizer failures." + "reason": "Position-id key warnings and sinq save/load quantization breakage are different problems." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird special-token registration is unrelated to MLukeTokenizer task AttributeError failures." + "reason": "Different tokenizer-specific failures with no shared underlying defect." }, { - "left": "issue:43525", - "right": "issue:46701", + "left": "issue:43819", + "right": "issue:44811", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is not the same as the DiffusionGemma image-text pipeline failure." + "reason": "DAC forward/latents mismatch and Whisper batch_decode behavior are unrelated." }, { - "left": "issue:43819", - "right": "issue:44811", + "left": "issue:43525", + "right": "issue:46701", "accept": false, - "reason": "DAC latent reconstruction mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." + "reason": "Llama4 config attribute error and DiffusionGemma pipeline failure are unrelated." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Both are vision-model issues, but the concrete bugs and code paths differ." + "reason": "Both involve vision models, but they describe different symptoms and likely different root causes." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Siglip2 zero-probability output and processor.tokenizer/custom image handling regressions are different failures." + "reason": "SigLIP2 zero-probabilities and processor/tokenizer regression are different bugs." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed dependency issues are separate import problems." + "reason": "Gemma processor import dependency breakage and Windows ROCm torch.distributed issue are unrelated." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "Trainer lifecycle failure and Gemma4 device-map forward mismatch are not the same bug." + "reason": "Trainer lifecycle failure and Gemma4 device-mismatch forward bug are unrelated." }, { - "left": "issue:45542", - "right": "issue:46178", + "left": "issue:43531", + "right": "issue:43742", "accept": false, - "reason": "Both are backend mapping errors, but they involve different missing optional dependencies and different error surfaces." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM loading key error are unrelated." }, { - "left": "issue:43531", - "right": "issue:43742", + "left": "issue:45542", + "right": "issue:46178", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key errors are unrelated." + "reason": "Different optional-backend dependency errors; not the same backend mapping bug." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Both concern tokenizer behavior changes, but one is a generic v5 setting change and the other is a specific CodeLlama round-trip regression." + "reason": "Generic v5 cleanup behavior change vs CodeLlama leading-space regression are not clearly the same bug." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization regressions are not the same as FSDP CPU RAM efficient loading breakage." + "reason": "Custom model initialization breakage and FSDP loading breakage are unrelated." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:43653", + "right": "issue:44568", "accept": false, - "reason": "Different tokenizer regressions: custom processor/image-token handling vs Kimi-K2.5 codec and warning behavior." + "reason": "Different tokenizer special-token bugs affecting different models." }, { - "left": "issue:43653", - "right": "issue:44568", + "left": "issue:43295", + "right": "issue:45356", "accept": false, - "reason": "Both are special-token bugs, but they affect different tokenizers and different token types." + "reason": "Both are tokenizer regressions, but the affected flows and described failure modes differ." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "BLIP2 dtype propagation and RTDetrV2 output divergence are unrelated model behaviors." + "reason": "BLIP2 dtype propagation and RTDetrV2 output divergence are unrelated." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Tokenizer consistency issues and a MacOS bus error in CLIP loading are clearly different failures." - }, - { - "left": "issue:43452", - "right": "issue:43577", - "accept": false, - "reason": "gguf_file loading breakage is unrelated to BLIP2 dtype casting staying float32." + "reason": "Slow/fast tokenizer inconsistency and MacOS CLIP bus error are unrelated." }, { - "left": "issue:44843", - "right": "issue:44991", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Offline hub lookup regression and est-roberta tokenizer loading failure are different problems." + "reason": "Chunked generation with compiled forward and Qwen3 recurrent-state reset are different generation bugs." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:43452", + "right": "issue:43577", "accept": false, - "reason": "Chunked generation with compiled forward and recurrent-state reset with cache are different generation bugs." + "reason": "gguf_file loading breakage and BLIP2 dtype mismatch are unrelated." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Auxiliary-loss normalization and the T5 apex RMSNorm memory leak are unrelated." + "reason": "Aux-loss normalization and T5 apex RMSNorm memory leak are unrelated." } ] }, @@ -7412,7 +7386,6 @@ "issue:43742", "issue:43756", "issue:43782", - "issue:43825", "issue:43927", "issue:44361", "issue:44464", @@ -7426,8 +7399,10 @@ "issue:44704", "issue:44734", "issue:44821", + "issue:44843", "issue:44964", "issue:44977", + "issue:44991", "issue:45405", "issue:45440", "issue:45491", @@ -7441,12 +7416,13 @@ ], "soft_pairs": [ "issue:43756|issue:46402", + "issue:44843|issue:44991", "issue:43653|issue:43927", "issue:43688|issue:45701", "issue:44479|issue:44625", "issue:45491|issue:46002", - "issue:43688|issue:45440", "issue:43550|issue:44361", + "issue:43688|issue:45440", "issue:44484|issue:44704", "issue:33666|issue:36010", "issue:43742|issue:44821", @@ -7456,13 +7432,13 @@ "issue:41720|issue:44464", "issue:44464|issue:44977", "issue:43504|issue:43720", - "issue:44977|issue:46693", "issue:43504|issue:43550", - "issue:43643|issue:44561", + "issue:44977|issue:46693", "issue:43504|issue:44610", "issue:43577|issue:44554", "issue:43720|issue:43927", "issue:33666|issue:36331", + "issue:43643|issue:44561", "issue:43606|issue:43927", "issue:43278|issue:43688", "issue:43782|issue:46693", @@ -7470,229 +7446,228 @@ "issue:39692|issue:43493", "issue:43262|issue:43295", "issue:44554|issue:46365", - "issue:43720|issue:44361", - "issue:43493|issue:43825" + "issue:43720|issue:44361" ], - "prepared_review_unit_hash": "60d411667e3e2fef9e9c4f9bc7c7eec11cdd91bac52732f8e8ee3ea37b249a82", + "prepared_review_unit_hash": "bb9b12373a7dd161aec76aca1ca8772431c32f93e4e39479391ab4e34a147209", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36842, - "estimated_input_tokens": 9211, - "estimated_eval_tokens": 18678 + "serialized_chars": 37428, + "estimated_input_tokens": 9357, + "estimated_eval_tokens": 18970 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs look like near-topic matches at best, not true duplicates: they involve different models, different failure modes, or different code paths. No safe canonical item emerges from this heterogeneous cluster.", - "confidence": 0.95, - "canonical_issue_reason": "No safe canonical issue; the cluster is heterogeneous and does not center on one underlying bug.", + "summary": "This set is highly heterogeneous: the items cover unrelated bugs, regressions, model-specific issues, and docs/examples. No soft-edge pair looks like the same underlying change, so I reject all candidate duplicates and do not nominate a single canonical issue/PR.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a strong representative because the items span unrelated reports across loading, generation, tokenization, and docs.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Different bugs: Smollm3 RoPE-layer handling vs Sam3Model text-encoder weights not loading." + "reason": "Different model families and failure modes: Smollm3 RoPE layer mismatch vs Sam3 text encoder weights not loading." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but one is offline-model_info patching and the other is a specific tokenizer failure for EMBEDDIA/est-roberta; not the same bug." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Both involve token-related behavior, but one is BigBirdTokenizer special-token registration and the other is DiaConfig save/load of custom IDs." + "reason": "BigBirdTokenizer special-token decode bug is unrelated to DiaConfig losing custom token IDs on save/load." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Aux-loss normalization in MoE models is unrelated to a transformers-version tokenization change." + "reason": "Aux-loss normalization in MoE models is unrelated to a broad tokenization-version change report." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression for Qwen VL models is a different code path from Qwen3.5 num_labels propagation." + "reason": "Video input regression for Qwen VL models is unrelated to num_labels propagation in Qwen3.5 config." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention and the other is classification-head init/checkpoint behavior." + "reason": "Gemma3 mixed-length NaNs in attention are unrelated to classification-head NaNs during checkout/init." }, { - "left": "issue:43688", - "right": "issue:45440", + "left": "issue:43550", + "right": "issue:44361", "accept": false, - "reason": "Different model families and different defects: loss normalization vs DeepSeekV3 implementation divergence." + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." }, { - "left": "issue:43550", - "right": "issue:44361", + "left": "issue:43688", + "right": "issue:45440", "accept": false, - "reason": "Torch.compile/SDPA issues in Bamba are unrelated to MLukeTokenizer task AttributeError." + "reason": "MoE auxiliary-loss scaling bug is unrelated to DeepseekV3MoE implementation divergence." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "Save-pretrained shard-size behavior is unrelated to AutoProcessor passing kwargs to cached_file." + "reason": "A save_pretrained shard-size question is unrelated to AutoProcessor not forwarding kwargs to cached_file." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU training question for Qwen2-VL is not the same as the GenerationMixin import error." + "reason": "Qwen2-VL multi-GPU training is unrelated to an import error for GenerationMixin." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "A MobileLLM load key error is a different failure than AutoImageProcessor loading from a URL." + "reason": "MobileLLM key-error loading is unrelated to AutoImageProcessor failing from URL." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error is a distinct training crash from a general Qwen2-VL multi-GPU training request." + "reason": "Trainer _gpu_broadcast_one NoneType error is a generic training crash, not the same as Qwen2-VL multi-GPU training setup issue." }, { "left": "issue:44734", "right": "issue:46612", "accept": false, - "reason": "Different generation bugs: server KV-cache continuation indexing vs beam-search cache reorder not running for certain models." + "reason": "KV-cache continuation indexing in transformers serve is unrelated to beam-search cache reorder omissions for specific model classes." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "lm_head.weight serialization inconsistency is unrelated to sinq quantized model save/load breakage." + "reason": "Inconsistent lm_head.weight serialization is unrelated to quantized model save/load breakage." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Auto device mapping cudaErrorAssert on Qwen3 is a different issue from compiled-forward chunked generation inconsistency." + "reason": "Qwen3 auto device mapping cudaErrorAssert is unrelated to chunked generation inconsistencies under compiled forward." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Both are generation regressions, but one is compiled-forward chunking and the other is flash-attention behavior in Qwen3.5." + "reason": "Chunked generation under compiled forward is unrelated to Qwen3.5 flash-attention generation failure." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Beit legacy-field loading and BitNet packed-weight unloading are unrelated model-loading bugs." - }, - { - "left": "issue:44977", - "right": "issue:46693", - "accept": false, - "reason": "Flash-attention correctness/generation failure is not the same bug as a performance regression with FA2 and long-context generation." + "reason": "BEiT legacy-field loading failure is unrelated to BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Different models and failure modes: pretrained preset loading vs torch.compile with SDPA." + "reason": "BEiT pretrained preset loading failure is unrelated to Bamba torch.compile SDPA behavior." }, { - "left": "issue:43643", - "right": "issue:44561", + "left": "issue:44977", + "right": "issue:46693", "accept": false, - "reason": "AutoConfig missing fields under trust_remote_code is not the same as removing is_torch_fx_available breaking remote-code models." + "reason": "Qwen3.5 flash-attention correctness bug is unrelated to a general FA2 long-context performance regression." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Legacy-field preset loading and OmDet-Turbo processor/model size mismatch are unrelated." + "reason": "BEiT legacy-field load bug is unrelated to OmDet-Turbo processor/model image-size mismatch." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype propagation bug is separate from an MPS attention correctness issue." + "reason": "Blip2 dtype propagation bug is unrelated to MPS attention correctness when value head dim differs from query." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight loading in BitNet is unrelated to DiaConfig custom-token-ID loss." + "reason": "BitNet accelerate-loading unpack bug is unrelated to DiaConfig losing custom token IDs." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Qwen2-VL multi-GPU training and CustomTrainer.compute_loss signature mismatch are different trainer issues." + "reason": "Qwen2-VL multi-GPU training is unrelated to CustomTrainer.compute_loss signature changes." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch in Bark-small does not match DiaConfig save/load token ID loss." + "reason": "CPU offload device mismatch for bark-small is unrelated to DiaConfig token-ID loss." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Embedding dtype drift during train/eval is unrelated to auxiliary-loss normalization." + "reason": "BF16-to-FP32 dtype drift in evaluation is unrelated to MoE auxiliary-loss normalization." }, { "left": "issue:43782", "right": "issue:46693", "accept": false, - "reason": "Qwen3VL weight_only loading failure is a different issue from FA2 long-context performance regression." + "reason": "Qwen3VL weight_only load error is unrelated to FA2 long-context performance regression." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Phi-4 multimodal loading failure is unrelated to the PEFT version bump/release issue." + "reason": "Phi-4 multimodal load failure is unrelated to a PEFT version pin being ahead of PyPI release." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "SigLIP2 doc-example errors are not the same as an HF-vs-JAX implementation discrepancy." + "reason": "SigLIP2 docs example errors are unrelated to SigLIP2 implementation discrepancy with the original JAX version." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio apply_chat_template sampling-rate defaulting and processor.tokenizer/image-input regression are different processor bugs." + "reason": "Audio chat-template sampling-rate default bug is unrelated to processor.tokenizer access / passing images to tokenizer regression." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness is unrelated to importing Trainer being blocked by datasets." + "reason": "MPS attention correctness issue is unrelated to Trainer import failure caused by datasets installation." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError share no concrete code path." - }, - { - "left": "issue:43493", - "right": "issue:43825", - "accept": false, - "reason": "SigLIP2 implementation divergence is unrelated to a pipeline error-message wording regression." + "reason": "BitNet packed-weight loading issue is unrelated to MLukeTokenizer task AttributeError." } ] }, @@ -7725,7 +7700,6 @@ "issue:43278", "issue:43295", "issue:43299", - "issue:43334", "issue:43452", "issue:43493", "issue:43504", @@ -7736,6 +7710,7 @@ "issue:43646", "issue:43701", "issue:43824", + "issue:43825", "issue:43844", "issue:43927", "issue:43994", @@ -7766,245 +7741,245 @@ "issue:46710" ], "soft_pairs": [ + "issue:43493|issue:43825", "issue:44514|issue:45290", "issue:43525|issue:43572", "issue:43452|issue:44451", - "issue:44977|issue:46209", "issue:44625|issue:45584", + "issue:44977|issue:46209", "issue:45823|issue:46050", "issue:28218|issue:33357", - "issue:42796|issue:44977", "issue:43994|issue:46192", + "issue:42796|issue:44977", "issue:43278|issue:43844", "issue:43493|issue:43844", "issue:44849|issue:46209", "issue:39692|issue:43295", - "issue:43646|issue:44466", "issue:33357|issue:33666", + "issue:43646|issue:44466", "issue:44568|issue:46489", - "issue:46209|issue:46710", - "issue:44488|issue:44843", "issue:44849|issue:45357", + "issue:46209|issue:46710", "issue:43278|issue:43701", "issue:44792|issue:45405", + "issue:44488|issue:44843", "issue:43575|issue:45405", "issue:43927|issue:44361", - "issue:44568|issue:45812", "issue:44589|issue:45520", "issue:43257|issue:43299", "issue:43824|issue:44589", + "issue:44568|issue:45812", "issue:43504|issue:43606", - "issue:43606|issue:44112", - "issue:43334|issue:43572" + "issue:43606|issue:44112" ], - "prepared_review_unit_hash": "284fc736c4d63d051564fe1879585b9533c850ae5632d0e81ee0a73366f4fdbe", + "prepared_review_unit_hash": "3aa4dbfc9e7a287aeaf415f82604fffc341cadc71f9338846dd1a1edef16e08b", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36731, - "estimated_input_tokens": 9183, - "estimated_eval_tokens": 18622 + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a noisy mix of mostly unrelated regressions: tokenizer quirks, model-loading failures, config-field bugs, and a few training/inference issues. None of the soft pairs look like the same concrete bug/change, so I would not collapse any of them.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45812 is the most representative and highest-scope issue in the tokenizer-regression subset: it describes a broad silent v4\u2192v5 tokenization failure affecting all Granite models, which is more central than the narrower model-specific loading/config reports around it.", + "summary": "This cluster is highly heterogeneous. The soft-similarity pairs are mostly false positives across different models, code paths, and failure modes; none looks like the same underlying bug/change closely enough to merge.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45812 is the broadest tokenizer-regression report in the set and has the strongest downstream impact among the tokenizer-related items, so it is the best umbrella representative if one issue must anchor the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the best single anchor for this cluster because it is broad, user-visible, and clearly about a concrete tokenizer regression rather than a one-off model-specific edge case.", + "best_issue_reason": "45812 is the most general and widely relevant issue here: it describes a silent tokenizer regression affecting all Granite models, which makes it the best candidate for duplicate triage within this otherwise mixed set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "Both mention model/pipeline correctness, but one is a SigLIP2 implementation mismatch and the other is a pipeline task-support error; different subsystems and symptoms." + }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both involve `apply_chat_template`, but one is a batched-padding crash and the other is a tool-call/no-content crash; different edge cases, not the same bug." + "reason": "Both are apply_chat_template crashes, but the triggers differ: batched Qwen2.5-VL with padding=False versus assistant tool-call messages with no content." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Similar missing-token-id theme, but different configs/models (`Llama4Config` vs `StableLmConfig`) and different missing fields; not one shared defect." + "reason": "These are similar migration-style config errors, but they involve different models and missing fields (pad_token_id vs pad_token_idx), so they are not the same bug." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "Both are loading failures, but the models and failure modes differ; no sign of the same underlying loader bug." + "reason": "One is a gguf_file/from_pretrained breakage, the other is a model-specific load failure for ScandiBERT; same broad area, but not the same code-path problem." }, { - "left": "issue:44977", - "right": "issue:46209", + "left": "issue:44625", + "right": "issue:45584", "accept": false, - "reason": "Separate regressions: Qwen3.5 generation/flash-attn behavior vs FP8 inference failure on B200." + "reason": "Different models and unrelated failure modes: Qwen3.5 num_labels propagation versus Whisper generation after align_special_tokens." }, { - "left": "issue:44625", - "right": "issue:45584", + "left": "issue:44977", + "right": "issue:46209", "accept": false, - "reason": "Unrelated symptoms and code paths: label propagation in Qwen3.5 vs Whisper empty-transcription handling after special-token alignment." + "reason": "These are separate runtime issues: Qwen3.5 flash-attention generation versus FP8 inference failure on B200 hardware." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "One is a device-map forward mismatch; the other is quantization save/load corruption. Different problems." + "reason": "One is a device-map forward mismatch, the other is quantized save/load breakage; different stages and likely different fixes." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug vs Mac bus error in a CLIP model; completely different issues." + "reason": "The tokenizer-added-space bug and the MacOS CLIP bus error are unrelated." }, { - "left": "issue:42796", - "right": "issue:44977", + "left": "issue:43994", + "right": "issue:46192", "accept": false, - "reason": "Tokenizer BPE-marker handling vs Qwen3.5 flash-attention generation; not the same defect." + "reason": "SigLIP2 nonsense outputs and silently ignored return_tensors are distinct API bugs with different triggers." }, { - "left": "issue:43994", - "right": "issue:46192", + "left": "issue:42796", + "right": "issue:44977", "accept": false, - "reason": "SigLIP2 output quality issue vs `return_tensors` being ignored when `text_kwargs` is passed; unrelated." + "reason": "Tokenizer BPE-marker handling and Qwen3.5 flash-attention generation are unrelated issues." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "BF16\u2192FP32 dtype drift and a ZeRO-3 gradient anomaly are different training issues." + "reason": "Embedding dtype drift in evaluate and gradient explosion under ZeRO-3 are different training/runtime problems." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 HF/JAX discrepancy vs DeepSpeed gradient growth; no shared concrete bug." + "reason": "A SigLIP2 discrepancy versus a ZeRO-3 gradient regression are clearly different bugs." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Different regressions: Qwen3.5 hidden-states behavior vs FP8 inference support on B200." + "reason": "Qwen3.5 output_hidden_states behavior and FP8 inference on B200 do not share the same underlying cause." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Documentation/example errors vs a processor/tokenizer regression in custom model code; not the same issue." + "reason": "Documentation example errors are not the same as the processor/tokenizer API regression in custom model code." }, { - "left": "issue:43646", - "right": "issue:44466", + "left": "issue:33357", + "right": "issue:33666", "accept": false, - "reason": "Custom model initialization breakage vs serialization of tied weights; separate code paths." + "reason": "A MacOS bus error in CLIP loading and Qwen2-VL multi-GPU training are unrelated." }, { - "left": "issue:33357", - "right": "issue:33666", + "left": "issue:43646", + "right": "issue:44466", "accept": false, - "reason": "Mac bus error in a CLIP model vs Qwen2-VL multi-GPU training; unrelated." + "reason": "Both are v5 regressions, but one is custom model initialization and the other is device-dependent serialization of tied weights." }, { "left": "issue:44568", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and behaviors; not enough evidence of one shared bug." + "reason": "Both are tokenizer regressions, but they affect different models and different mechanisms; they are not the same concrete bug." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Output-hidden-states behavior and incorrect visual encoder key saving are separate Qwen3.5 issues." }, { "left": "issue:46209", "right": "issue:46710", "accept": false, - "reason": "FP8 inference failure vs accuracy regression after upgrade; different symptoms and likely different causes." + "reason": "FP8 inference on B200 and a DeepSeek-R1-Distill-Llama accuracy regression after upgrade are different problems." }, { - "left": "issue:44488", - "right": "issue:44843", + "left": "issue:43278", + "right": "issue:43701", "accept": false, - "reason": "Model-loading failure for `sleng-bert` vs offline-mode breakage in `_patch_mistral_regex`; distinct issues." + "reason": "Embedding dtype changes and resume_from_checkpoint key mismatches do not share the same code path." }, { - "left": "issue:44849", - "right": "issue:45357", - "accept": false, - "reason": "Hidden-states regression in Qwen3.5 vs incorrect visual encoder keys during save/load." - }, - { - "left": "issue:43278", - "right": "issue:43701", + "left": "issue:44792", + "right": "issue:45405", "accept": false, - "reason": "Embedding dtype change vs checkpoint resume key mismatch; no shared underlying bug." + "reason": "A failing Janus image-generation test and an unreleased peft version bump are unrelated." }, { - "left": "issue:44792", - "right": "issue:45405", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "A failed Janus test vs an unreleased PEFT version bump; not the same defect." + "reason": "Both involve model loading/tokenizers, but one is a specific model load failure and the other is an offline-mode regression in a regex patch." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM on a Qwen model vs dependency-version pinning; unrelated." + "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to a dependency version bump." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Custom token IDs lost on save/load vs tokenizer AttributeError on tasks; different model-specific bugs." - }, - { - "left": "issue:44568", - "right": "issue:45812", - "accept": false, - "reason": "Both are tokenizer regressions, but one is missing BOS/EOS on mdeberta-v3 and the other is wrong token IDs on Granite; not the same code-path problem." + "reason": "DiaConfig save/load token-ID loss and MLukeTokenizer task AttributeError are different failures." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage lookup failure vs missing `flash_attn` import key on Python 3.13; unrelated." + "reason": "These are different backend compatibility errors: Float8 storage lookup versus flash_attn import metadata on Python 3.13." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen MOE loading, but one is accelerate+deepspeed conversion and the other is a dev-version model-load break; too different to merge." + "reason": "Both concern Qwen3 MoE models, but one is weight conversion under accelerate+deepspeed and the other is a loading regression for Qwen3VL MoE models." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "ImportError for a Qwen2.5-VL class vs an FP8 storage type error; no relation." + "reason": "An import error for a model class and a Float8 storage object lookup failure are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different model families and different concrete behaviors." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field model preset loading vs CPU-offload device mismatch; different failures." + "reason": "A legacy-field loading failure in BEiT semantic segmentation is unrelated to the Bark CPU offload device mismatch." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both mention device mismatch/offload, but one is a real model failure and the other is a stale CI test; not the same bug." - }, - { - "left": "issue:43334", - "right": "issue:43572", - "accept": false, - "reason": "Both are missing token-id style config regressions, but they hit different models and fields; no evidence of one underlying fix." + "reason": "These both mention devices, but one is a real forward-path mismatch and the other is a stale CI test; not the same underlying bug." } ] }, @@ -8078,9 +8053,10 @@ "issue:46710" ], "soft_pairs": [ + "issue:43334|issue:43572", "issue:43334|issue:45759", - "issue:45245|issue:46710", "issue:43572|issue:44291", + "issue:45245|issue:46710", "issue:44493|issue:46285", "issue:43504|issue:43653", "issue:44464|issue:46424", @@ -8099,203 +8075,185 @@ "issue:44977|issue:46701", "issue:28218|issue:34567", "issue:15354|issue:38175", - "issue:44554|issue:45446", "issue:43065|issue:46369", + "issue:44554|issue:45446", "issue:44423|issue:44734", + "issue:43493|issue:43701", "issue:41720|issue:46710", - "issue:45406|issue:46121", - "issue:43493|issue:43701" + "issue:45406|issue:46121" ], - "prepared_review_unit_hash": "800dad8951afac8e31ba86bbeedd6c487c567b82c70d9336d92beaedf0391731", + "prepared_review_unit_hash": "0e37f31d798ce2663d66e87772e6398793b552ae114f72a7dd3c9dae91f6342f", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35860, - "estimated_input_tokens": 8965, - "estimated_eval_tokens": 18186 + "soft_pair_count": 28, + "serialized_chars": 36157, + "estimated_input_tokens": 9040, + "estimated_eval_tokens": 18336 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous; the soft-similarity pairs mostly link unrelated transformer bugs that share only broad subsystem vocabulary. No soft edge looks like a true duplicate or a single mergeable PR pair.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44464 is the closest thing to a representative issue because it is a concrete runtime regression in generation/forward execution, but the overall cluster is too diverse to support a strong true-canonical choice.", + "summary": "These issues are broadly unrelated, with only loose thematic overlap around model/config loading regressions. No soft pair looks like the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43334 is the most representative of the small subset of load-time config compatibility reports (Qwen multimodal config missing a pad token attribute), but overall the cluster is too heterogeneous to indicate a true duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44464 is the best anchor only by being the clearest concrete bug report; it is still not a good duplicate umbrella for the rest of the cluster.", + "best_issue_reason": "issue:43334 is the clearest, most concrete bug report among the set and sits near the most plausible thematic center of the cluster (model/config loading regressions).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43334", - "right": "issue:45759", + "right": "issue:43572", "accept": false, - "reason": "Different bugs: one is a missing pad_token_id on Qwen3-VL, the other is AutoModelForCausalLM.from_config not unwrapping composite text_config." + "reason": "Both are missing-attribute config load failures, but they affect different models, different fields, and different code paths." }, { - "left": "issue:45245", - "right": "issue:46710", + "left": "issue:43334", + "right": "issue:45759", "accept": false, - "reason": "Unrelated failure modes: category-count limit error vs accuracy regression after upgrade." + "reason": "Both involve Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is incorrect text_config unwrapping; not the same bug." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Different config/API issues: StableLm pad_token_idx regression vs unexpected _is_hf_initialized argument in init_empty_weights." + "reason": "One is a missing config field after a version update; the other is an init_empty_weights/_is_hf_initialized constructor mismatch. Different failures." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "Integer category-limit error and an accuracy regression are unrelated bugs." }, { "left": "issue:44493", "right": "issue:46285", "accept": false, - "reason": "Different code paths: unexpected position-id key handling vs CLIP text model structure divergence." + "reason": "Unexpected position-id key handling and CLIP text model structure inconsistency are separate issues." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Different tokenizer/model bugs: legacy preset field loading vs mask token registration and decode output." + "reason": "Legacy-field model loading failure and tokenizer special-token registration are different problems." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Both touch generation/cache, but one is compiled-forward chunked generation output mismatch and the other is cache-length handling for quantized cache." + "reason": "Compiled-forward chunked-generation inconsistency is not the same as cache_length being ignored except for quantized cache." }, { "left": "issue:43937", "right": "issue:46701", "accept": false, - "reason": "Different problems: invalid GenerationConfig vs image-text-to-text pipeline failure for DiffusionGemma." + "reason": "Invalid GenerationConfig for GLM-5 and a diffusion pipeline failure are different code paths and symptoms." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Different concerns: missing mm_token_type_ids default vs multi-GPU inference OOM." - }, - { - "left": "issue:44464", - "right": "issue:45584", - "accept": false, - "reason": "Different paths: compiled-forward chunked generation bug vs Whisper empty transcription after align_special_tokens." - }, - { - "left": "issue:45229", - "right": "issue:45823", - "accept": false, - "reason": "Same model family, but one is OOM and the other is a device_map auto placement mismatch." + "reason": "Defaulting mm_token_type_ids and CUDA OOM in multi-GPU inference are unrelated." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "Different bugs: ZeRO-3 gradient growth vs Phi-4 multimodal loading failure." + "reason": "DeepSpeed/ZeRO gradient growth and Phi-4 multimodal loading failure are not the same bug." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Both are quantization-related, but one is offloading behavior and the other is save/load corruption after sinq quantization." - }, - { - "left": "issue:43994", - "right": "issue:44493", - "accept": false, - "reason": "Different symptoms and layers: bad SigLIP2 outputs vs unexpected key in position-id handling." - }, - { - "left": "issue:29942", - "right": "issue:30333", - "accept": false, - "reason": "Completely unrelated: Flash Attention 2 test failures vs MLFlow reporting after failed training jobs." + "reason": "Quantized offloading behavior and sinq save/load breakage are different quantization-related issues, not one concrete defect." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "Different code paths: flash_attention.py crash on s_aux=None vs sinq model persistence breakage." + "reason": "Flash-attention sink-less crash and sinq serialization breakage have no shared underlying code-path failure." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Tokenizer parity issue vs Trainer distributed broadcast TypeError; not the same bug." + "reason": "Tokenizer slow/fast inconsistency and _gpu_broadcast_one NoneType errors are unrelated." }, { - "left": "issue:42490", - "right": "issue:43696", + "left": "issue:29942", + "right": "issue:30333", "accept": false, - "reason": "Unrelated: trailing-slash local path import warning vs CUDA out-of-memory on GPT-oss-20b." + "reason": "Flash Attention 2 test failures and MLFlow reporting on failed jobs are separate concerns." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Different issues in compilation/import paths: non-template-node compile error vs flash_attn import KeyError on Python 3.13." + "reason": "Template-node compilation errors and flash_attn import key errors are unrelated." }, { "left": "issue:44977", "right": "issue:46701", "accept": false, - "reason": "Different models and failures: Qwen3.5 flash-attention generation vs DiffusionGemma image-text pipeline." + "reason": "Qwen3.5 flash-attention generation issues and DiffusionGemma pipeline failure are different models and mechanisms." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing regression and TrainerState token-count tracking are unrelated." + "reason": "Tokenizer added-space behavior and TrainerState token counting are unrelated." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "JIT scripting GeneratorExp support vs SigLIP2 zero-probability outputs; no shared code bug." + "reason": "TorchScript GeneratorExp export failure and SIGLIP2 zero-probability outputs are unrelated." }, { - "left": "issue:44554", - "right": "issue:45446", + "left": "issue:43065", + "right": "issue:46369", "accept": false, - "reason": "MPS attention correctness with mismatched head dims vs incorrect PyTorch version check for AuxRequest import." + "reason": "Sam3PixelDecoder dummy convolution and sam3-litetext memory overflow are different issues." }, { - "left": "issue:43065", - "right": "issue:46369", + "left": "issue:44554", + "right": "issue:45446", "accept": false, - "reason": "Same model area, but one is a dummy Conv2d implementation detail and the other is a memory overflow report." + "reason": "MPS attention correctness and a PyTorch version guard for flex_attention imports are not the same bug." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Different serve crashes: multimodal .to() AttributeError vs KV-cache continuation indexing bug." + "reason": "Both are serve-related, but one crashes on multimodal tensor conversion and the other on KV-cache continuation indexing; distinct failures." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." }, { "left": "issue:41720", "right": "issue:46710", "accept": false, - "reason": "CUDA assert under auto device mapping vs post-upgrade accuracy regression; not the same underlying bug." + "reason": "CUDA assert under auto device mapping and an accuracy regression after upgrade are not the same underlying issue." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Different regressions: missing Gemma4Processor._tokenizer vs rope validation TypeError on JSON-loaded lists." - }, - { - "left": "issue:43493", - "right": "issue:43701", - "accept": false, - "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." + "reason": "Gemma4Processor missing _tokenizer and rope-parameter JSON list handling are unrelated." } ] }, @@ -8372,25 +8330,25 @@ "issue:43404|issue:46710", "issue:45072|issue:45198", "issue:37428|issue:43957", - "issue:45588|issue:45759", "issue:43295|issue:43450", + "issue:45588|issue:45759", "issue:43540|issue:45200", "issue:45200|issue:45823", - "issue:43295|issue:43873", "issue:44186|issue:46192", + "issue:43295|issue:43873", "issue:45823|issue:46192", "issue:43295|issue:44610", "issue:45520|issue:46178", "issue:43232|issue:44849", - "issue:43450|issue:43575", "issue:43572|issue:44987", + "issue:43450|issue:43575", "issue:43550|issue:43844", "issue:43572|issue:43957", "issue:44554|issue:45356", "issue:37428|issue:43824", "issue:42490|issue:43493", - "issue:44466|issue:45127", "issue:43653|issue:44112", + "issue:44466|issue:45127", "issue:38175|issue:43493", "issue:39692|issue:46192", "issue:43278|issue:43856", @@ -8400,184 +8358,214 @@ "issue:42915|issue:43824", "issue:43950|issue:46285" ], - "prepared_review_unit_hash": "1aa55f8fa433e29038ac8db32192702c289e57c3f029fa5dee8c33b9c7eef0ea", + "prepared_review_unit_hash": "7b35ca44567f1eee8f0fda422d471df6df07dbbf8104e50247e4ceb1bbefe7f1", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 37040, - "estimated_input_tokens": 9260, - "estimated_eval_tokens": 18776 + "serialized_chars": 37041, + "estimated_input_tokens": 9261, + "estimated_eval_tokens": 18778 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated issues across tokenization, model loading, multimodal processors, generation, dtype/device handling, and model-specific regressions. None of the soft edges look like safe duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue stands out; the reports do not share one underlying bug or fix path, so this is not a true duplicate cluster.", - "canonical_pr_reason": null, - "best_issue_reason": "If a representative issue is needed, issue 43950 is the strongest broad regression report (from_pretrained / buffer handling), but it only represents a small slice of this mixed set and should not be treated as the cluster\u2019s canonical target.", - "best_pr_reason": null, + "summary": "All proposed soft edges look like false positives: the items span unrelated tokenizer, model-loading, generation, quantization, multimodal-processing, and documentation bugs with no clear shared code path.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: every soft pair appears to describe a different bug or regression, so there is no duplicate cluster to collapse.", + "canonical_pr_reason": "No PRs present in the provided items.", + "best_issue_reason": "No best issue: none of the issues is clearly representative of a merged duplicate set.", + "best_pr_reason": "No PRs present in the provided items.", "soft_edge_verdicts": [ { "left": "issue:43404", "right": "issue:46710", "accept": false, - "reason": "Different bugs: tied lm_head weights in Mistral3 multimodal generation vs an accuracy regression in DeepSeek after a version upgrade." + "reason": "Different bugs and model families: Mistral3 tie-weight issue vs DeepSeek accuracy regression." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Unrelated failures: bfloat16 dtype mismatch in inference vs save_pretrained/tokenization failure in Wav2Vec2." + "reason": "CI dtype mismatch in inference vs Wav2Vec2 save/tokenization failure; unrelated symptoms." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Different code paths and symptoms: missing flash-attention import symbol vs meta-device loading breakage." + "reason": "Missing import symbol vs meta-device loading regression; different code paths." }, { - "left": "issue:45588", - "right": "issue:45759", + "left": "issue:43295", + "right": "issue:43450", "accept": false, - "reason": "Unrelated issues: flash_attention crash on s_aux=None vs AutoModelForCausalLM.from_config not unwrapping text_config." + "reason": "Processor.tokenizer regression with images vs batched video output shape bug." }, { - "left": "issue:43295", - "right": "issue:43450", + "left": "issue:45588", + "right": "issue:45759", "accept": false, - "reason": "Different APIs and bugs: processor.tokenizer/images regression vs video processor batched-shape bug." + "reason": "flash_attention crash on s_aux=None vs AutoModelForCausalLM config unwrapping; unrelated." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Different multimodal model issues: Qwen3OmniMoe video processing error vs Gemma 4 missing mm_token_type_ids default." + "reason": "Qwen3Omni video ValueError vs Gemma4 mm_token_type_ids defaulting issue." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Different failures in different models: token type ID defaulting vs device_map auto PLE device mismatch." + "reason": "Missing mm_token_type_ids vs device_map auto forward mismatch; different root causes." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:44186", + "right": "issue:46192", "accept": false, - "reason": "Different subsystems: custom model processor/tokenizer regression vs offloading/quantization behavior." + "reason": "Tokenizer crash on NER/batched padding vs return_tensors ignored with text_kwargs; different behaviors." }, { - "left": "issue:44186", - "right": "issue:46192", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "Different tokenizer problems: LayoutLMv2 NER/padding crash vs return_tensors being ignored with explicit text_kwargs." + "reason": "Processor/tokenizer bug vs quantization offloading failure; not the same bug." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Device placement bug in Gemma4 vs a tokenizer argument-handling bug; no shared underlying fix." + "reason": "Gemma4 device mismatch vs return_tensors handling; different APIs and failures." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Different multimodal processing bugs: processor/tokenizer regression vs OmDet-Turbo image-size mismatch." + "reason": "Images-to-tokenizer regression vs processor output size mismatch; unrelated preprocessing bugs." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Unrelated import/configuration issues: flash_attn key error on Python 3.13 vs BACKENDS_MAPPING rejecting tensorflow_text." + "reason": "flash_attn import KeyError on Python 3.13 vs BACKENDS_MAPPING error for tensorflow_text." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Different generation/model bugs: update_model_kwargs_for_generation after sync_gpus vs Qwen3.5 output_hidden_states regression." + "reason": "Generation kwarg update after sync_gpus vs Qwen3.5 hidden_states bug; different code paths." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in config vs model loading failure for a specific repo; unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape bug vs tensor-parallel OOM when loading a large model; not duplicates." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile+SDPA failure vs ZeRO-3 gradient increase; different training/runtime issues." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Config regression vs meta-device loading breakage; different underlying problems." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness bug vs Kimi-K2.5 tokenizer regression; unrelated." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different models; not the same bug." + "reason": "Flash-attention helper import error vs missing Qwen2_5_VL export; different missing symbols." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path import warning vs SigLIP2 parity discrepancy; unrelated problems." + "reason": "Local-path import warning vs SigLIP2 implementation discrepancy; unrelated." }, { - "left": "issue:44466", - "right": "issue:45127", + "left": "issue:43653", + "right": "issue:44112", "accept": false, - "reason": "Both touch weights, but one is serialization of tied lm_head.weight and the other is LoRA merge collapse with extended vocab; not the same code-path fix." + "reason": "BigBirdTokenizer special-token registration vs GraniteSpeech CI device override test." }, { - "left": "issue:43653", - "right": "issue:44112", + "left": "issue:44466", + "right": "issue:45127", "accept": false, - "reason": "Different issues: BigBirdTokenizer special-token registration vs a GraniteSpeech CI test flake." + "reason": "lm_head serialization inconsistency vs LoRA merge collapse after vocab extension; related theme but not the same code-path bug." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Related model family, but one is zero probabilities and the other is a broader HF-vs-JAX discrepancy; too little evidence they share one exact bug." + "reason": "Both are SigLIP2-related, but one reports zero probabilities and the other a JAX/HF discrepancy; not clearly the same bug." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "Documentation example errors vs tokenizer argument handling; unrelated." + "reason": "Documentation example errors vs return_tensors/text_kwargs handling; different issues." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Different regressions: embedding dtype changes during eval vs memory usage during Qwen3 MoE training." + "reason": "Embedding dtype drift on evaluate vs MoE memory inefficiency; unrelated." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Different problems: Qwen2.5-VL video position IDs vs FP8 inference failing on B200." + "reason": "qwen2.5-vl video position-id bug vs FP8 inference failure on B200; different models and hardware paths." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Same API surface, but different concrete bugs: all-zero assistant masks vs crash on tool-call messages with no content." + "reason": "assistant_masks all zero vs crash on tool-call messages; same API but distinct bugs." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing after added tokens vs embedding reinitialization in post_init; unrelated." + "reason": "Added-token spacing bug vs embedding reinitialization after resize; unrelated tokenizer/model-init issues." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Different models and failure modes: FineGrainedFP8Config load failure vs missing Qwen2.5-VL import symbol." + "reason": "FineGrainedFP8Config failure on Qwen3Moe vs Qwen2_5_VL import error; unrelated." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Both involve loading/model structure, but one is non-persistent buffer corruption and the other is CLIP text model structure inconsistency; not the same fix." + "reason": "Non-persistent buffer corruption in from_pretrained vs CLIPTextModel structure inconsistency; different loading/model-shape bugs." } ] }, @@ -8657,41 +8645,41 @@ "issue:44442|issue:45198", "issue:44792|issue:45563", "issue:43819|issue:44112", - "issue:44898|issue:45412", "issue:44479|issue:46192", + "issue:44898|issue:45412", "issue:43827|issue:43901", "issue:45200|issue:45325", "issue:44561|issue:46234", "issue:44610|issue:44857", "issue:43388|issue:43526", "issue:43335|issue:43550", - "issue:41720|issue:46209", "issue:43819|issue:45083", + "issue:41720|issue:46209", "issue:44857|issue:45593", - "issue:45759|issue:46285", "issue:44206|issue:45584", + "issue:45759|issue:46285", "issue:45520|issue:46170", "issue:43606|issue:45083", "issue:38175|issue:45412", "issue:43335|issue:43653", "issue:29942|issue:32090", - "issue:43550|issue:44112", - "issue:37428|issue:43425" + "issue:37428|issue:43425", + "issue:43550|issue:44112" ], - "prepared_review_unit_hash": "e78e4aa510670d96b9df0b8dddeaae4cdb29a96bf2f6124aaa46e40e1a9f051f", + "prepared_review_unit_hash": "b3cc25802611c3dc09ae59ecf397499a8a400ff4d512db57b698147ddd0cb765", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35516, + "serialized_chars": 35515, "estimated_input_tokens": 8879, "estimated_eval_tokens": 18014 }, @@ -8700,174 +8688,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous. The soft pairs mostly share superficial keywords (models, tokenizers, configs, CI, flash attention) but not the same concrete bug or code path, so none should be merged as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 43824 is the strongest representative here: it reports a concrete import regression, has the most discussion/inbound references among the set, and describes a clear user-facing failure.", + "summary": "This cluster is highly heterogeneous: the soft-similarity pairs mostly share broad themes like import errors, model-specific regressions, or processor/config issues, but they do not look like the same underlying bug or change. I found no safe PR duplicate merge here; the strongest issue anchor is a highly discussed, concrete bug report, but it is not representative of the whole set.", + "confidence": 0.39, + "canonical_issue_reason": "issue:43824 has the strongest signal in the cluster (high discussion/inbound refs) and a concrete, reproducible import failure, making it the best single issue anchor despite the cluster\u2019s overall heterogeneity.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43824 is the best overall issue candidate in this heterogeneous cluster because it is specific, reproducible, and high-signal compared with the other more disparate reports.", + "best_issue_reason": "issue:43824 is the clearest, most actionable issue in the set; it has strong engagement and a specific failure mode, while most other items are unrelated one-off regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43452", "right": "issue:46512", "accept": false, - "reason": "Different problems: gguf / from_pretrained loading failure versus stale config.dtype after from_config(dtype=...)." + "reason": "Different problems: gguf/from_pretrained loading breakage vs config.dtype not being updated by from_config." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Different models and failures: missing Qwen2_5_VL symbol import versus DeepSeek detokenization regression." + "reason": "Both involve Qwen-related behavior, but one is an import failure and the other is a detokenization bug in a different path." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Same symptom class (missing pad_token_id) but different model families and likely different fixes; not the same underlying bug." + "reason": "Same symptom class, but different model configs and likely separate missing-attribute fixes." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Different modalities and code paths: audio chat-template sampling rate versus batched video processor output shape." + "reason": "Audio chat-template sampling-rate defaulting and batched video shape handling are unrelated code paths." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer-related, but one is FastSpeech2Conformer loading and the other is Wav2Vec2 save/tokenization behavior." + "reason": "Both are speech/audio tokenizer issues, but they affect different models and different failure points." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Different generate-related issues: Janus image generation test failure versus stale warning for num_return_sequences." + "reason": "Janus image-generation test failure and a stale generation warning are not the same bug." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "Unrelated: DAC latent/STE mismatch versus GraniteSpeech CI device-override test flakiness." + "reason": "DAC latent mismatch vs GraniteSpeech CI device override are unrelated." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:44479", + "right": "issue:46192", "accept": false, - "reason": "Different bugs: Perceiver non-default resolution failure versus RT-DETR memory not being released on deletion." + "reason": "Video-input regression in specific Qwen models vs return_tensors being ignored in text_kwargs are different bugs." }, { - "left": "issue:44479", - "right": "issue:46192", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "Video input regression for Qwen VL models versus return_tensors being ignored when text_kwargs is passed." + "reason": "Perceiver resolution failure and RT-DETR memory release are unrelated." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they concern different APIs and different stale references." + "reason": "Both are docs-related, but they reference different deprecated/changed behaviors and are not the same change." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Different Qwen2.5-VL issues: missing/default mm_token_type_ids versus incorrect rope position scaling." + "reason": "Gemma multimodal token ids defaulting and Qwen2.5-VL rope indexing are separate model-specific regressions." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both involve missing import_utils helpers, but they are different functions and separate regressions." + "reason": "Both mention missing import_utils helpers, but they are different helper removals with different breakages." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Different components: OmDet-Turbo preprocessing size mismatch versus LwDetrImageLoss AMP crash." + "reason": "Processor output size mismatch and AMP crash in a loss are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Both mention label handling, but one is gather_for_metrics batch truncation and the other is BeitImageProcessorFast reduce_labels behavior." + "reason": "Both concern labels, but one is gather_for_metrics batching and the other is BeitImageProcessorFast label reduction." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Config construction bug versus torch.compile/SDPA runtime failure; not the same underlying issue." + "reason": "SwitchTransformers sparse-layer construction and Bamba torch.compile SDPA failure are different code paths." }, { - "left": "issue:41720", - "right": "issue:46209", + "left": "issue:43819", + "right": "issue:45083", "accept": false, - "reason": "Different hardware/model failures: Qwen3 auto device mapping on A800 versus FP8 inference on B200." + "reason": "DAC latent mismatch and qwen3_omni_moe length helper behavior are unrelated." }, { - "left": "issue:43819", - "right": "issue:45083", + "left": "issue:41720", + "right": "issue:46209", "accept": false, - "reason": "Different helper/code paths: DAC latent mismatch versus qwen3_omni_moe feature-length calculation behavior." + "reason": "Both are GPU inference failures, but on different models, hardware, and likely different root causes." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Different training-loss bugs: AMP/CUDA crash versus auxiliary losses not used when denoising is off." + "reason": "Loss AMP crash and missing auxiliary losses when denoising is off are separate training issues." }, { - "left": "issue:45759", - "right": "issue:46285", + "left": "issue:44206", + "right": "issue:45584", "accept": false, - "reason": "Different multimodal model-structure problems: Qwen composite config unwrapping versus CLIP text-model structure inconsistency." + "reason": "Unsupported feature-extractor arg and Whisper empty-transcription handling are unrelated." }, { - "left": "issue:44206", - "right": "issue:45584", + "left": "issue:45759", + "right": "issue:46285", "accept": false, - "reason": "Different failures: unsupported center arg in a feature extractor versus Whisper empty-transcription generation after token alignment." + "reason": "Composite Qwen config unwrapping and CLIP text-model structure inconsistency are distinct model-loading issues." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Separate platform/import problems: flash_attn lookup on Python 3.13 versus Windows ROCm torch.distributed dependency issues." + "reason": "Python 3.13 flash_attn import key error and Windows ROCm torch.distributed dependency failure are unrelated platform issues." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "Unrelated: CPU-offload device mismatch in bark-small versus qwen3_omni_moe output-length helper behavior." + "reason": "CPU offload device mismatch and qwen3_omni_moe feature length calculation are not the same bug." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Different bugs: Siglip2 zero probabilities versus RT-DETR memory retention after deletion." + "reason": "Zero probabilities in SigLIP2 and RT-DETR memory retention are unrelated." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Different subsystems: SwitchTransformers sparse-layer config bug versus BigBirdTokenizer special-token registration." + "reason": "Sparse-layer config bug and tokenizer special-token registration bug are different subsystems." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Different failures: Flash Attention 2 test breakage versus Trainer _gpu_broadcast_one NoneType type error." + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are unrelated." }, { - "left": "issue:43550", - "right": "issue:44112", + "left": "issue:37428", + "right": "issue:43425", "accept": false, - "reason": "Different scope: Bamba torch.compile/SDPA failure versus a stale GraniteSpeech CI test." + "reason": "Flash-attention import error and generic Torch 2.10 incompatibility are related only at a broad dependency level, not the same concrete fix." }, { - "left": "issue:37428", - "right": "issue:43425", + "left": "issue:43550", + "right": "issue:44112", "accept": false, - "reason": "Different import/version problems: missing _flash_supports_window_size versus generic Torch 2.10 incompatibility." + "reason": "Bamba torch.compile SDPA failure and GraniteSpeech CI stale test are unrelated." } ] }, @@ -8943,227 +8931,227 @@ "issue:28282|issue:33357", "issue:43611|issue:43931", "issue:44448|issue:44743", - "issue:45759|issue:46512", "issue:43335|issue:43504", + "issue:45759|issue:46512", "issue:44855|issue:45083", "issue:43720|issue:44112", "issue:43278|issue:43825", "issue:29127|issue:32090", "issue:43335|issue:45083", "issue:43335|issue:44112", - "issue:37428|issue:44351", "issue:43335|issue:45593", + "issue:37428|issue:44351", "issue:44514|issue:44521", - "issue:45588|issue:45859", "issue:34689|issue:35141", + "issue:45588|issue:45859", "issue:45127|issue:45245", - "issue:46431|issue:46566", "issue:45290|issue:46326", + "issue:46431|issue:46566", "issue:42490|issue:43854", "issue:45200|issue:45588", "issue:45310|issue:45478", - "issue:43819|issue:44493", - "issue:44448|issue:44779", "issue:39692|issue:43526", + "issue:44448|issue:44779", + "issue:43606|issue:44857", "issue:45459|issue:46234", - "issue:43606|issue:44857" + "issue:43819|issue:44493" ], - "prepared_review_unit_hash": "23e51a7a4a4e588a8aa420846908734ba653a1868c7313fc4be96c179a773747", + "prepared_review_unit_hash": "bbf8110eabb273742de19bd6d339641a3372f4eb85dfad4a80cd846dece9b1ad", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35629, - "estimated_input_tokens": 8908, - "estimated_eval_tokens": 18072 + "serialized_chars": 35623, + "estimated_input_tokens": 8906, + "estimated_eval_tokens": 18068 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issue reports with similar wording. I found one clear duplicate pair: the two identical Qwen3.5 MoE `from_pretrained` errors. No pull requests are present.", + "summary": "This cluster is mostly a set of unrelated issue reports with one clear duplicate pair: the Qwen3.5 MoE from_pretrained error. All other soft pairs appear to describe different bugs or regressions, even when they touch nearby code paths.", "confidence": 0.93, - "canonical_issue_reason": "issue:45310 is the earliest exact-match report for the Qwen3.5 MoE `from_pretrained` failure; issue:45478 is a later duplicate of the same bug.", + "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is the earlier, more specific report for the Qwen3.5 MoE `from_pretrained` failure and matches 45478 exactly.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the cleanest duplicate anchor in the set because it exactly matches issue:45478 and is the earlier report.", + "best_issue_reason": "Issue 45310 is the strongest representative because it is the exact duplicate target of 45478 and has the clearest, most direct problem statement.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Different failures: missing PyTorch ImportError vs MacOS bus error on a community CLIP model." + "reason": "Both are loading/runtime failures, but one is a missing PyTorch import and the other is a MacOS bus error with a CLIP model; different underlying bugs." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "Both are model-loading problems, but they involve different root causes and models." + "reason": "Both are model-loading regressions, but they involve different failure modes and different affected models (`base_model_prefix` vs Qwen3-VL weight shape mismatch)." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus output drift vs Qwen3.5 recurrent-state reset are different code paths and bugs." + "reason": "One is an output difference for Pegasus between v4 and v5; the other is recurrent states resetting in modular_qwen3_5 with cache. Not the same bug." }, { - "left": "issue:45759", - "right": "issue:46512", + "left": "issue:43335", + "right": "issue:43504", "accept": false, - "reason": "Different issues: missing `text_config` unwrapping vs stale `config.dtype` after `from_config(dtype=...)`." + "reason": "SwitchTransformers config creating sparse layers and BEiT pretrained preset loading are unrelated issues in different model families." }, { - "left": "issue:43335", - "right": "issue:43504", + "left": "issue:45759", + "right": "issue:46512", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and BEiT legacy-field loading are unrelated." + "reason": "Both concern `from_config`, but one is about unwrapping `text_config` in composite multimodal configs while the other is about stale `config.dtype` after dtype-aware initialization." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 import parsing error vs helper output-length behavior in `qwen3_omni_moe`." + "reason": "Python 3.13 import parsing failure in DebertaV2 is unrelated to a helper returning unexpected lengths in qwen3_omni_moe." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Both are CI-related, but one is BitNet packed-weight loading and the other is a stale GraniteSpeech device override test." + "reason": "Both mention CI/device issues, but one is packed-weight unpacking during accelerate loading and the other is a stale device override test in GraniteSpeech." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype mismatch during eval is unrelated to a pipeline error message about translation tasks." + "reason": "Embedding dtype drift during evaluate is unrelated to a pipeline error message about translation tasks in v5." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "LayoutLMv3 box-validation messaging and `_gpu_broadcast_one` NoneType handling are different bugs." + "reason": "LayoutLMv3 box-validation messaging and Trainer `_gpu_broadcast_one` NoneType handling are distinct error-handling bugs." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "Config creating a sparse layer and a helper length-calculation issue are not the same underlying defect." + "reason": "Different model components and symptoms: sparse-layer config bug vs feature-extraction length helper behavior." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "Different model areas and failure modes; one is config logic, the other is a CI regression." + "reason": "Sparse-layer creation in SwitchTransformers and GraniteSpeech CI device override are unrelated." }, { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:43335", + "right": "issue:45593", "accept": false, - "reason": "Both are import-related, but one is a missing internal symbol and the other is `HybridCache` export/import availability." + "reason": "SwitchTransformers config bug and D-FINE auxiliary-loss behavior are unrelated model-specific issues." }, { - "left": "issue:43335", - "right": "issue:45593", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "Sparse-layer config bug and missing auxiliary losses in D-FINE are unrelated." + "reason": "Both are import errors, but they concern different symbols and different breakages (`_flash_supports_window_size` vs `HybridCache`)." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both involve chat templates, but one is batched padding crash and the other is zero assistant masks for multimodal inputs." + "reason": "Both are Qwen2.5-VL chat-template issues, but one is a batching crash with `padding=False` while the other is incorrect all-zero assistant masks for multimodal inputs." }, { - "left": "issue:45588", - "right": "issue:45859", + "left": "issue:34689", + "right": "issue:35141", "accept": false, - "reason": "Flash-attention `s_aux=None` crash and Qwen3_5 rotary embedding CPU-offload incompatibility are separate defects." + "reason": "Llama 3.2 Vision model loading regression and output embedding reinitialization after resizing are different problems." }, { - "left": "issue:34689", - "right": "issue:35141", + "left": "issue:45588", + "right": "issue:45859", "accept": false, - "reason": "Model loading breakage for Llama 3.2 Vision and embedding reinitialization after resizing are different bugs." + "reason": "Both involve offload/flash-related behavior, but one is an `s_aux=None` crash in flash attention integration and the other is a CPU-offload incompatibility in rotary embedding forward." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge collapse with extended vocab is unrelated to the 2^24 category limit runtime error." + "reason": "LoRA merge collapse with extended vocab is unrelated to the category-count ceiling error." }, { - "left": "issue:46431", - "right": "issue:46566", + "left": "issue:45290", + "right": "issue:46326", "accept": false, - "reason": "Missing chat template support in a processor and DiffusionGemma weights staying on meta are unrelated." + "reason": "Both touch chat templating, but one crashes on assistant tool-call messages without content and the other is a `continue_final_message` bug for mistral_common tokenizers." }, { - "left": "issue:45290", - "right": "issue:46326", + "left": "issue:46431", + "right": "issue:46566", "accept": false, - "reason": "Tool-call assistant-content crash and `continue_final_message` behavior for mistral_common are distinct chat-template issues." + "reason": "Missing chat template support in `apply_chat_template` is unrelated to DiffusionGemma weights staying on meta with auto device map." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "A local-path trailing-slash warning is not the same as a GLM-4.7 loading failure in unit tests." + "reason": "A trailing-slash local-path import warning is not the same as a GLM-4.7-Flash model-loading failure in unit tests." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 token-type defaults and flash-attention `s_aux` handling are unrelated." + "reason": "Gemma 4 `mm_token_type_ids` defaulting and flash attention `s_aux=None` are unrelated issues." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Exact same title and same Qwen3.5 MoE `from_pretrained` failure; the later issue is a duplicate of the earlier report." + "reason": "Exact same Qwen3.5 MoE `from_pretrained` error on the same version range; 45478 is a clear duplicate of 45310." }, { - "left": "issue:43819", - "right": "issue:44493", + "left": "issue:39692", + "right": "issue:43526", "accept": false, - "reason": "DAC `from_latents` mismatch and unexpected position-id keys are different model behaviors." + "reason": "SigLIP2 documentation example errors and BEiT Fast image processor label reduction are different concerns." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus v4/v5 output drift and DeepSeek tokenizer regression are different components." + "reason": "Pegasus output differences and DeepSeek tokenizer regressions are separate model/tokenizer bugs." }, { - "left": "issue:39692", - "right": "issue:43526", + "left": "issue:43606", + "right": "issue:44857", "accept": false, - "reason": "SigLIP2 documentation/example errors and BEiT `reduce_labels` bug do not share the same underlying issue." + "reason": "CPU offload device mismatch in Bark-small is unrelated to mixed-precision CUDA crash in LwDetrImageLoss." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer decode-error masking and missing `is_falcon_mamba_ssm_available` export are unrelated." + "reason": "Tokenizer protobuf error masking and a missing `is_falcon_mamba_ssm_available` utility function are unrelated." }, { - "left": "issue:43606", - "right": "issue:44857", + "left": "issue:43819", + "right": "issue:44493", "accept": false, - "reason": "CPU offload device mismatch in Bark and float16 AMP crash in LwDetr are different failures." + "reason": "DAC `from_latents` mismatch due to missing STE and unexpected position-id keys are different defects." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: all items are issues, and the only clear duplicate pair in the provided set is issue:45310 / issue:45478, which do share the same title and bug description. The soft-edge rejections are conservative overall and do not overclaim mergeability or duplication." + "feedback": "The cluster summary is grounded in the packet, and the soft-edge verdicts are conservative. The only accepted pair, issue:45310 vs issue:45478, is strongly supported by the identical titles and close timing; the remaining pairs are reasonably treated as different bugs or regressions." }, "evaluator_used": true, "retried": false, @@ -9192,6 +9180,7 @@ "issue:29942", "issue:31501", "issue:32090", + "issue:33290", "issue:33357", "issue:34689", "issue:35141", @@ -9215,7 +9204,6 @@ "issue:44493", "issue:44561", "issue:44617", - "issue:44779", "issue:44855", "issue:44857", "issue:44977", @@ -9240,10 +9228,10 @@ "issue:44617|issue:46697", "issue:44206|issue:45823", "issue:43701|issue:45083", - "issue:44857|issue:45072", - "issue:44360|issue:44561", "issue:32090|issue:33357", + "issue:44857|issue:45072", "issue:42490|issue:43575", + "issue:44360|issue:44561", "issue:38175|issue:44493", "issue:43550|issue:45083", "issue:29942|issue:35141", @@ -9261,204 +9249,216 @@ "issue:44857|issue:45588", "issue:45290|issue:46655", "issue:37428|issue:42175", - "issue:44561|issue:44779" + "issue:32090|issue:33290" ], - "prepared_review_unit_hash": "e68db3c7b4376e35a30159504df806f863f970bc3d849c5267b051698cc13bff", + "prepared_review_unit_hash": "57c08377205bdc8484916ba51a2dce8b43645d1c2f4e413ecd4777ce286569b1", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36133, - "estimated_input_tokens": 9034, - "estimated_eval_tokens": 18324 + "serialized_chars": 36110, + "estimated_input_tokens": 9028, + "estimated_eval_tokens": 18312 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports that only share weak lexical similarity (training, tokenizer, FlashAttention, save/load, etc.). I found no pair that clearly describes the same underlying bug or fix, so no soft-edge duplicates are accepted.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45920 is the broadest, highest-impact regression report in the set, with multiple affected models and broad tokenizer fallout; it is the least niche representative issue here.", + "summary": "These issues are mostly unrelated across different models, processors, training flows, and regressions. The soft-similarity links are superficial, so none of the candidate pairs should be merged.", + "confidence": 0.96, + "canonical_issue_reason": "No single issue cleanly represents the cluster: it mixes unrelated bugs in training, tokenization, Flash Attention, model loading, serialization, and image/video processors.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the strongest standalone issue to anchor the cluster because it has broad scope, clear user impact, and relatively high activity.", + "best_issue_reason": "No best issue among these; the cluster is too heterogeneous to have a meaningful canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "Both are training-startup failures, but the described failure modes are different and not clearly the same bug." + "reason": "Both are training-start failures, but one is a resume_from_checkpoint epoch-iterator problem and the other is a FineWeb SLM startup issue; different code paths and root causes." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing behavior and Trainer GPU broadcast NoneType errors are unrelated code paths." + "reason": "Tokenizer spacing after an added token is unrelated to the Trainer _gpu_broadcast_one NoneType error." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "One is a SigLIP2 docs/example problem; the other is a position-id regression across models." + "reason": "A SigLIP2 docs/example mismatch plus quantization failure is not the same bug as a position-id warning across many models." }, { "left": "issue:44617", "right": "issue:46697", "accept": false, - "reason": "CUDA OOM for Sam3Video and over-padding in Swin2SR are different model/processor bugs." + "reason": "Sam3Video CUDA OOM and Swin2SR over-padding are different models and different failure modes." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "Unsupported feature-extractor arg and Gemma4 device mismatch are distinct failures." + "reason": "Unsupported center argument in LasrFeatureExtractor is unrelated to Gemma4 device-map forward mismatches." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint key mismatch and qwen3_omni_moe feature-length behavior are unrelated." + "reason": "A checkpoint key mismatch is not the same as unexpected feature-extraction length behavior in qwen3_omni_moe." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast of NoneType and MacOS bus error in a CLIP model are distinct runtime problems." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Both mention dtype-related crashes, but they affect different modules and concrete failure paths." + "reason": "float16 AMP crash in LwDetrImageLoss and dtype mismatches in SwitchTransformers/TimmWrapperModel are separate issues." }, { - "left": "issue:44360", - "right": "issue:44561", - "accept": false, - "reason": "A missing ReLU in the DSA indexer and a removed utility breaking trust_remote_code are unrelated." - }, - { - "left": "issue:32090", - "right": "issue:33357", + "left": "issue:42490", + "right": "issue:43575", "accept": false, - "reason": "Trainer broadcast NoneType and MacOS bus error on CLIP are different runtime issues." + "reason": "A local-path import warning is unrelated to an out-of-memory failure when loading a very large Qwen2 model with tensor parallelism." }, { - "left": "issue:42490", - "right": "issue:43575", + "left": "issue:44360", + "right": "issue:44561", "accept": false, - "reason": "Local-path import warning and tensor-parallel OOM are not the same bug." + "reason": "Missing ReLU in the DSA indexer and removal of is_torch_fx_available breaking trust_remote_code models are unrelated changes." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and position-id key warnings are unrelated." + "reason": "Unexpected zero probabilities in siglip2-base-patch16-224 is not the same as a generic position-id key regression." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA failure and feature-extractor output-length behavior do not describe the same code defect." + "reason": "Bamba-9B-v2 torch.compile/SDPA failure and qwen3_omni_moe feature-length behavior involve different subsystems and bugs." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures and embedding reinitialization after resize are unrelated." + "reason": "Flash Attention 2 test failures are unrelated to output embeddings being reinitialized during post_init after resize_token_embeddings." }, { "left": "issue:27957", "right": "issue:34689", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and Llama 3.2 Vision model loading are different issues." + "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 90B Vision loading regression are different concrete issues." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "A dtype drift between train/eval and an undefined video-token helper are different bugs." + "reason": "A BF16-to-FP32 dtype change in evaluation is unrelated to undefined variables in the video branch of multimodal token counting." }, { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "Tokenizer ID regression and return_tensors being ignored are separate API problems." + "reason": "Wrong token IDs across multiple tokenizers is a different tokenizer bug than silently ignoring return_tensors when text_kwargs is passed." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "A v4-v5 output difference for Pegasus and a Qwen3.5 flash-attention generation issue are unrelated." + "reason": "Pegasus v4/v5 output differences and Qwen3.5 flash-attention generation errors are unrelated model-specific regressions." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Qwen2.5-VL iterable-dataset multi-image finetuning and BeitImageProcessorFast label reduction are different failures." + "reason": "IterableDataset finetuning with multiple images and BeitImageProcessorFast reduce_labels returning one label are distinct bugs." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat-template tool-call handling and qwen2.5-vl video position IDs are separate multimodal issues." + "reason": "Chat-template tool-call crashes and Qwen2.5-VL video vision_position_ids issues are different code paths." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Eval-mode gradient checkpointing and train/eval dtype changes are not the same underlying problem." + "reason": "Both mention eval-mode behavior, but BF16/FP32 drift is not the same as gradient checkpointing being disallowed in eval mode." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Flash-attention generation behavior and Gemma4 multi-GPU OOM are different symptoms and code paths." + "reason": "Flash-attention generation failure in Qwen3.5 is unrelated to Gemma4 multi-GPU CUDA OOM." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "German DistilBertModel issue and ViT torch.jit generator-expression export failure are unrelated." + "reason": "German DistilBertModel issue and GeneratorExp/JIT export failure are separate, old, unrelated problems." }, { "left": "issue:44855", "right": "issue:46399", "accept": false, - "reason": "Python 3.13 decorator/comment parsing and missing flash-kernel decorator registration are distinct import/runtime issues." + "reason": "Python 3.13 IndentationError/JIT parsing failure is unrelated to apply_rotary_pos_emb registration errors in Qwen3.5." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Custom token IDs lost on save/load and wrong custom module loading are both serialization-related, but not the same bug." + "reason": "Losing custom token IDs after save/load and loading the wrong custom module after save_pretrained are both persistence-related, but they are different mechanisms and bugs." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP float16 crash in LwDetrImageLoss and flash_attention s_aux=None crash are different failures." + "reason": "AMP crashes in LwDetrImageLoss and flash_attention.py AttributeError on s_aux=None are different failures." }, { - "left": "issue:45290", + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "A missing internal flash-attention import and missing TensorFlow backend in a torch install are unrelated packaging/import issues." + }, + { + "left": "issue:43329", "right": "issue:46655", "accept": false, - "reason": "Tool-call chat templating and video padding in SmolVLM are unrelated." + "reason": "Undefined helper variables in multimodal token counting are unrelated to incorrect padding calculation in SmolVLM's video processor." }, { - "left": "issue:37428", - "right": "issue:42175", + "left": "issue:44493", + "right": "issue:44561", "accept": false, - "reason": "Missing flash-attention helper import and missing TensorFlow backend installation are unrelated packaging/import issues." + "reason": "Unexpected position-id keys and removal of is_torch_fx_available breaking trust_remote_code models are not the same regression." }, { - "left": "issue:44561", - "right": "issue:44779", + "left": "issue:38175", + "right": "issue:46192", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to return_tensors being ignored when text_kwargs is explicitly passed." + }, + { + "left": "issue:43381", + "right": "issue:45823", "accept": false, - "reason": "Removal of is_torch_fx_available and DeepSeek tokenizer regressions are different root causes." + "reason": "Gradient checkpointing in eval mode and Gemma4 PLE device mismatch are different problems." } ] }, @@ -9487,7 +9487,6 @@ "issue:29942", "issue:30064", "issue:32090", - "issue:33290", "issue:33357", "issue:34567", "issue:39692", @@ -9512,8 +9511,10 @@ "issue:44448", "issue:44466", "issue:44479", + "issue:44561", "issue:44568", "issue:44743", + "issue:44779", "issue:44871", "issue:44977", "issue:45198", @@ -9528,24 +9529,23 @@ "issue:45712", "issue:45803", "issue:45823", - "issue:46121", - "issue:46693" + "issue:46121" ], "soft_pairs": [ - "issue:32090|issue:33290", - "issue:43873|issue:45229", "issue:25251|issue:32090", + "issue:44561|issue:44779", + "issue:43873|issue:45229", "issue:44448|issue:44466", - "issue:43976|issue:45440", "issue:42490|issue:43873", - "issue:45588|issue:45823", + "issue:43976|issue:45440", "issue:43278|issue:43575", + "issue:45588|issue:45823", "issue:41720|issue:45229", "issue:33357|issue:39692", - "issue:44568|issue:45491", "issue:43381|issue:43701", - "issue:44977|issue:45563", + "issue:44568|issue:45491", "issue:29127|issue:34567", + "issue:44977|issue:45563", "issue:44871|issue:44977", "issue:45198|issue:45803", "issue:45372|issue:45588", @@ -9556,229 +9556,214 @@ "issue:42915|issue:42994", "issue:43412|issue:45588", "issue:45372|issue:45406", - "issue:45459|issue:45712", "issue:29942|issue:30064", - "issue:43756|issue:45440", + "issue:45459|issue:45712", "issue:43653|issue:44442", - "issue:44206|issue:44479", - "issue:45491|issue:46693", - "issue:45372|issue:45459", - "issue:29127|issue:33357" + "issue:43756|issue:45440", + "issue:44206|issue:44479" ], - "prepared_review_unit_hash": "850ae6e54210f1b6256e4302e547aea3b46dc0faaa53fe696e81a480756d4bc4", + "prepared_review_unit_hash": "78a8056c17bf7a6b78895f2b31cd141c47b5138a42565e37266b2f72338fa992", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37417, - "estimated_input_tokens": 9355, - "estimated_eval_tokens": 18966 + "soft_pair_count": 29, + "serialized_chars": 36545, + "estimated_input_tokens": 9137, + "estimated_eval_tokens": 18530 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the titles point to unrelated bugs across tokenizers, Gemma/Qwen, quantization, audio/vision, and trainer internals. None of the soft pairs look like the same underlying bug or change; the apparent similarity is mostly broad subsystem overlap.", - "confidence": 0.69, - "canonical_issue_reason": "No strong duplicate nucleus is visible from the titles alone. If one issue must anchor the cluster, issue 43873 is the broadest and most active report here (quantization/offloading, open, with high discussion/inbound refs), so it is the most plausible canonical issue.", + "summary": "This set looks like a collection of unrelated Transformers bug reports across tokenization, model loading, training, inference, quantization, and docs. The suggested soft pairs share superficial wording only; none appear to describe the same concrete underlying bug or merge cleanly into one fix.", + "confidence": 0.92, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the best representative issue because it is broad, currently open, and has the most evidence of being a live, discussion-heavy bug rather than a narrow one-off report.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:32090", - "right": "issue:33290", + "left": "issue:25251", + "right": "issue:32090", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error and Adafactor+DeepSpeed OOM are different failure modes with no concrete shared code path." + "reason": "Both mention runtime/output problems, but one is about pipeline top_k changing list nesting and the other is a Trainer NoneType broadcast error. Different code paths and failure modes." }, { - "left": "issue:43873", - "right": "issue:45229", + "left": "issue:44561", + "right": "issue:44779", "accept": false, - "reason": "Both mention quantization/memory, but one is offloading behavior and the other is Gemma4 multi-GPU CUDA OOM; different bug surfaces." + "reason": "Both are v5 regressions involving remote or tokenizer-related breakage, but one is about removal of is_torch_fx_available and the other is a DeepSeek tokenizer correctness issue. Not the same bug." }, { - "left": "issue:25251", - "right": "issue:32090", + "left": "issue:43873", + "right": "issue:45229", "accept": false, - "reason": "Pipeline top_k nesting output change is unrelated to Trainer broadcast errors." + "reason": "Both involve memory/offloading behavior, but one is offloading with quantization generally and the other is a Gemma4 31B multi-GPU CUDA OOM. Different models and likely different root causes." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Both are v5 output/serialization regressions, but one is Pegasus generation output drift and the other is lm_head.weight serialization; not the same code-path bug." - }, - { - "left": "issue:43976", - "right": "issue:45440", - "accept": false, - "reason": "Python version compatibility and DeepSeekV3 implementation divergence are unrelated." + "reason": "Both reference v5 behavior changes, but one is a generation/output difference for pegasus and the other is inconsistent lm_head.weight serialization. Distinct bugs." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Trailing-slash import warning is a loading-path issue, not the quantization/offloading bug described in 43873." + "reason": "A local-path import warning is unrelated to quantization/offloading behavior. No shared underlying defect." }, { - "left": "issue:45588", - "right": "issue:45823", + "left": "issue:43976", + "right": "issue:45440", "accept": false, - "reason": "Flash-attention s_aux=None crash and Gemma4 PLE device mismatch are separate runtime failures." + "reason": "One is a Python version compatibility regression, the other is a DeepseekV3MoE implementation divergence. Different layers of the stack." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Embedding dtype drift in eval and Qwen2-57B-A14B-Instruct tensor-parallel OOM are different problems." + "reason": "Both concern dtype/OOM-like model loading behavior, but one is an embedding dtype mismatch during eval and the other is tensor-parallel OOM for a specific Qwen2 model. Not the same bug." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Both mention Gemma4/processor-related runtime issues, but one is a flash_attention crash on s_aux=None and the other is a PLE device mismatch during forward. Different failure points." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Different models and symptoms: Qwen3 auto device mapping cuda assert vs Gemma4 multi-GPU OOM." + "reason": "Both mention CUDA/device-map failures, but one is a Qwen3 auto device mapping cudaErrorAssert and the other is Gemma4 31B multi-GPU OOM. Different symptoms and likely causes." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS bus error on CLIP is not the same bug as SigLIP2 documentation example errors." - }, - { - "left": "issue:44568", - "right": "issue:45491", - "accept": false, - "reason": "mdeberta add_special_tokens behavior and Gemma3 NaN embeddings are unrelated tokenizer vs attention/embedding bugs." + "reason": "One is a MacOS bus error for a community CLIP model; the other is a SigLIP2 docs example with model/processor mismatch and quantization failure. Not the same bug." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch do not share the same underlying defect." + "reason": "Both involve checkpoint/resume/state handling, but one is about gradient checkpointing in eval mode and the other is a resume_from_checkpoint key mismatch. Different mechanisms." }, { - "left": "issue:44977", - "right": "issue:45563", + "left": "issue:44568", + "right": "issue:45491", "accept": false, - "reason": "Qwen3.5 flash-attention generation regression is unrelated to a stale warning for num_return_sequences." + "reason": "Both involve tokenizer/model configuration issues, but one is missing BOS/EOS addition and the other is NaN embeddings from sliding-window attention with mixed-length batches. Different code paths." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error-message clarity and TrainerState num_input_tokens_seen updates are unrelated trainer/layout issues." + "reason": "One is a LayoutLMv3 error-message clarity issue; the other is TrainerState num_input_tokens_seen not updating. No shared bug." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "One is a Qwen3.5 flash-attention generation regression; the other is a stale warning for paged generate num_return_sequences. Different behavior and fix surface." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id config mismatch and Qwen3.5 flash-attention generation failure are different bugs." + "reason": "One is an eos_token_id configuration mismatch in Gemma-3; the other is a Qwen3.5 flash-attention generation issue. Unrelated." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and Fuyu bare-except error handling are distinct tokenizer/model issues." + "reason": "Both are save/convert-related failures, but one is Wav2Vec2 save_pretrained/tokenization and the other is a bare except in FuyuBatchFeature.convert_to_tensors(). Different components and bug classes." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import failure and flash_attention.py s_aux=None crash are both Gemma4-adjacent but target different layers and symptoms." + "reason": "Both touch Gemma4 processing, but one is an ImportError from mistral_common while loading the processor and the other is a flash_attention AttributeError on s_aux=None. Separate defects." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated." + "reason": "Both involve model/tokenizer implementation correctness, but one is BigBirdTokenizer special-token registration and the other is Qwen3.5 recurrent state reset with cache. Different subsystems." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/padding and Wav2Vec2 save/tokenization failure are different tokenizer paths." + "reason": "Both are tokenizer/input handling bugs, but LayoutLMv2 NER padding/truncation crashes are unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError on tasks is not the same failure as Wav2Vec2 save_pretrained/tokenization issues." + "reason": "MLukeTokenizer AttributeError on tasks is a tokenizer runtime bug; Wav2Vec2 save_pretrained/tokenization failure is a different model and failure mode." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "base_model_prefix loading break and convert_rope_params_to_dict TypeError are unrelated API/validation bugs." + "reason": "One is a Transformers 5.0 base_model_prefix loading regression; the other is a rope-validation helper TypeError on JSON-loaded lists. Different code paths." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Both involve quantization, but one is Qwen3Moe+FineGrainedFP8Config failure and the other is quantized model saving failure; different code paths." + "reason": "Both involve quantization, but one is Qwen3Moe with FineGrainedFP8Config and the other is quantized model saving. Not the same underlying issue." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "ExecuTorch export failure and flash-attention s_aux=None runtime crash are separate problems." + "reason": "Gemma3n executorch export failures and a flash_attention s_aux=None crash are separate export/runtime problems." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor loading import error and serve-time missing _tokenizer attribute are related to Gemma4 but not the same concrete bug." - }, - { - "left": "issue:45459", - "right": "issue:45712", - "accept": false, - "reason": "Tokenizer protobuf error masking and leftover dummy classes in dummy_pt_objects.py are unrelated." + "reason": "Both concern Gemma4 processor-related failures, but one is a missing dependency import and the other is a missing _tokenizer attribute in transformers serve. Distinct bugs." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are unrelated." + "reason": "Flash Attention 2 test failures are unrelated to void segmentation map processing in an image processor." }, { - "left": "issue:43756", - "right": "issue:45440", + "left": "issue:45459", + "right": "issue:45712", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and DeepSeekV3 implementation divergence are different model-specific bugs." + "reason": "Both involve error-handling or import-related behavior, but one is protobuf decode error handling in tokenizers and the other is dummy PT classes leaking into dir(transformers). Different issues." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBirdTokenizer special-token registration and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are distinct tokenizer issues." - }, - { - "left": "issue:44206", - "right": "issue:44479", - "accept": false, - "reason": "Unsupported center arg in LasrFeatureExtractor and Qwen video-input regression are unrelated." + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer-loading bugs." }, { - "left": "issue:45491", - "right": "issue:46693", + "left": "issue:43756", + "right": "issue:45440", "accept": false, - "reason": "Gemma3 NaN embeddings on mixed-length batches and FA2 long-context performance regression are both long-context/FA2 themed, but one is a correctness failure and the other is a performance regression." + "reason": "Smollm3 RoPE layer-count mismatch and DeepseekV3MoE divergence are model architecture differences, not the same defect." }, { - "left": "issue:29127", - "right": "issue:33357", + "left": "issue:44206", + "right": "issue:44479", "accept": false, - "reason": "LayoutLMv3 box-information error messaging and CLIP bus error on MacOS are unrelated." + "reason": "The LasrFeatureExtractor center-arg regression and the Qwen2_5/Qwen3 video-input regression are different API breakages in different model families." } ] }, @@ -9806,6 +9791,7 @@ "issue:29127", "issue:29942", "issue:33290", + "issue:33357", "issue:34567", "issue:34689", "issue:35141", @@ -9822,7 +9808,6 @@ "issue:43701", "issue:43761", "issue:43824", - "issue:43873", "issue:43931", "issue:43994", "issue:44186", @@ -9835,13 +9820,13 @@ "issue:45198", "issue:45245", "issue:45250", + "issue:45372", "issue:45412", "issue:45440", + "issue:45459", "issue:45491", "issue:45563", - "issue:45588", "issue:45593", - "issue:45600", "issue:45834", "issue:45859", "issue:46121", @@ -9851,259 +9836,266 @@ "issue:46693" ], "soft_pairs": [ + "issue:45491|issue:46693", + "issue:29127|issue:33357", + "issue:45372|issue:45459", "issue:45245|issue:45563", "issue:43065|issue:45412", "issue:43065|issue:43994", "issue:44265|issue:45198", "issue:27957|issue:29942", "issue:43575|issue:43701", - "issue:43066|issue:44462", "issue:45440|issue:45593", "issue:29127|issue:35141", "issue:42490|issue:43381", + "issue:43066|issue:44462", "issue:45412|issue:46369", "issue:43994|issue:45412", - "issue:43421|issue:45491", "issue:43575|issue:44387", + "issue:43421|issue:45491", "issue:44442|issue:45072", - "issue:43526|issue:45834", "issue:42915|issue:43931", + "issue:43526|issue:45834", "issue:43526|issue:46693", "issue:43824|issue:44279", "issue:27957|issue:29127", - "issue:46121|issue:46399", "issue:27957|issue:33290", + "issue:46121|issue:46399", "issue:43526|issue:44186", "issue:43994|issue:46369", - "issue:45250|issue:46640", - "issue:43334|issue:43531", "issue:33290|issue:34689", + "issue:43334|issue:43531", + "issue:45250|issue:46640", "issue:33290|issue:35141", - "issue:45588|issue:45600", - "issue:43761|issue:45859", "issue:33290|issue:34567", - "issue:43381|issue:43873" + "issue:43761|issue:45859" ], - "prepared_review_unit_hash": "d3aa06e5c9478f83e64bccadfd665fdaf913db05f342d4d2cd569bbd1049facf", + "prepared_review_unit_hash": "de274acb319777c6826975ecd961608d3983e437a4c2256907cc5003332784a6", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 32, - "serialized_chars": 36374, - "estimated_input_tokens": 9094, - "estimated_eval_tokens": 18444 + "soft_pair_count": 33, + "serialized_chars": 36752, + "estimated_input_tokens": 9188, + "estimated_eval_tokens": 18632 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most links are only superficial keyword matches (Flash Attention 2, OOM, tokenizer, model loading, etc.) and do not describe the same underlying bug or change.", + "summary": "These are mostly unrelated issues that only share broad keywords or subsystems. I reject all soft edges because they do not appear to be the same underlying bug or change.", "confidence": 0.94, - "canonical_issue_reason": "Issue 45250 is the broadest Flash Attention 2 umbrella item, so it is the least-specific and most representative anchor in this set.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 45250 is the best issue anchor because it is the most generic Flash Attention 2 entry; the rest of the cluster is too mixed to justify a tighter canonical.", + "best_issue_reason": "issue:46693 is the clearest standalone report in the set: it is open, recent, and describes a concrete Flash Attention 2 long-context performance regression with discussion and inbound references.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both mention FA2, but one is a Gemma3 NaN embedding bug on mixed-length batches and the other is a long-context performance regression; different symptoms and code paths." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 box-validation messaging and a MacOS CLIP bus error are unrelated failures." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "One is a mistral_common import error affecting Gemma 4 processor loading; the other hides tokenizer errors when protobuf is missing. Different bugs." + }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Different bugs: category-limit RuntimeError vs stale generate warning." + "reason": "Category-limit runtime error and a stale paging warning are not the same problem." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Different code paths: Sam3PixelDecoder dummy conv vs RT-DETR memory leak." + "reason": "Sam3PixelDecoder dummy Conv2d and RT-DETR memory release behavior are different model issues." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Different models and failure modes; only broad vision-model overlap." + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to SigLIP2 producing bad outputs." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "Different export/checkpointing issue vs Wav2Vec2 save/tokenization failure." + "reason": "torch.export failure via torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { "left": "issue:27957", "right": "issue:29942", "accept": false, - "reason": "Model-specific FA2 support bug vs generic failing FA2 tests." + "reason": "XLMRoberta FA2 support and failing FA2 tests are not the same concrete bug; one is model support, the other is test/CI failure." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "OOM on tensor-parallel model load vs checkpoint key mismatch." - }, - { - "left": "issue:43066", - "right": "issue:44462", - "accept": false, - "reason": "Tokenizer decoder type bug vs AutoTokenizer ignoring tokenizer.json." + "reason": "Qwen2 load OOM under tensor parallelism and a resume_from_checkpoint key mismatch are unrelated." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "Different model logic: DeepSeekV3MoE divergence vs D-FINE auxiliary loss handling." + "reason": "DeepseekV3MoE remote divergence and D-FINE auxiliary-loss behavior are different code paths and symptoms." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 error messaging vs embedding reinit regression." + "reason": "Tokenizer/error-message clarity and embedding reinitialization after resize are unrelated." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Local-path import warning vs gradient checkpointing eval-mode restriction." + "reason": "Path-trailing-slash import warning and gradient-checkpointing eval-mode restriction are different issues." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are separate tokenizer-loading bugs." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "Both mention memory, but one is RT-DETR GC behavior and the other sam3-litetext OOM." + "reason": "Both mention memory, but one is an RT-DETR GC leak while the other is a Sam3-LiteText overflow; not the same bug." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "Different subsystems: model outputs/pipeline behavior vs model memory cleanup." + "reason": "SigLIP2 nonsensical results and RT-DETR memory retention are unrelated." }, { - "left": "issue:43421", - "right": "issue:45491", + "left": "issue:43575", + "right": "issue:44387", "accept": false, - "reason": "Tokenizer post-processor updates vs Gemma3 sliding-window NaNs." + "reason": "Both are OOM reports, but the model/context and root cause differ (tp load vs int4 quantization reserved memory)." }, { - "left": "issue:43575", - "right": "issue:44387", + "left": "issue:43421", + "right": "issue:45491", "accept": false, - "reason": "Both are OOM-related, but one is TP load and the other int4 reserved-memory regression." + "reason": "TokenizersBackend post-processor updates and Gemma3 NaN embeddings are unrelated." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Different loading/tokenizer bug vs bfloat16 dtype mismatch in inference." + "reason": "FastSpeech2ConformerTokenizer loading and bfloat16 dtype mismatches are different failures." }, { - "left": "issue:43526", - "right": "issue:45834", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs Kosmos2.5 long-OCR index error." + "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are distinct model-loading issues." }, { - "left": "issue:42915", - "right": "issue:43931", + "left": "issue:43526", + "right": "issue:45834", "accept": false, - "reason": "Different Qwen family issues: FP8 config failure vs weight-shape mismatch." + "reason": "BeitImageProcessorFast reduce_labels and Kosmos2.5 long-ocr index errors are unrelated." }, { "left": "issue:43526", "right": "issue:46693", "accept": false, - "reason": "Image label reduction bug vs FA2 long-context performance regression." + "reason": "A Beit image-processor label bug is unrelated to a Flash Attention 2 performance regression." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Specific Qwen2.5-VL import error vs generic dependency issue." + "reason": "A missing Qwen2_5_VL symbol import and a generic dependency issue are not the same underlying bug." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "Unrelated: Flash Attention 2 model support vs LayoutLMv3 error messaging." + "reason": "FA2 support for XLMRoberta and a LayoutLMv3 error-message complaint are unrelated." }, { - "left": "issue:46121", - "right": "issue:46399", + "left": "issue:27957", + "right": "issue:33290", "accept": false, - "reason": "ROPE validation TypeError vs missing decorator on rotary embedding registration." + "reason": "Flash Attention 2 support and an Adafactor/DeepSpeed OOM are different problems." }, { - "left": "issue:27957", - "right": "issue:33290", + "left": "issue:46121", + "right": "issue:46399", "accept": false, - "reason": "Flash Attention 2 support issue vs DeepSpeed Adafactor OOM." + "reason": "RoPE-parameter dict conversion TypeError and a Qwen3.5 kernel-registration error are unrelated." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Different tokenizer/image-processor bugs with different inputs and symptoms." + "reason": "Beit label reduction and LayoutLMv2Tokenizer NER/padding crashes are different tokenizer/image-processor bugs." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "Different model behavior: nonsensical outputs vs memory overflow." + "reason": "SigLIP2 bad outputs and Sam3-LiteText memory overflow are unrelated." }, { - "left": "issue:45250", - "right": "issue:46640", + "left": "issue:33290", + "right": "issue:34689", "accept": false, - "reason": "Too generic to treat as one bug; both mention FA2 but no concrete shared failure mode." + "reason": "Adafactor OOM in DeepSpeed and Llama 3.2 Vision loading breakage are different issues." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Qwen3-VL config attribute error vs Qwen3-MoE sliding-window issue." + "reason": "Qwen3-VL missing pad_token_id and Qwen3-MoE sliding_window issues share a family name but not the same bug." }, { - "left": "issue:33290", - "right": "issue:34689", + "left": "issue:45250", + "right": "issue:46640", "accept": false, - "reason": "Deepspeed/optimizer OOM vs model-loading regression for Llama 3.2 Vision." + "reason": "These are both Flash Attention 2 topics, but the evidence is too broad to show the same concrete bug or mergeable change." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "OOM in optimizer path vs embedding reinitialization bug." - }, - { - "left": "issue:45588", - "right": "issue:45600", - "accept": false, - "reason": "Flash-attention runtime crash vs auto_mappings CI breakage." - }, - { - "left": "issue:43761", - "right": "issue:45859", - "accept": false, - "reason": "CLIP hidden_states regression vs CPU-offload incompatibility." + "reason": "DeepSpeed OOM and embedding reinitialization after resize are unrelated." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "OOM in training stack vs TrainerState token-count update bug." + "reason": "OOM under Adafactor/DeepSpeed and num_input_tokens_seen not updating are unrelated training bugs." }, { - "left": "issue:43381", - "right": "issue:43873", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction vs quantization offload behavior." + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different failures." } ] }, @@ -10132,6 +10124,7 @@ "issue:42490", "issue:43262", "issue:43316", + "issue:43381", "issue:43493", "issue:43525", "issue:43611", @@ -10147,12 +10140,9 @@ "issue:44062", "issue:44077", "issue:44079", - "issue:44188", - "issue:44190", "issue:44206", "issue:44442", "issue:44479", - "issue:44496", "issue:44514", "issue:44743", "issue:44811", @@ -10170,6 +10160,8 @@ "issue:45381", "issue:45440", "issue:45479", + "issue:45588", + "issue:45600", "issue:45759", "issue:45907", "issue:46178", @@ -10177,12 +10169,14 @@ "issue:46697" ], "soft_pairs": [ + "issue:45588|issue:45600", + "issue:43381|issue:43873", "issue:44514|issue:45381", "issue:43646|issue:44077", - "issue:44977|issue:45245", "issue:43262|issue:44206", - "issue:44442|issue:44898", + "issue:44977|issue:45245", "issue:44743|issue:44811", + "issue:44442|issue:44898", "issue:34689|issue:43611", "issue:42490|issue:44743", "issue:45003|issue:45440", @@ -10192,8 +10186,8 @@ "issue:43723|issue:45479", "issue:43493|issue:43827", "issue:44077|issue:44479", - "issue:44062|issue:45907", "issue:45325|issue:46697", + "issue:44062|issue:45907", "issue:43873|issue:44871", "issue:43761|issue:44079", "issue:43749|issue:43931", @@ -10201,200 +10195,198 @@ "issue:43643|issue:44913", "issue:44855|issue:45137", "issue:44987|issue:46178", - "issue:43316|issue:45759", - "issue:43761|issue:44188", - "issue:44190|issue:44496" + "issue:43316|issue:45759" ], - "prepared_review_unit_hash": "7eec967297f3d663090ecf0a2326581fe922820a788173b32f2a1b3b2745e893", + "prepared_review_unit_hash": "be0a166c8899925b4101e13a82f941ccf92af3209faf1dbed73fdf900fd48b63", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36074, - "estimated_input_tokens": 9019, - "estimated_eval_tokens": 18294 + "serialized_chars": 36014, + "estimated_input_tokens": 9004, + "estimated_eval_tokens": 18264 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The candidate edges are mostly same-subsystem regressions, but none look like the same concrete bug/change closely enough to merge as duplicates. I would reject all soft edges and keep the issues separate.", - "confidence": 0.82, - "canonical_issue_reason": null, + "summary": "These soft links are overwhelmingly false positives: they connect issues from different models, features, and failure modes rather than the same underlying bug or change. I would not merge any of the proposed pairs.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the items span unrelated regressions across loading, processors, configs, quantization, docs, and model-specific behavior, so there isn\u2019t a cohesive duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue has to anchor the set, issue:43749 is the strongest standalone report: it describes a concrete loading regression, has the most discussion/reference activity, and reads like a well-formed bug report.", + "best_issue_reason": "No global best issue; every proposed pair is still about a different concrete code path, so none can serve as a representative duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "Different bugs: flash-attention runtime crash vs auto-mapping/CI reference cleanup." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Unrelated failures: gradient checkpointing in eval mode vs quantization/offloading behavior." + }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both involve Qwen2.5-VL, but one is a batching crash in `apply_chat_template()` and the other is incorrect video `vision_position_ids`; different failure modes and likely different fixes." + "reason": "Same product family but different code paths: batched chat template padding crash vs video position-id computation." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Both mention v5/custom model init, but one is a broad custom-initialization regression and the other is a specific `patchtsmixer` `post_init` rule change." - }, - { - "left": "issue:44977", - "right": "issue:45245", - "accept": false, - "reason": "Unrelated bugs: flash-attention generation failure versus an embedding/category cardinality error." + "reason": "Both initialization-related, but one is a broad v5 custom-model init regression and the other is a specific post_init policy issue." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Both are processor/feature-extractor issues, but the first is wrong default sampling rate propagation and the second is an unsupported `center` argument crash." + "reason": "Different audio preprocessing bugs: default sample rate handling vs unsupported center argument." }, { - "left": "issue:44442", - "right": "issue:44898", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "Different model families and different code paths: tokenizer loading vs image-classification interpolation behavior." + "reason": "Completely unrelated: flash-attention generation bug vs category-cardinality runtime error." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Recurrent cache-state reset and Whisper `batch_decode()` special-token handling are separate bugs in different components." + "reason": "Different subsystems and symptoms: recurrent cache state reset vs Whisper batch_decode token handling." }, { - "left": "issue:34689", - "right": "issue:43611", + "left": "issue:44442", + "right": "issue:44898", "accept": false, - "reason": "Both are loading regressions, but the underlying causes are different: Llama 3.2 vision loading breakage versus `base_model_prefix` handling in v5." + "reason": "Tokenizer loading failure vs Perceiver image classification/interpolation bug; no shared underlying defect." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but for different versions and different causes; not the same concrete bug." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "A trailing-slash local-path warning is unrelated to cache/recurrent-state behavior." + "reason": "Local-path import warning and Qwen recurrent-state reset are unrelated." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe `sys.modules[]` access and DeepSeekV3MoE divergence are unrelated bugs." + "reason": "sys.modules access bug vs DeepSeekV3MoE implementation divergence; different code paths." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and FastSpeech2Conformer tokenizer loading are unrelated failures." + "reason": "BitNet packed-weight loading issue is unrelated to FastSpeech2Conformer tokenizer loading." }, { "left": "issue:42175", "right": "issue:46234", "accept": false, - "reason": "Packaging/backends mismatch for TensorFlow versus a missing import-utils availability helper; not the same bug." + "reason": "Package-backend install issue vs missing import_utils symbol; not the same bug." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config/API discrepancies, but they affect different models and different missing fields." + "reason": "Both config API inconsistencies, but for different models and different missing fields." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer-loading regression and zero-loss sequence-classification bug are unrelated." + "reason": "Tokenizer loading regression vs sequence-classification loss bug; unrelated code paths." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch and docs still referencing removed `pipeline()` APIs are separate issues." + "reason": "SigLIP2 implementation mismatch vs docs still referencing removed pipeline APIs." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "A `post_init` policy change for one model is not the same as the Qwen video-input regression." + "reason": "Specific post_init restriction vs Qwen video-input regression; different features and fixes." }, { - "left": "issue:44062", - "right": "issue:45907", + "left": "issue:45325", + "right": "issue:46697", "accept": false, - "reason": "`AddedToken` constructor duplication and `list_repo_files` not forwarding `revision` are unrelated API bugs." + "reason": "Qwen2.5-VL rope/temporal position bug vs Swin2SR over-padding bug." }, { - "left": "issue:45325", - "right": "issue:46697", + "left": "issue:44062", + "right": "issue:45907", "accept": false, - "reason": "Qwen2.5-VL rope/position-id scaling and Swin2SR over-padding are different image/video preprocessing bugs." + "reason": "AddedToken kwargs collision vs repo listing missing revision forwarding; unrelated." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma eos-token config inconsistency do not share the same root cause." + "reason": "Quantization/offloading issue vs Gemma eos_token_id config mismatch." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "One is a model forward-path output bug, the other is a `ModelOutput` key assignment issue; related conceptually but not the same defect." + "reason": "CLIPVision hidden_states not returned vs generic ModelOutput key assignment bug; related area but different defects." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM-efficient loading and Qwen3-VL weight-shape mismatch are different loading failures." + "reason": "FSDP CPU RAM efficient loading breakage vs Qwen3-VL shape-mismatch loading error." }, { - "left": "issue:43931", - "right": "issue:45216", + "left": "issue:44987", + "right": "issue:46178", "accept": false, - "reason": "A shape-mismatch load error is not the same as a `save_pretrained()` checkpoint regression." + "reason": "Model loading regression vs backend registry error; no shared underlying bug." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Missing fields from `trust_remote_code=True` config loading and `rotary_pct` reverting on reload are distinct config-persistence bugs." + "reason": "Missing fields with trust_remote_code vs rotary_pct not persisting on reload; different config behaviors." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 import/`torch.jit.script` parsing failure and DeepSpeed ZeRO3 deque underflow are unrelated." + "reason": "Python 3.13 torch.jit parsing/import issue vs DeepSpeed ZeRO3 deque error." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Model-loading regression for a specific repo and TensorFlow-text backend registration failure are not the same issue." + "reason": "Different failure modes and code paths; one is model loading, the other is backend availability resolution." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both touch composite configs, but one is a Gemma3TextConfig API discrepancy and the other is missing `text_config` unwrapping for Qwen multimodal configs." - }, - { - "left": "issue:43761", - "right": "issue:44188", - "accept": false, - "reason": "Hidden-states being `None` and attention-kernel divergence under `torch.compile` are separate regressions." - }, - { - "left": "issue:44190", - "right": "issue:44496", - "accept": false, - "reason": "Dataset script loading and an unrecognized model/config error are unrelated." + "reason": "Config API discrepancy for Gemma3TextConfig vs composite Qwen multimodal from_config unwrapping." } ] }, @@ -10441,19 +10433,22 @@ "issue:43701", "issue:43716", "issue:43742", + "issue:43761", "issue:43825", "issue:44077", "issue:44162", + "issue:44188", + "issue:44190", "issue:44442", "issue:44464", "issue:44496", "issue:44534", - "issue:44610", + "issue:44568", "issue:44792", "issue:44871", "issue:44913", + "issue:45479", "issue:45584", - "issue:45632", "issue:45698", "issue:45803", "issue:46050", @@ -10461,12 +10456,12 @@ "issue:46234", "issue:46291", "issue:46365", - "issue:46369", "issue:46421", - "issue:46424", "issue:46506" ], "soft_pairs": [ + "issue:43761|issue:44188", + "issue:44190|issue:44496", "issue:43232|issue:43388", "issue:44442|issue:45803", "issue:42907|issue:45698", @@ -10475,10 +10470,11 @@ "issue:43575|issue:43742", "issue:27957|issue:32090", "issue:25251|issue:28282", - "issue:46234|issue:46291", + "issue:44568|issue:45479", "issue:43550|issue:43701", - "issue:44464|issue:46421", + "issue:46234|issue:46291", "issue:43606|issue:43701", + "issue:44464|issue:46421", "issue:16998|issue:22355", "issue:43295|issue:43825", "issue:43295|issue:44871", @@ -10492,206 +10488,210 @@ "issue:45803|issue:46365", "issue:44162|issue:44792", "issue:45803|issue:46170", - "issue:45632|issue:46424", - "issue:43404|issue:43441", - "issue:44610|issue:46369" + "issue:43404|issue:43441" ], - "prepared_review_unit_hash": "5efbd9ac07db277920e981b7a5122e786d920be0e972868bc66f532eff861dfb", + "prepared_review_unit_hash": "30ba16acf0d29ecb2a8c0496f0b2a0ba25374e0047b414fa90aafd56c18f37ac", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35306, - "estimated_input_tokens": 8827, - "estimated_eval_tokens": 17910 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36350, + "estimated_input_tokens": 9088, + "estimated_eval_tokens": 18432 }, "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 spanning loading, generation, caching, training, and model-specific regressions. I did not find any soft-edge pair that looks like the same underlying bug/change.", - "confidence": 0.78, - "canonical_issue_reason": "No single canonical duplicate stands out because the items are heterogeneous. If one representative is needed, issue #45698 is the closest fit: it is open, fairly active, and describes a broadly applicable save/load custom-module regression.", + "summary": "This cluster is a loose mix of unrelated Transformers issues across tokenization, generation, model loading, training, and backend/runtime regressions. None of the soft-edge pairs look like the same underlying bug or change, so all are rejected as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "#45698 is the most broadly relevant and active issue in the set, with a general from_pretrained/save_pretrained failure that could affect many custom models.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Different failures: CLIPVisionModel hidden_states regression vs attention-kernel divergence in a separate torch.compile branch." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an example script is unrelated to an unrecognized-model/config.json error." + }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Different subsystems and symptoms: generation kwargs after sync_gpus vs metric gathering dropping tuple labels." + "reason": "Generation kwargs after sync_gpus and metric gathering in evaluation are different code paths and symptoms." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure is unrelated to Fuyu's tensor conversion error handling bug." + "reason": "Tokenizer load failure and a bare-except bug in Fuyu tensor conversion are unrelated issues." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load flows, but one is dequantized model saving and the other is wrong custom-module loading; not the same bug." + "reason": "Both involve save/load, but one is dequantized model saving and the other is custom module resolution after save_pretrained." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3D parallel launch failure and tensor-parallel OOM are both parallelism-related, but they are different concrete problems." + "reason": "A 3d_parallel example failure is not the same bug as tensor-parallel OOM loading Qwen2-57B." }, { "left": "issue:43701", "right": "issue:46506", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to CPU/CUDA output divergence in PPDocLayoutV3." + "reason": "Checkpoint key mismatch is unrelated to CPU/CUDA output divergence in PPDocLayoutV3." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2 and a key error loading MobileLLM are different model-loading failures." + "reason": "Tensor-parallel OOM and a key error loading MobileLLM are different loading failures." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "FlashAttention support for XLMRoberta is unrelated to a Trainer broadcast TypeError." + "reason": "FlashAttention support for XLMRoberta is unrelated to a Trainer broadcast NoneType error." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k output shape regression is unrelated to missing PyTorch import errors." + "reason": "Pipeline top_k output-shape behavior has nothing to do with missing PyTorch at import time." }, { - "left": "issue:46234", - "right": "issue:46291", + "left": "issue:44568", + "right": "issue:45479", "accept": false, - "reason": "Both are import-utils related, but one is a missing availability helper and the other is a kernels dependency/version-check failure." + "reason": "Tokenizer special-token regression and degenerate sequence-classification loss are separate bugs." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile + SDPA on Bamba is unrelated to checkpoint resume key mismatches." + "reason": "torch.compile SDPA failure for Bamba is unrelated to resume_from_checkpoint key mismatches." }, { - "left": "issue:44464", - "right": "issue:46421", + "left": "issue:46234", + "right": "issue:46291", "accept": false, - "reason": "Both mention chunked generation, but one is output inconsistency under compilation while the other is a recompilation/performance issue." + "reason": "A missing import_utils helper and a kernels dependency/version import error are different availability/import problems." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch for bark-small is unrelated to resume_from_checkpoint key mismatches." + "reason": "CPU offload device mismatch is unrelated to checkpoint resume key handling." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Inconsistent generated outputs and spurious recompilation are distinct chunked-generation issues." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "model_max_length on DeBERTa-V3 and missing transformers.onnx are unrelated issues." + "reason": "model_max_length behavior for DeBERTa-V3 is unrelated to the transformers.onnx import error." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "v4.57.5 custom processor/tokenizer regression is unrelated to a pipeline error message about translation tasks in v5." + "reason": "Custom model/processor regression and a pipeline error-message regression are different bugs." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Custom multimodal processor regression and Gemma-3 eos_token_id config mismatch are different bugs." + "reason": "Processor/tokenizer regression is unrelated to Gemma3 eos_token_id configuration mismatch." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Gemma eos_token_id inconsistency is not the same as Whisper empty-transcription generation failure after special-token alignment." + "reason": "Gemma3 EOS config inconsistency is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image preprocessor dtype mismatch are unrelated." + "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image-preprocessor dtype mismatch are separate model issues." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "Backend packaging issue with pip extras is unrelated to a config API discrepancy." + "reason": "Backend install packaging and Gemma3TextConfig API mismatch are unrelated." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Optional post_init validation for patchtsmixer is unrelated to non-persistent buffer initialization junk." + "reason": "patchtsmixer post_init allowance and junk non-persistent buffers are different v5 regressions." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Gemma eos_token_id config mismatch and sinq quantized save/load breakage are different model-specific problems." + "reason": "Gemma3 EOS token config mismatch is unrelated to quantized-model save/load breakage." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage in Transformers 5.0.0 is unrelated to GPTNeoX rotary_pct serialization/reload behavior." + "reason": "Custom model initialization breakage is not the same as GPTNeoXConfig rotary_pct not persisting." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "GGUF from_pretrained breakage is not the same as a generic missing model_type config error." + "reason": "gguf_file loading regression and missing model_type config are different model-loading failures." }, { "left": "issue:45803", "right": "issue:46365", "accept": false, - "reason": "Fuyu tensor conversion error handling has no relation to importing Trainer failing because of datasets." + "reason": "Fuyu bare-except behavior is unrelated to importing Trainer with datasets installed." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 being broken and a janus model test failure are different model-specific issues." + "reason": "ESM2 breakage and Janus image-generation test failure are unrelated model-specific problems." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "KeyboardInterrupt swallowing in Fuyu is unrelated to Windows/ROCm torch.distributed import problems." - }, - { - "left": "issue:45632", - "right": "issue:46424", - "accept": false, - "reason": "Local cache path collisions from trust_remote_code are unrelated to max_cache_length only applying to Quantized Cache." + "reason": "Fuyu exception handling bug is unrelated to a Windows ROCm torch.distributed import issue." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Mistral3 weight tying bug and Ministral-3 FlashAttention failure are different code-path problems, despite the shared family." - }, - { - "left": "issue:44610", - "right": "issue:46369", - "accept": false, - "reason": "OmDet-Turbo input-size mismatch is unrelated to sam3-litetext memory overflow." + "reason": "Mistral-3 weight tying and Ministral-3 FlashAttention failures affect different code paths and symptoms." } ] }, @@ -10739,11 +10739,11 @@ "issue:44479", "issue:44514", "issue:44534", + "issue:44610", "issue:44617", "issue:44734", "issue:44743", "issue:44792", - "issue:44855", "issue:44913", "issue:44964", "issue:44977", @@ -10751,7 +10751,6 @@ "issue:45200", "issue:45406", "issue:45542", - "issue:45588", "issue:45600", "issue:45632", "issue:45698", @@ -10759,24 +10758,27 @@ "issue:46178", "issue:46202", "issue:46234", - "issue:46421", + "issue:46369", + "issue:46424", "issue:46617", "issue:46640", "issue:46697" ], "soft_pairs": [ + "issue:45632|issue:46424", "issue:43650|issue:46202", + "issue:44610|issue:46369", "issue:43761|issue:44479", "issue:43572|issue:46178", "issue:45542|issue:46234", "issue:43540|issue:44514", + "issue:43504|issue:45698", + "issue:43388|issue:43688", "issue:45127|issue:45200", "issue:43065|issue:44617", "issue:43352|issue:46640", - "issue:43504|issue:45698", - "issue:43388|issue:43688", - "issue:44792|issue:45600", "issue:44792|issue:44977", + "issue:44792|issue:45600", "issue:43673|issue:44743", "issue:41950|issue:43475", "issue:43525|issue:43531", @@ -10785,190 +10787,188 @@ "issue:44534|issue:44913", "issue:44734|issue:45406", "issue:46082|issue:46697", - "issue:43931|issue:44060", "issue:45632|issue:46617", - "issue:44855|issue:45588", - "issue:43827|issue:43994", - "issue:43673|issue:46421" + "issue:43931|issue:44060", + "issue:43827|issue:43994" ], - "prepared_review_unit_hash": "d7e2bbfa64a1fc5dbb62bc1176bd8697cf4eb9146e4041d83125677685dda37b", + "prepared_review_unit_hash": "9ff6ecd7fd3ac383d155a2b4cd150d51b52b83b32430ed2f30f7b8ce669e26e8", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 25, - "serialized_chars": 35275, - "estimated_input_tokens": 8819, - "estimated_eval_tokens": 17894 + "serialized_chars": 35141, + "estimated_input_tokens": 8786, + "estimated_eval_tokens": 17828 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very noisy: it mixes unrelated issues across loading, cache, multimodal processors, flash-attention, configs, and docs. None of the soft pairs look like true duplicates, so I would not merge any of them.", - "confidence": 0.41, - "canonical_issue_reason": "issue:45698 is the most active open loading/serialization report in the set, so it is the closest available representative, but the cluster is too heterogeneous for a strong canonical choice.", + "summary": "The set does not form one coherent duplicate cluster; the soft pairs are mostly unrelated issues that only share broad subsystem keywords. I rejected all soft edges as distinct bugs/changes.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical duplicate exists across the full set. If a representative issue must be chosen, issue 45698 is the broadest and most active open loading bug in the group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45698 is the best global anchor only by activity and breadth; semantically it does not unify the rest of the items well.", + "best_issue_reason": "Issue 45698 is the best overall representative because it is open, broadly phrased, and has the highest discussion/reference activity among the items.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both involve loading/cache wording, but one is a trust_remote_code cache-path collision and the other is max_cache_length behavior; different bugs." + }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "Unrelated topics: a vague data-related issue versus a Gemma pipeline roundtrip/tokenization problem." + "reason": "Unrelated topics: a placeholder issue versus a Gemma pipeline roundtrip/tokenization bug." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Different models and failure modes: input-size mismatch in OmDet-Turbo versus OOM in sam3-litetext." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Both are regressions, but one is CLIP vision hidden-state output and the other is Qwen video-input handling; different code paths." + "reason": "CLIPVision hidden_states regression is unrelated to Qwen video-input regression." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "Missing pad_token_idx in a model config is unrelated to tensorflow_text backend mapping errors." + "reason": "StableLmConfig missing pad_token_idx is unrelated to a tensorflow_text backend mapping error." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "TensorBoard/TensorFlow backend setup is not the same bug as a missing import_utils feature flag." + "reason": "Tensorboard/TensorFlow backend availability issue is unrelated to a missing import_utils helper." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Different multimodal failures: Qwen3OmniMoe video validation versus Qwen2_5_VL batched chat-template padding." + "reason": "Both are video/input processing bugs, but they affect different Qwen processors and different code paths." }, { - "left": "issue:45127", - "right": "issue:45200", + "left": "issue:43504", + "right": "issue:45698", "accept": false, - "reason": "LoRA/extended-vocab collapse and defaulting mm_token_type_ids are distinct problems." + "reason": "Legacy preset loading in Beit is a different bug from incorrect custom-module loading after save_pretrained." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "A dummy Conv2d placeholder is unrelated to CUDA OOM in Sam3Video." + "reason": "Distributed metrics label truncation is unrelated to auxiliary-loss normalization." }, { - "left": "issue:43352", - "right": "issue:46640", + "left": "issue:45127", + "right": "issue:45200", "accept": false, - "reason": "Both mention Flash Attention 2, but one is a model-specific unsupported guard and the other is a generic flash-attention issue." + "reason": "LoRA/extended-vocab collapse with tied embeddings is not the same as missing mm_token_type_ids defaults." }, { - "left": "issue:43504", - "right": "issue:45698", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "Legacy-field model preset loading and wrong custom-module loading are different failure modes." + "reason": "A dummy Conv2d implementation issue is not the same as a SAM3 video memory overflow." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:43352", + "right": "issue:46640", "accept": false, - "reason": "Metric label truncation and auxiliary-loss normalization are unrelated." + "reason": "Both mention Flash Attention 2, but the specific unsupported-model bug is not clearly the same as the generic issue." }, { "left": "issue:44792", - "right": "issue:45600", + "right": "issue:44977", "accept": false, - "reason": "A Janus test failure and a stale auto_mappings entry are not the same bug." + "reason": "Janus test failure and Qwen3.5 flash-attention generation bug are different models and symptoms." }, { "left": "issue:44792", - "right": "issue:44977", + "right": "issue:45600", "accept": false, - "reason": "Test failure versus Qwen3.5 flash-attention generation bug; no shared concrete fix." + "reason": "A failing image-generation test is unrelated to stale auto_mappings references to removed configs." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both involve cache/state, but one is missing GenerationMixin cache in chunked prefill and the other resets recurrent states with seq_len>1." + "reason": "Chunked-prefill cache missing and recurrent states resetting are both cache-related but describe different mechanisms and code paths." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Pipeline processor selection and a missing encoder output attribute are unrelated." + "reason": "Video-classification processor mismatch is unrelated to a missing fpn_position_embeddings attribute in SAM3 video output." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Config attribute absence and sliding_window behavior in Qwen3-MoE are different issues." + "reason": "Missing pad_token_id in Llama4Config is unrelated to Qwen3-MoE sliding_window behavior." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading in a training script is unrelated to loading a multimodal model." + "reason": "Local dataset loading in a training script is unrelated to loading a multimodal model preset." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding resize/reinit and trailing-slash local path import warnings are unrelated." + "reason": "Embedding reinitialization on resize is unrelated to import warnings from a trailing-slash local path." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk and GPTNeoX rotary_pct reload behavior are different serialization/config bugs." + "reason": "Non-persistent buffer initialization corruption is unrelated to GPTNeoX rotary_pct persistence." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "KV-cache tensor indexing in serve is unrelated to Gemma4Processor missing _tokenizer." + "reason": "KV-cache continuation indexing and missing _tokenizer on Gemma4Processor are unrelated serve failures." }, { "left": "issue:46082", "right": "issue:46697", "accept": false, - "reason": "LlamaConfig head_dim validation is not the same as Swin2SR image over-padding." - }, - { - "left": "issue:43931", - "right": "issue:44060", - "accept": false, - "reason": "Model weight-shape mismatch and an incorrect tied-weights warning are distinct loading/tie-weight issues." + "reason": "LlamaConfig head_dim validation is unrelated to Swin2SR image over-padding." }, { "left": "issue:45632", "right": "issue:46617", "accept": false, - "reason": "Both touch trust_remote_code caching, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache lookup; different failure modes." + "reason": "Both mention trust_remote_code/cache, but one is a leaf-directory collision and the other is a FileNotFoundError from cache loading; not the same bug." }, { - "left": "issue:44855", - "right": "issue:45588", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Python 3.13 torchscript parsing error and flash_attention s_aux=None AttributeError are unrelated." + "reason": "Qwen3-VL weight-shape mismatch is unrelated to an incorrect tied-weights warning in Qwen3-Next." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still mentioning pipeline() and a SigLIP2 model producing bad outputs are not the same bug." - }, - { - "left": "issue:43673", - "right": "issue:46421", - "accept": false, - "reason": "Chunked-prefill missing cache and cache lazy-init recompilation are related only at a high level, not the same concrete defect." + "reason": "Docs still referencing pipeline() is unrelated to a SigLIP2 model producing nonsensical outputs." } ] }, @@ -11003,6 +11003,7 @@ "issue:43540", "issue:43550", "issue:43644", + "issue:43673", "issue:43761", "issue:44077", "issue:44112", @@ -11011,7 +11012,6 @@ "issue:44462", "issue:44464", "issue:44479", - "issue:44521", "issue:44610", "issue:44617", "issue:44734", @@ -11031,17 +11031,19 @@ "issue:45542", "issue:45563", "issue:45584", + "issue:45588", "issue:45600", - "issue:45632", "issue:45803", "issue:45907", - "issue:46234", "issue:46369", + "issue:46421", "issue:46489", "issue:46655", "issue:46697" ], "soft_pairs": [ + "issue:44855|issue:45588", + "issue:43673|issue:46421", "issue:43761|issue:44242", "issue:44617|issue:45584", "issue:45563|issue:45584", @@ -11050,9 +11052,9 @@ "issue:44933|issue:45362", "issue:43335|issue:44855", "issue:45381|issue:46369", - "issue:43550|issue:44190", "issue:44077|issue:44877", "issue:44479|issue:44734", + "issue:43550|issue:44190", "issue:41720|issue:45245", "issue:43335|issue:44190", "issue:44464|issue:45563", @@ -11069,24 +11071,22 @@ "issue:43540|issue:44734", "issue:44938|issue:45542", "issue:45276|issue:45584", - "issue:44462|issue:45907", - "issue:44521|issue:45381", - "issue:45632|issue:46234" + "issue:44462|issue:45907" ], - "prepared_review_unit_hash": "0542ffbfef9362d73387638e0d5fba4104d8bc5e573323b5f2bf9706562dbc2b", + "prepared_review_unit_hash": "1d1f31db77022d49c8f36c38cab2f7385aac50e6054c80c46e21848f2c5fabcf", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36657, + "serialized_chars": 36660, "estimated_input_tokens": 9165, "estimated_eval_tokens": 18586 }, @@ -11095,198 +11095,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous and does not form a broad duplicate set. The only strong duplicate-looking edge is the DeepSeek tokenizer regression pair; the rest are different bugs across unrelated models, configs, or subsystems.", - "confidence": 0.92, - "canonical_issue_reason": "issue:44779 is the better canonical because it is the broader, earlier report of the DeepSeek tokenizer regression in v5; issue:46489 reads like a narrower follow-up describing the same underlying problem and referencing a gap in an earlier fix.", + "summary": "This set is mostly a mix of unrelated issues from different models/features. The only strong duplicate-style match is the DeepSeek tokenizer regression pair (44779/46489); the rest look like distinct bugs or follow-up symptoms in different code paths.", + "confidence": 0.71, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44779 is the best representative issue for the only accepted duplicate pair and has the clearest general title/description for the underlying tokenizer regression.", + "best_issue_reason": "issue:44779 is the best issue anchor because issue:46489 reads like a follow-up to the same DeepSeek tokenizer regression; the rest of the cluster is too heterogeneous to support a single canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Different subsystems and failures: Python 3.13 import/torch.jit parsing vs flash-attention sink handling." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill/cache, but one is a missing cache in v5.0 and the other is lazy-init causing recompilation; too different to treat as the same bug." + }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "Different models and failures: CLIPVision hidden_states handling vs MoE load-balancing loss logic." + "reason": "CLIPVision hidden_states regression and MoE load-balancing loss omission are unrelated issues." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Unrelated symptoms and code paths: Sam3Video OOM vs Whisper empty-transcription generation failure." + "reason": "CUDA OOM in Sam3Video is not the same bug as Whisper empty-transcription generation failure." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Paged-generate warning cleanup is unrelated to Whisper's align_special_tokens failure." + "reason": "A stale warning for num_return_sequences is unrelated to Whisper failing on empty transcription." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "One is a global buffer-init regression; the other is a GPTNeoX config reload defaulting issue." + "reason": "Non-persistent buffer initialization junk and GPTNeoXConfig reload behavior are different code paths." }, { "left": "issue:25251", "right": "issue:27957", "accept": false, - "reason": "No overlap in bug or subsystem: pipeline top_k nesting vs XLM-RoBERTa Flash Attention 2 support." + "reason": "Pipeline top_k nesting bug and XLMRoberta Flash Attention 2 support are unrelated." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Different issues: missing image_utils import vs Qwen3.5 chat crash." + "reason": "Missing image_utils import and Qwen3.5 chat crash are different problems." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers config layer construction and DebertaV2 JIT parse error are unrelated." + "reason": "Sparse-layer config creation and Python 3.13 import parsing are unrelated." }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Qwen2.5-VL video position IDs and Sam3-lite text memory overflow are different bugs." - }, - { - "left": "issue:43550", - "right": "issue:44190", - "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to local dataset loading in the image-classification script." + "reason": "Qwen video vision-position-id regression and Sam3-lite memory overflow are different bugs." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Both mention config semantics, but they are about different models and different failure modes." + "reason": "Strict config loading for granite_speech and optional post_init on patchtsmixer are separate config issues." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression in Qwen VL models vs serve KV-cache continuation indexing bug are different code paths." + "reason": "Both are generation regressions, but one is video-input position ids and the other is KV-cache continuation tensor indexing." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to local dataset loading in the example script." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "A800 auto device-mapping assert and category-count overflow are unrelated." + "reason": "Auto device mapping on Qwen3 and category cardinality limits are unrelated." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to the dataset-loading script issue." + "reason": "Sparse-layer config bug and dataset loading failure are unrelated." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is unrelated to the stale num_return_sequences warning." + "reason": "Chunked generation inconsistency and a stale warning are different issues." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Sam3PixelDecoder dummy conv and OmDet-Turbo processor/model size mismatch are different bugs." + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor resolution mismatch are different bugs." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge with tied embeddings is unrelated to Qwen3.5 visual encoder save_pretrained keys." + "reason": "LoRA merge with tied embeddings and incorrect visual encoder keys on save_pretrained are not the same problem." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "Documentation example errors are unrelated to the local dataset loading bug." + "reason": "SigLIP2 documentation errors and local dataset loading are unrelated." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "CI failures in GraniteSpeech and auto_mappings reference removal are different root causes." + "reason": "Stale GraniteSpeech CI test and removed Sam3LiteText auto_mappings references are different CI breakages." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both involve video, but one is wrong vision_position_ids and the other is incorrect video padding calculation." + "reason": "Both are video/image processor bugs, but the model, symptom, and likely fix differ (vision_position_ids vs padding calculation)." }, { "left": "issue:43526", "right": "issue:46697", "accept": false, - "reason": "BEiT reduce_labels bug and Swin2SR over-padding bug are unrelated image-processor issues." + "reason": "Beit reduce_labels returning one label and Swin2SR over-padding are different image processor bugs." }, { "left": "issue:44779", "right": "issue:46489", "accept": true, - "reason": "Both report DeepSeek tokenizer output regressions on v5+; 46489 looks like a narrower follow-up of the same underlying tokenizer bug." + "reason": "Both describe the same DeepSeek tokenizer regression on v5+, with 46489 explicitly reading like a follow-up gap in the earlier fix." }, { "left": "issue:44617", "right": "issue:46369", "accept": false, - "reason": "Both mention memory pressure, but Sam3Video CUDA OOM and sam3-litetext memory overflow are different models and scenarios." + "reason": "Sam3Video CUDA OOM and Sam3-lite memory overflow are related by model family only, not the same concrete bug." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Processor/tokenizer API regression and Gemma4 resize_token_embeddings propagation are distinct issues." + "reason": "Processor/tokenizer regression and resize_token_embeddings not updating Gemma4 embedding fields are different bugs." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "XLNet device placement bug and Fuyu tensor-conversion exception handling are unrelated." + "reason": "XLNet device placement bug and a bare-except in Fuyu tensor conversion are unrelated." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError is different from serve KV-cache continuation indexing." + "reason": "Qwen3OmniMoe video input validation and serve KV-cache indexing failure are unrelated." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Python 3.14 import/load failure is unrelated to the TensorBoard-without-TF backend error." + "reason": "Python 3.14 import failure and tensorboard-without-TensorFlow backend error are different problems." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Gemma4 embedding resize propagation and Whisper empty transcription handling are unrelated." + "reason": "Gemma4 embedding resize bug and Whisper empty-transcription failure are unrelated." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to list_repo_files revision forwarding in CI." - }, - { - "left": "issue:44521", - "right": "issue:45381", - "accept": false, - "reason": "apply_chat_template assistant_mask bug is not the same as Qwen2.5-VL video position_ids corruption." - }, - { - "left": "issue:45632", - "right": "issue:46234", - "accept": false, - "reason": "Cache-path collision for trust_remote_code is unrelated to the missing import_utils helper." + "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files missing revision forwarding are unrelated repository/API bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only accepted soft pair is issue:44779 vs issue:46489, and the explanation matches their titles (same DeepSeek tokenizer regression, with 46489 framed as a follow-up/gap in a prior fix). The other soft-edge rejections are conservative and stay within what the titles support, without overstating evidence." + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted soft edge, issue:44779 \u2194 issue:46489, is well supported by the titles as the same DeepSeek tokenizer regression with the later issue reading like a follow-up gap. The rest of the pairwise rejections are reasonable given the different models, subsystems, or symptoms described in the issue titles." }, "evaluator_used": true, "retried": false, @@ -11317,16 +11317,13 @@ "issue:42175", "issue:42796", "issue:43295", - "issue:43381", "issue:43404", "issue:43479", - "issue:43526", "issue:43531", "issue:43643", "issue:43650", "issue:43673", "issue:43854", - "issue:43901", "issue:43976", "issue:44016", "issue:44112", @@ -11347,22 +11344,26 @@ "issue:45245", "issue:45335", "issue:45375", + "issue:45381", "issue:45446", "issue:45584", - "issue:45593", "issue:45600", + "issue:45632", "issue:46082", "issue:46121", + "issue:46234", "issue:46424", "issue:46506" ], "soft_pairs": [ + "issue:44521|issue:45381", + "issue:45632|issue:46234", "issue:44246|issue:45003", "issue:43854|issue:44964", "issue:44964|issue:45600", - "issue:46082|issue:46121", "issue:43643|issue:44496", "issue:44112|issue:44964", + "issue:46082|issue:46121", "issue:27957|issue:35141", "issue:42175|issue:44279", "issue:45245|issue:45584", @@ -11382,206 +11383,204 @@ "issue:43976|issue:44987", "issue:43295|issue:45335", "issue:44683|issue:45446", - "issue:44496|issue:45042", - "issue:43526|issue:43901", - "issue:43381|issue:45593" + "issue:44496|issue:45042" ], - "prepared_review_unit_hash": "508ee6489b154e2795e30a2e591e2dff46f54823109acdac5386a436004d11c2", + "prepared_review_unit_hash": "5e6409a9076183c36c4eb08f35f3f7f4ec5e0be8edbe4df865451768d2b7e1ed", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36055, - "estimated_input_tokens": 9014, - "estimated_eval_tokens": 18284 + "serialized_chars": 35502, + "estimated_input_tokens": 8876, + "estimated_eval_tokens": 18008 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a collection of superficially similar but distinct issues (imports, configs, model loading, cache behavior, tokenizer bugs). I would reject every soft pair; there is no clear duplicate cluster here. No PRs are present.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44964 is the most concrete, user-facing model-loading regression in the set and the best representative if a single issue must anchor this group, even though it is not a true duplicate of the others.", + "summary": "All soft-edge candidates are false positives: the items cover unrelated bugs, regressions, and feature requests across different models, pipelines, and import paths rather than the same underlying change or defect.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue stands out because none of the soft pairs look like true duplicates of the same bug or change.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44964 has a specific affected model, a reproducible failure, and a clear triage target, making it the most actionable issue among these superficially related tickets.", + "best_issue_reason": "No global best issue for deduplication purposes: the set is heterogeneous, so any single issue would be arbitrary rather than representative of a duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/Qwen-related, but one is about assistant masks in apply_chat_template and the other is wrong vision position ids; different bugs." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Both involve imports, but one is a trust_remote_code cache-path collision and the other is a missing utility symbol; not the same failure." + }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Different bugs: import latency vs unsafe sys.modules access in modeling_utils." + "reason": "Import-related symptoms differ: intermittent slow import vs unsafe sys.modules access; no shared code-path bug is evident." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both are model-loading failures, but for different models and likely different root causes." + "reason": "Both are model-loading failures, but they concern different models and different root causes." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Model loading regression vs broken auto_mappings/CI due to removed configs." - }, - { - "left": "issue:46082", - "right": "issue:46121", - "accept": false, - "reason": "Unrelated config-validation bugs: Llama head_dim handling vs rope-parameter type conversion." + "reason": "One is a model load regression; the other is stale auto_mappings references breaking CI. Different problems." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model/config.json." + "reason": "Both are config/model-loading issues, but one is missing fields under trust_remote_code and the other is an unrecognized model_type error." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI test flake in GraniteSpeech is not the same bug as a Phi-4 multimodal load failure." + "reason": "CI stale device override test failure is unrelated to the Phi-4 multimodal model loading issue." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "One is LlamaConfig head_dim validation; the other is rope-validation list conversion. Different validation paths." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Flash Attention support and embedding resize/tie_word_embedding are separate code paths." + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to token-embedding resize reinitialization." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Package-extras/backend dependency problem is not the same as a generic dependency issue." + "reason": "One is a missing TensorFlow backend in a torch install extra; the other is a general dependency issue without the same root cause." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Category-count overflow and Whisper empty-transcription generation failure are unrelated." + "reason": "Runtime category-limit error and Whisper empty-transcription failure are unrelated." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Different models and symptoms: untied lm_head in Mistral3 vs a Qwen3.5 save_pretrained regression." + "reason": "Different regressions: untied lm_head weights in Mistral3 vs incorrect save_pretrained checkpointing in Qwen3.5." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Specific missing Qwen2_5_VL export vs a broader AutoImageProcessor import failure." + "reason": "Both are import errors, but they reference different missing symbols and different modules." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE-marker handling and multimodal assistant mask generation are different bugs." + "reason": "Tokenizer BPE-marker handling is unrelated to multimodal assistant_mask generation." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to serve KV-cache tensor indexing." + "reason": "Tokenizer padding/crash in NER inputs is not the same code-path as serve KV-cache continuation indexing." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "The first is just a placeholder title; the second is an audio feature extraction bug." + "reason": "ADD THE DATA is not a substantive bug report and does not match the fbank feature issue." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Two distinct import errors affecting different symbols and modules." + "reason": "Different import failures: Qwen2.5 VL image processor vs layer_type_validation missing from configuration_utils." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer broadcast NoneType error is unrelated to output-embedding reinitialization." + "reason": "Trainer broadcast NoneType error is unrelated to embedding reinitialization after resize_token_embeddings." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Different multimodal config initialization/serialization issues with different models and fields." + "reason": "One is default vision/audio config initialization from None; the other is a missing deepstack_visual_indexes field dropped by strict parsing." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline image-processor lookup bug and notebook syntax error are unrelated." + "reason": "Video-classification pipeline using image processors is unrelated to a notebook syntax error." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is not the same as stale auto_mappings references breaking CI." + "reason": "Qwen3-MoE sliding_window behavior and removed Sam3LiteText auto_mappings entries are separate issues." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "Generation cache missing in chunked_prefill vs max_cache_length only applying to quantized cache are different cache semantics." + "reason": "Chunked-prefill cache-mixin regression is different from max_cache_length only applying to quantized cache." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "Reading position_ids after buffer registration is unrelated to CPU/CUDA output divergence." + "reason": "Registering position_ids as a buffer causing read errors is unrelated to CPU/CUDA output divergence." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error has no relation to _torch_extract_fbank_features." + "reason": "Notebook syntax error and _torch_extract_fbank_features() bug are unrelated." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version compatibility failure vs a specific model-loading regression at 5.1.0." + "reason": "Python version compatibility problem and a specific model load failure are not the same defect." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Processor/tokenizer regression and T5Gemma embedding-resize bug are different code paths." + "reason": "Regression in processor.tokenizer/image passing is unrelated to t5gemma resize_token_embeddings not updating decoder.embed_tokens." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Torch>=2.9 flex_attention failure and a version-check bug for AuxRequest import are not the same issue." + "reason": "Compiled flex_attention on torch>=2.9 and an incorrect version check for AuxRequest import are different code paths." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Unrecognized model/config loading and PIL backend torchvision dependency are unrelated." - }, - { - "left": "issue:43526", - "right": "issue:43901", - "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to a docs/behavior mismatch." - }, - { - "left": "issue:43381", - "right": "issue:45593", - "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and D-FINE auxiliary-loss omission are separate problems." + "reason": "Unrecognized model/model_type and PIL backend requiring torchvision are separate import/load issues." } ] }, @@ -11611,18 +11610,18 @@ "issue:42907", "issue:43295", "issue:43299", + "issue:43381", "issue:43441", + "issue:43526", "issue:43577", "issue:43650", "issue:43673", "issue:43824", - "issue:43854", "issue:43901", "issue:43976", "issue:44016", "issue:44062", "issue:44075", - "issue:44242", "issue:44279", "issue:44336", "issue:44351", @@ -11634,7 +11633,6 @@ "issue:44933", "issue:45020", "issue:45083", - "issue:45200", "issue:45216", "issue:45278", "issue:45335", @@ -11642,6 +11640,7 @@ "issue:45405", "issue:45584", "issue:45588", + "issue:45593", "issue:45759", "issue:45812", "issue:46082", @@ -11655,9 +11654,11 @@ "issue:46693" ], "soft_pairs": [ - "issue:42907|issue:43441", + "issue:43526|issue:43901", + "issue:43381|issue:45593", "issue:45083|issue:46506", "issue:43901|issue:44496", + "issue:42907|issue:43441", "issue:43976|issue:44279", "issue:43673|issue:44336", "issue:43441|issue:45588", @@ -11682,225 +11683,226 @@ "issue:43299|issue:45357", "issue:42796|issue:43650", "issue:41950|issue:45020", - "issue:44779|issue:45812", - "issue:43299|issue:43854", - "issue:44242|issue:45200" + "issue:44779|issue:45812" ], - "prepared_review_unit_hash": "855313a130f9d62127667796397ce43824600262a111e0c7f0aea67a4e9ca812", + "prepared_review_unit_hash": "1795f04033fdc3935b413467df27000dcdc197108fe61a9fa051ebe5f4fda8ef", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36630, - "estimated_input_tokens": 9158, - "estimated_eval_tokens": 18572 + "serialized_chars": 36628, + "estimated_input_tokens": 9157, + "estimated_eval_tokens": 18570 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is a loose grab-bag of Transformers bug reports and regressions, but none of the soft pairs look like the same underlying issue closely enough to merge.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue is safe: the reports span unrelated models, tokenizer behaviors, pipelines, docs, and environment/version regressions.", + "summary": "Most pairs are clearly different bugs touching different models, APIs, or subsystems. The only plausible duplicate is the import/export regression pair around `HybridCache` and the broader 4.57.0\u21925.5.0 import failures.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45278 is the broadest representative of the import/export regression cluster, and it plausibly subsumes specific missing-symbol reports like 44351.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the broadest umbrella report about recent-version breakage, but it is still not a duplicate of the rest.", + "best_issue_reason": "Issue 45278 is the strongest umbrella issue in this set because it describes the wider import breakage rather than a single symbol failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42907", - "right": "issue:43441", + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Different surfaces: Beit image processor label reduction vs pipeline docs mentioning `return_all_scores`." + }, + { + "left": "issue:43381", + "right": "issue:45593", "accept": false, - "reason": "Different failures: save/dequantization vs FlashAttention runtime crash on a different model family." + "reason": "Unrelated failures: gradient checkpointing in eval mode vs D-FINE auxiliary-loss behavior." }, { "left": "issue:45083", "right": "issue:46506", "accept": false, - "reason": "Different code paths and symptoms; one is a helper math bug, the other is CPU/CUDA output divergence." + "reason": "Different code paths and symptoms: Qwen3 Omni feature-length helper vs CPU/CUDA discrepancy in PPDocLayoutV3." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Docs mismatch for a pipeline vs model loading failure from an unrecognized config." + "reason": "Docs/API mismatch in text classification pipeline vs model-loading config error for an OLMo checkpoint." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family name, but different bugs: dequantized save failure vs FlashAttention crash." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Different problem class: Python version compatibility vs a dependency/package issue." + "reason": "General dependency/version complaint is too broad to equate with a specific Python support regression." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Generation cache bug and ANSI output formatting are unrelated." + "reason": "Completely different areas: generation cache behavior vs ANSI output in loading report." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both mention flash attention, but they target different models and different crash conditions." + "reason": "Both mention flash attention, but one is a Ministral model-specific failure and the other is an integration crash on `s_aux=None`." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Unrecognized model config vs missing text_config unwrapping for multimodal configs; not the same bug." + "reason": "Both are multimodal/model-loading related, but the concrete bugs differ: unrecognized model type vs `text_config` unwrapping in `from_config`." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error is unrelated to AddedToken argument handling." + "reason": "Notebook syntax error and tokenizer constructor argument conflict are unrelated." }, { "left": "issue:44792", "right": "issue:46424", "accept": false, - "reason": "A failed model test is not the same as cache_length being ignored except for one cache type." + "reason": "Janus image-generation test failure is unrelated to cache-length configuration handling." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Different import failures in different modules with different missing symbols." + "reason": "Both are import errors, but for different missing symbols/modules and no clear shared code-path." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Separate Qwen regressions: loading failure vs save_pretrained serialization bug." + "reason": "Same Qwen3.5/Qwen3VL family, but one is model loading and the other is `save_pretrained` checkpoint serialization." }, { "left": "issue:46123", "right": "issue:46369", "accept": false, - "reason": "Partial-batch bookkeeping bug vs memory overflow in a different model." + "reason": "Partial-batch mask generation bug and SAM memory overflow are different failures." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "Generic dependency issue vs torch-version incompatibility; not the same root cause." + "reason": "Generic dependency issue is not the same as a specific torch 2.4.1 incompatibility." }, { "left": "issue:44521", "right": "issue:46693", "accept": false, - "reason": "Multimodal chat-template mask bug and long-context FA2 performance regression are different problems." + "reason": "Assistant-mask generation for multimodal chat templates is unrelated to FA2 long-context performance regression." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Embedding resize behavior is unrelated to Whisper empty-transcription generation failure." + "reason": "Different models and behaviors: T5Gemma embedding resize vs Whisper empty transcription generation." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "Config validation error vs kernel registration crash; different code paths." + "reason": "Llama config validation and Qwen3.5 kernel registration crash are different bugs." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Dependency issue and unreleased PEFT minimum version are unrelated." + "reason": "A broad dependency problem is not the same as the specific unreleased PEFT version bump." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype propagation bug is unrelated to Qwen3.5 checkpoint serialization." + "reason": "Dtype handling in Blip2 loading and Qwen3.5 save_pretrained serialization are unrelated." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Different documentation issues for different APIs and models." + "reason": "Different documentation problems with different root causes: SigLIP2 example errors vs stale pipeline docs." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Multimodal processor/tokenizer regression vs Deepseek tokenizer regression; same era, different bug." + "reason": "Custom processor/tokenizer API regression and Deepseek tokenizer output regression are different issues." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "Optimizer args ignored and tie_word_embeddings warning are unrelated." + "reason": "SGD optimizer args being ignored is unrelated to a tied-embeddings warning." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python support mismatch is not the same as a PEFT version bump." + "reason": "Python support mismatch is not the same as a PEFT minimum-version packaging issue." }, { "left": "issue:46082", "right": "issue:46512", "accept": false, - "reason": "Config validation and stale config.dtype after from_config are distinct issues." + "reason": "One is config validation for `head_dim`; the other is stale `model.config.dtype` after `from_config`." }, { "left": "issue:44351", "right": "issue:45278", - "accept": false, - "reason": "One import error is a specific symptom; the other is a broad version-upgrade report without confirmed same root cause." + "accept": true, + "reason": "`HybridCache` import failure looks like a specific instance of the broader 4.57.0\u21925.5.0 import/export breakage described in 45278." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 issues: model loading vs incorrect visual-encoder keys on save." + "reason": "Both involve Qwen3.5, but loading a composite config and saving incorrect visual encoder keys are separate bugs." }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "Tokenizer BPE-marker bug has nothing to do with the non-technical 'ADD THE DATA' report." + "reason": "Tokenizer BPE-marker handling and a placeholder issue titled `ADD THE DATA` are not comparable duplicate candidates." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Pipeline processor lookup bug vs remote_code model breakage; only broad ecosystem overlap." + "reason": "Video-classification processor lookup and remote-code breakage are unrelated." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but for different model families and likely different fixes." - }, - { - "left": "issue:43299", - "right": "issue:43854", - "accept": false, - "reason": "Different multimodal model-loading failures with no concrete shared code path." - }, - { - "left": "issue:44242", - "right": "issue:45200", - "accept": false, - "reason": "Load-balancing loss bug is unrelated to mm_token_type_ids defaults for text-only fine-tuning." + "reason": "Two tokenizer regressions, but for different model families and likely different code paths." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45278 is clearly the broad umbrella import-regression issue, and 44351 is a specific HybridCache import failure that plausibly fits it. The other soft-edge rejections are conservative and tied to distinct titles/symptoms. No overstatement stands out." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -11921,6 +11923,7 @@ "issue:41950", "issue:42175", "issue:42994", + "issue:43299", "issue:43317", "issue:43441", "issue:43479", @@ -11928,14 +11931,13 @@ "issue:43650", "issue:43653", "issue:43673", - "issue:43723", "issue:43761", - "issue:43824", + "issue:43854", "issue:43901", "issue:43976", "issue:43994", "issue:44016", - "issue:44162", + "issue:44242", "issue:44246", "issue:44261", "issue:44263", @@ -11950,6 +11952,7 @@ "issue:44779", "issue:45081", "issue:45103", + "issue:45200", "issue:45276", "issue:45372", "issue:45397", @@ -11962,11 +11965,11 @@ "issue:45823", "issue:45834", "issue:45907", - "issue:46209", - "issue:46291", - "issue:46399" + "issue:46209" ], "soft_pairs": [ + "issue:43299|issue:43854", + "issue:44242|issue:45200", "issue:43479|issue:43525", "issue:44466|issue:45276", "issue:41950|issue:45823", @@ -11981,8 +11984,8 @@ "issue:42994|issue:43317", "issue:33290|issue:45656", "issue:45081|issue:45907", - "issue:45600|issue:45712", "issue:44261|issue:44263", + "issue:45600|issue:45712", "issue:44623|issue:44749", "issue:42175|issue:43976", "issue:44671|issue:44779", @@ -11990,202 +11993,199 @@ "issue:44623|issue:45397", "issue:43441|issue:44683", "issue:43901|issue:44246", - "issue:43673|issue:43761", "issue:45542|issue:45800", - "issue:46291|issue:46399", - "issue:43723|issue:44162", - "issue:43824|issue:43976" + "issue:43673|issue:43761" ], - "prepared_review_unit_hash": "c9ca9fabd9e0feb46c9fc315fb1266466a95fc3c6194db7f37a404de3fcd3ce4", + "prepared_review_unit_hash": "3068fc3fc8b799cb04769cd0da0a7013dc9c1b0915d8bf0c5cdcd4d3d49a2df4", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36130, - "estimated_input_tokens": 9033, - "estimated_eval_tokens": 18322 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35201, + "estimated_input_tokens": 8801, + "estimated_eval_tokens": 17858 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items span unrelated tokenizer, config, model, pipeline, dependency, and performance regressions. None of the soft pairs look like the same underlying bug, so no duplicate merge is justified.", - "confidence": 0.96, - "canonical_issue_reason": "No true canonical issue stands out because the candidate set is not one duplicate cluster; the soft pairs are mostly broad-similarity false positives across different subsystems and models.", + "summary": "No duplicate cluster emerges here: the soft pairs are mostly superficial matches across version regressions, model families, or broad subsystems rather than the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single best issue for deduping purposes; the artifacts do not share a concrete bug or fix target strongly enough to justify choosing a representative canonical issue.", + "best_issue_reason": "issue:44779 is the strongest standalone bug report in the set: it is a concrete regression with a clear v4\u2192v5 behavior change and a specific tokenizer/model impact.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different failures: Qwen3VL MoE loading regression vs a GLM-4.7-Flash unit-test/model-load issue; no shared code path is evident." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "One is missing load-balancing loss when router logits are off; the other is Gemma 4 text-only fine-tuning needing default mm_token_type_ids. Different mechanisms." + }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Different model-config bugs: Phi4 multimodal default subconfig initialization vs Llama4 missing pad_token_id." + "reason": "Both are config-related, but one is bad default vision/audio config initialization and the other is missing pad_token_id on Llama4Config. Not the same bug." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Both touch weights, but one is device-dependent serialization and the other is resize_token_embeddings propagation; different code paths." + "reason": "Serialization of tied lm_head weights vs resize_token_embeddings not updating Gemma4 embeddings/output layers. Different operations and symptoms." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Pipeline processor lookup bug vs Gemma4 forward-time device mismatch; unrelated failures." + "reason": "Video-classification pipeline processor lookup and Gemma4 PLE device mismatch are unrelated issues." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "Tokenizer special-token registration issue vs D-FINE auxiliary-loss logic; no shared underlying bug." + "reason": "Tokenizer special-token registration/empty decode is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Torch flex_attention compile failure vs missing TensorFlow backend dependency; unrelated." + "reason": "Compiled flex_attention on torch>=2.9 vs tensorboard-only backend import issue; no common code path." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "Placeholder issue vs FP8 inference failure on B200; not the same change or bug." + "reason": "A vague data request is unrelated to FP8 inference failing on B200." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error vs init_empty_weights unexpected-argument TypeError; different layers." + "reason": "Notebook syntax error is unrelated to init_empty_weights passing unexpected _is_hf_initialized." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is a different tokenizer-loading bug from the Mistral regex patch crash." + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are both tokenizer-related but distinct bugs in different paths." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "Ministral FlashAttention regression vs Kosmos2.5 OCR index error; unrelated model-specific issues." + "reason": "Ministral FlashAttention failure and Kosmos2.5 OCR long-input index error are different model-specific problems." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer task AttributeError vs D-FINE denoising/aux-loss behavior; no overlap." + "reason": "MLukeTokenizer AttributeError is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotation handling vs Gemma4 processor import dependency; different subsystems and fixes." + "reason": "Auto-docstring annotation handling and a mistral_common import error affect different subsystems." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized-model save failure vs dequantized-model load/offload failure; related theme but different stages and bugs." + "reason": "Quantized-model saving failure and dequantized offload loading failure are both quantization-adjacent but not the same concrete bug." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Adafactor OOM in DeepSpeed vs optimizer.step being invoked twice; distinct optimizer/deepspeed issues." + "reason": "AdaFactor OOM with DeepSpeed and optimizer.step being called twice with DeepSpeed are different optimizer/engine issues." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex crash vs list_repo_files revision forwarding; unrelated API bugs." + "reason": "Tokenizer regex crash and list_repo_files not forwarding revision are unrelated." }, { - "left": "issue:45600", - "right": "issue:45712", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "Sam3Lite auto-mapping cleanup vs leftover dummy classes check failure; different maintenance problems." + "reason": "MLA q_a_layernorm config precision issue and GlmMoeDsaIndexer torch.split return values are different model internals." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:45600", + "right": "issue:45712", "accept": false, - "reason": "MLA q_a_layernorm epsilon issue vs torch.split return-value issue in GlmMoeDsaIndexer; unrelated." + "reason": "auto_mappings cleanup/CI breakage and leftover dummy_pt_objects classes are separate maintenance issues." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "processor.save_pretrained missing files vs Chinese report of slow filtering after upgrade; not the same bug." + "reason": "processor.save_pretrained missing files and a Chinese performance regression report do not describe the same bug." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "TensorFlow backend packaging bug vs Python version compatibility regression; different environment issues." + "reason": "Missing TensorFlow in the torch extra and Python version support problems are unrelated packaging/environment issues." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM regression vs DeepSeek tokenizer regression; different models and code paths." + "reason": "Both are v5 regressions, but CamemBERT masked-LM output issues and Deepseek tokenizer correctness are different code paths and models." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs drift around return_all_scores is unrelated to SigLIP2 producing nonsensical outputs." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing save_pretrained files vs Gemma-4 zero3 from_pretrained; unrelated." + "reason": "Processor save_pretrained missing files and Gemma-4 zero3 from_pretrained are different behaviors." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "FlashAttention regression on a model vs compiled flex_attention failure on torch>=2.9; different attention implementations." + "reason": "Both mention attention, but Ministral FlashAttention failure and compiled flex_attention failure on torch>=2.9 are distinct problems." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs/behavior mismatch for return_all_scores vs intermittent slow import; not the same bug." - }, - { - "left": "issue:43673", - "right": "issue:43761", - "accept": false, - "reason": "Chunked-prefill cache regression vs CLIPVision hidden_states regression; unrelated." + "reason": "Documentation mismatch and slow import behavior are unrelated." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "Missing TensorFlow backend vs torch/transformers version incompatibility; different dependency problems." - }, - { - "left": "issue:46291", - "right": "issue:46399", - "accept": false, - "reason": "kernels dependency/version import failure vs Qwen3.5 rotary-pos-emb registration crash; distinct bugs." - }, - { - "left": "issue:43723", - "right": "issue:44162", - "accept": false, - "reason": "AutoTokenizer loading regression vs ESM2 broken workflow; unrelated." + "reason": "Tensorboard/TensorFlow backend mis-installation is unrelated to torch 2.4.1 vs transformers 5.8.0 incompatibility." }, { - "left": "issue:43824", - "right": "issue:43976", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "Missing Qwen2_5_VL import vs Python 3.9/3.10 compatibility problem; not the same issue." + "reason": "Cache-missing during chunked_prefill and CLIPVision hidden_states=None are different regression reports." } ] }, @@ -12209,11 +12209,9 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ - "issue:36246", "issue:41628", "issue:41950", "issue:42491", - "issue:42757", "issue:42886", "issue:43299", "issue:43525", @@ -12221,10 +12219,13 @@ "issue:43531", "issue:43550", "issue:43643", + "issue:43723", "issue:43784", + "issue:43824", "issue:43881", "issue:43976", "issue:44016", + "issue:44162", "issue:44188", "issue:44220", "issue:44242", @@ -12232,16 +12233,13 @@ "issue:44295", "issue:44297", "issue:44336", - "issue:44351", "issue:44373", "issue:44485", "issue:44496", "issue:44589", - "issue:44779", "issue:44871", "issue:45003", "issue:45103", - "issue:45292", "issue:45362", "issue:45397", "issue:45542", @@ -12250,6 +12248,7 @@ "issue:45593", "issue:45823", "issue:46132", + "issue:46291", "issue:46399", "issue:46424", "issue:46506", @@ -12259,13 +12258,16 @@ "issue:46726" ], "soft_pairs": [ + "issue:43723|issue:44162", + "issue:46291|issue:46399", + "issue:43824|issue:43976", "issue:43525|issue:43881", "issue:44016|issue:45823", "issue:43525|issue:43643", "issue:44589|issue:45362", "issue:44188|issue:44336", - "issue:46655|issue:46726", "issue:44336|issue:44373", + "issue:46655|issue:46726", "issue:41628|issue:43976", "issue:43784|issue:45542", "issue:44246|issue:44496", @@ -12273,9 +12275,9 @@ "issue:45561|issue:45563", "issue:46614|issue:46665", "issue:45103|issue:46399", - "issue:42491|issue:43299", - "issue:44242|issue:45593", "issue:43526|issue:45003", + "issue:44242|issue:45593", + "issue:42491|issue:43299", "issue:43531|issue:45362", "issue:44246|issue:44871", "issue:44336|issue:44485", @@ -12283,207 +12285,204 @@ "issue:44336|issue:46424", "issue:42886|issue:46132", "issue:43550|issue:46506", - "issue:36246|issue:42757", - "issue:41950|issue:44220", - "issue:44779|issue:45292", - "issue:36246|issue:44351" + "issue:41950|issue:44220" ], - "prepared_review_unit_hash": "094c089fc9cc959fefd1501c94c6b49457ca56613547473c5cd7bb829231b116", + "prepared_review_unit_hash": "6052b48cf71ae98994277d8f4c4e10836ee45e6fa0388e8e02671e9555d010f0", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35961, - "estimated_input_tokens": 8991, - "estimated_eval_tokens": 18238 + "serialized_chars": 35355, + "estimated_input_tokens": 8839, + "estimated_eval_tokens": 17934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose grab-bag of Transformer issues with only superficial similarity (mostly import/load failures). None of the soft pairs appear to be the same underlying bug or a single mergeable fix.", - "confidence": 0.84, - "canonical_issue_reason": "issue:41628 is the broadest representative of the import-error subset, but the cluster is not a tight duplicate set.", + "summary": "These issues are largely unrelated Transformers bugs across different models, tokenizers, processors, and runtime paths. None of the soft pairs look like the same underlying bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.83, + "canonical_issue_reason": "41628 is the cleanest, most concrete regression report in this set: a straightforward top-level import failure with a clear user-facing symptom. It\u2019s a good anchor for the processor/import-related subset, even though the cluster is not truly duplicate-heavy.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41628 is the most generic and reusable representative among these heterogeneous reports, especially the import-related ones.", + "best_issue_reason": "41628 is the best single issue candidate because it is specific, actionable, and easy to triage; the failure mode is crisp and well-scoped.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading regression vs a broad ESM2 workflow breakage; different model/path and no shared concrete bug." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both mention kernels, but one is a pipelines import/version constraint and the other is a Qwen3.5 registration crash; different code paths." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Specific missing model export vs a Python-version compatibility problem; not the same bug." + }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Different failures: one is a missing Llama config field, the other is a model loading failure for a specific vision model." + "reason": "Llama4Config missing field vs glm-4v loading failure; different models and failure mechanisms." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Unrelated problems: notebook syntax error vs device placement mismatch in Gemma4 forward." + "reason": "Notebook syntax error vs Gemma4 device-mapping runtime mismatch; unrelated." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "One is a config attribute error; the other is trust_remote_code returning incomplete AutoConfig fields." + "reason": "Missing pad_token_id on a specific config vs missing fields from trust_remote_code AutoConfig; related in theme only, not the same defect." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Different code paths and symptoms: missing Float8 storage vs Qwen3.5 chat crash." + "reason": "Storage class import/serialization issue vs Qwen3.5 chat crash; different bug surfaces." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence on torch.compile is unrelated to ANSI codes in loading_report." + "reason": "torch.compile attention-kernel divergence vs terminal ANSI output in loading_report; unrelated." }, { - "left": "issue:46655", - "right": "issue:46726", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Both involve vision-related errors, but one is a SmolVLM video padding bug and the other is AutoProcessor backend-missing behavior." + "reason": "ANSI escape generation bug vs a wrong docstring; not the same code issue." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:46655", + "right": "issue:46726", "accept": false, - "reason": "Loading-report ANSI output is unrelated to a docstring mistake for position_ids." + "reason": "SmolVLM video padding calculation bug vs AutoProcessor backend-missing import/reporting issue; different processor failures." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Both mention imports/versioning, but one is a missing AutoImageProcessor import and the other is a Python-version compatibility issue." + "reason": "Top-level import failure for AutoImageProcessor vs Python version support issue; unrelated." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "One is a missing nn import in sentence-transformers integration; the other is a TensorBoard/TensorFlow backend initialization bug." + "reason": "Missing nn import in sentence-transformers integration vs undefined tf backend from tensorboard-only install; different packages and causes." }, { - "left": "issue:44242", + "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Load-balancing loss with router logits is unrelated to unrecognized model loading for an Olmo Hybrid model." + "reason": "Import slow-path performance issue vs unrecognized model/config loading failure; not the same bug." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer class serialization mismatch is a different issue from gemma-4 zero3 from_pretrained behavior." + "reason": "tokenizer.save_pretrained metadata mismatch vs Gemma-4 zero3 from_pretrained failure; different behaviors." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "Test-worker file races and stale paged generate warnings are unrelated." + "reason": "pytest-xdist race condition vs stale generate() warning; unrelated test/runtime issues." }, { "left": "issue:46614", "right": "issue:46665", "accept": false, - "reason": "A stale phonemizer backend after language switching is unrelated to from_pretrained behavior inside Python classes." + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained behavior inside Python classes; too different to be duplicates." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Both involve decorators/introspection, but one is auto-docstring annotation handling and the other is Qwen3.5 kernel registration missing a decorator." + "reason": "auto_docstring crash under future annotations vs kernels registration error in Qwen3.5; unrelated." }, { - "left": "issue:42491", - "right": "issue:43299", + "left": "issue:43526", + "right": "issue:45003", "accept": false, - "reason": "Same model family, but one is hf4.x LoRA compatibility and the other is sliding_window loading on v5; not the same bug." + "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access; different modules and symptoms." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Different training losses: router load-balancing loss vs D-FINE auxiliary losses when denoising is off." + "reason": "MoE load-balancing loss not added vs D-FINE auxiliary losses disabled; both are loss-related but on different models/code paths." }, { - "left": "issue:43526", - "right": "issue:45003", + "left": "issue:42491", + "right": "issue:43299", "accept": false, - "reason": "BEiT label reduction bug is unrelated to unsafe sys.modules access in modeling_utils." + "reason": "Both are Qwen MoE loading regressions, but they involve different model families/version contexts and not clearly the same concrete bug." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Qwen3-MoE sliding_window logic is unrelated to Qwen3.5 chat crashes." + "reason": "Qwen3-MoE sliding_window issue vs Qwen3.5 chat crash; not the same failure mode." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import latency and eos_token_id config inconsistency are unrelated." + "reason": "Import performance issue vs Gemma-3 eos_token_id inconsistency; unrelated." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI rendering in loading_report is unrelated to GLM-5 RoPE implementation discussion." + "reason": "ANSI output bug vs GLM-5 RoPE implementation discussion; different areas." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Reading a registered buffer is a different problem from gemma-4 zero3 from_pretrained." + "reason": "position_ids buffer read error vs Gemma-4 zero3 loading problem; not the same defect." }, { "left": "issue:44336", "right": "issue:46424", "accept": false, - "reason": "Terminal ANSI codes and cache max length handling are unrelated." + "reason": "ANSI codes in loading_report vs cache max_cache_length behavior; unrelated." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline cache loading failure is unrelated to AttentionInterface.register changing function behavior." + "reason": "Offline tokenizer cache loading failure vs AttentionInterface.register behavior change; different mechanisms." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "torch.compile SDPA failure in Bamba-9B-v2 is unrelated to CPU/CUDA divergence in PPDocLayoutV3." - }, - { - "left": "issue:36246", - "right": "issue:42757", - "accept": false, - "reason": "Both are import errors, but they concern different symbols from different packages and different root causes." + "reason": "torch.compile+SDPA failure in Bamba vs CPU/CUDA result mismatch in PPDocLayoutV3; different models and bugs." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification using image processors is unrelated to _torch_extract_fbank_features()." - }, - { - "left": "issue:44779", - "right": "issue:45292", - "accept": false, - "reason": "Deepseek tokenizer regression is unrelated to resize_token_embeddings not updating output embeddings." - }, - { - "left": "issue:36246", - "right": "issue:44351", - "accept": false, - "reason": "Both are import-name errors, but they involve different missing exports (Qwen image processor vs HybridCache) and distinct fixes." + "reason": "Video-classification processor lookup bug vs _torch_extract_fbank_features issue; different pipelines and code paths." } ] }, @@ -12508,6 +12507,7 @@ "cluster_id": "cluster-10000-479", "nodes": [ "issue:36246", + "issue:42757", "issue:43257", "issue:43262", "issue:43295", @@ -12518,12 +12518,11 @@ "issue:43673", "issue:43901", "issue:44038", - "issue:44060", - "issue:44077", "issue:44162", "issue:44242", "issue:44246", "issue:44295", + "issue:44351", "issue:44466", "issue:44479", "issue:44496", @@ -12531,6 +12530,7 @@ "issue:44623", "issue:44734", "issue:44749", + "issue:44779", "issue:44857", "issue:44877", "issue:44987", @@ -12556,6 +12556,9 @@ "issue:46726" ], "soft_pairs": [ + "issue:44779|issue:45292", + "issue:36246|issue:42757", + "issue:36246|issue:44351", "issue:43901|issue:45003", "issue:43257|issue:45137", "issue:45003|issue:45561", @@ -12566,221 +12569,232 @@ "issue:43673|issue:44561", "issue:44162|issue:45200", "issue:44479|issue:46655", + "issue:43262|issue:44246", "issue:45375|issue:45759", "issue:43475|issue:46369", - "issue:43262|issue:44246", "issue:44242|issue:44857", "issue:44295|issue:44749", "issue:45137|issue:45600", + "issue:44877|issue:45375", "issue:36246|issue:43596", "issue:45292|issue:45584", - "issue:44877|issue:45375", "issue:43295|issue:45292", "issue:44561|issue:46399", "issue:44466|issue:45335", "issue:44623|issue:46726", "issue:44987|issue:45459", "issue:44038|issue:46506", - "issue:43352|issue:45250", "issue:44479|issue:46170", - "issue:44060|issue:44077" + "issue:43352|issue:45250" ], - "prepared_review_unit_hash": "b8df5edda52ac48964ecdc75c662af4f8d3328e3c273cff8e99831c6ec75592c", + "prepared_review_unit_hash": "28aed8c2dcde9c5b715a50887977c35d92a2418aed66cbbb9fa2b176ce54160a", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35662, - "estimated_input_tokens": 8916, - "estimated_eval_tokens": 18088 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36758, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous; the soft pairs share broad themes like video processors, DeepSpeed, or Qwen models, but not the same concrete bug. I would not merge any of the proposed soft edges. If a single issue must act as the closest anchor, issue 44479 is the best representative of the clearest subgroup (video-input regression across multiple Qwen variants).", - "confidence": 0.93, - "canonical_issue_reason": "Issue 44479 is the clearest canonical-style anchor: it describes a broad, version-specific video-input regression across multiple related model families, with a concrete symptom and scope.", + "summary": "The items are mostly heterogeneous and do not form a strong duplicate cluster. The soft matches are mostly superficial topical overlap (imports, cache, Qwen, DeepSpeed, video, embeddings) rather than the same concrete bug.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 36246 is the most concrete, self-contained bug report in the set: a specific missing export/import path with a clear failure mode and likely fix surface. The rest of the issues are broader or on different code paths.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the most suitable representative among these issues because it is the broadest, most explicit regression report and is easiest to use as a duplicate target for nearby video-input reports.", + "best_issue_reason": "Issue 36246 is the best representative issue because it is narrowly scoped, reproducible, and actionable compared with the more ambiguous or broad reports in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Both mention transformer behavior regressions, but one is a DeepSeek tokenizer correctness issue and the other is resize_token_embeddings propagation; different code paths and failures." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for unrelated symbols and dependencies: a missing transformers vision processor export vs huggingface_hub.is_offline_mode." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing-name import errors, but they concern different exported objects and likely different fixes." + }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for TextClassificationPipeline vs unsafe sys.modules access in modeling_utils; unrelated bugs." + "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access in modeling_utils." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Both involve DeepSpeed, but one is Qwen3 MOE weight conversion and the other is ZeRO3 deque/index failures; different failure modes." + "reason": "Both involve DeepSpeed, but one is Qwen3 MoE weight conversion and the other is a ZeRO3 deque/indexing failure; not the same bug." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to pytest-xdist race conditions in testing utils." + "reason": "Unsafe sys.modules access and pytest-xdist file races are distinct issues in different subsystems." }, { "left": "issue:44734", "right": "issue:46424", "accept": false, - "reason": "Both mention cache, but one is serve/KV-cache continuation indexing and the other is cache length semantics for quantized cache; not the same bug." + "reason": "Both mention KV/cache behavior, but one is a serve continuation indexing crash and the other is cache length config handling; different concrete failures." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model/config loading vs sys.modules access; no shared code-path bug." + "reason": "Unrecognized model/config loading and unsafe sys.modules access are unrelated." }, { "left": "issue:43540", "right": "issue:46655", "accept": false, - "reason": "Both are video-input issues, but they hit different models and different reported causes; too broad to treat as one bug." + "reason": "Both are video-processing bugs, but they affect different models and different failure points." }, { "left": "issue:45859", "right": "issue:46399", "accept": false, - "reason": "Qwen3_5Moe CPU-offload incompatibility is a different issue from use_kernels registration/decorator failure." + "reason": "Both are Qwen3.5-related, but one is CPU offload incompatibility and the other is a registration/decorator crash with use_kernels=True." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Chunked-prefill cache missing and trust_remote_code breakage from removed helper are unrelated." + "reason": "GenerationMixin chunked_prefill cache missing and removal of is_torch_fx_available are unrelated regressions." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids handling are separate model-specific issues." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids handling are unrelated model-specific bugs." }, { "left": "issue:44479", "right": "issue:46655", "accept": false, - "reason": "Both are video regressions, but one is a broad Qwen-family v5.3.0 regression and the other is a SmolVLM padding bug; different underlying code paths." + "reason": "Both are video-input regressions, but they target different models and likely different preprocessing code paths." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat template sample-rate defaulting and import performance have no substantive overlap." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Both are Qwen multimodal config problems, but one is a missing vision config field and the other is from_config text_config unwrapping; distinct bugs." + "reason": "Both concern Qwen multimodal config handling, but one is a missing config field and the other is failure to unwrap text_config; different root causes." }, { "left": "issue:43475", "right": "issue:46369", "accept": false, - "reason": "SAM3 vision encoder attribute error and SAM3-litetext memory overflow are not the same failure." - }, - { - "left": "issue:43262", - "right": "issue:44246", - "accept": false, - "reason": "Audio chat-template sampling-rate default and slow import are unrelated." + "reason": "SAM 3 Video output attribute loss and SAM3 LiteText memory overflow are different model bugs." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Mixture-of-experts load-balancing loss omission and LwDetrImageLoss AMP crash are different subsystems and symptoms." + "reason": "Load-balancing loss logic and an AMP/CUDA crash in LwDetrImageLoss are unrelated." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Position_ids buffer access error and a Chinese report of slower filtering after upgrade are not the same bug." + "reason": "A position_ids buffer access error and a Chinese performance regression are not the same bug." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "ZeRO3 deque failure and auto_mappings references to removed configs are unrelated." + "reason": "ZeRO3 deque underflow and broken auto_mappings references are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict config loading for granite_speech and missing Qwen visual indexes are different config-validation issues." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Qwen2.5-VL import error and BertModel ZeRO3 init index error are unrelated." + "reason": "Importing a missing processor class and a DeepSpeed zero3/BertModel index error are unrelated." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Token embedding resize propagation and Whisper empty-transcription generation failure are different issues." - }, - { - "left": "issue:44877", - "right": "issue:45375", - "accept": false, - "reason": "Strict config loading for granite_speech and a missing Qwen3_5MoeVisionConfig field are distinct config bugs." + "reason": "resize_token_embeddings propagation and Whisper empty-transcription handling are different model behaviors." }, { "left": "issue:43295", "right": "issue:45292", "accept": false, - "reason": "Processor.tokenizer/image passing regression is unrelated to output embedding resize propagation." + "reason": "A regression in processor.tokenizer/image passing is unrelated to token embedding resizing behavior." }, { "left": "issue:44561", "right": "issue:46399", "accept": false, - "reason": "Removed is_torch_fx_available breaks trust_remote_code models; the Qwen3.5 kernel crash is a separate registration issue." + "reason": "Removal of an FX helper breaking trust_remote_code models is unrelated to Qwen3.5 rotary embedding registration." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Both concern embedding weights, but one is device-dependent serialization and the other is resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Both involve embeddings/weights, but one is lm_head serialization and the other is decoder.embed_tokens not updating after resize." }, { "left": "issue:44623", "right": "issue:46726", "accept": false, - "reason": "Missing files on processor.save_pretrained and unrecognized image processor when backend is absent are different processor issues." + "reason": "processor.save_pretrained file omissions and AutoProcessor failing without a vision backend are different processor-loading/saving issues." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model loading failure for physical-intelligence/fast and tokenizer error masking from protobuf handling are unrelated." + "reason": "A model-loading failure for physical-intelligence/fast and tokenizer error masking when protobuf is missing are unrelated." }, { "left": "issue:44038", "right": "issue:46506", "accept": false, - "reason": "Qwen3-VL-Moe regression and PPDocLayoutV3 CPU/CUDA divergence do not describe the same bug." - }, - { - "left": "issue:43352", - "right": "issue:45250", - "accept": false, - "reason": "A model not supporting Flash Attention 2.0 and a generic Flash Attention issue are too broad and not the same concrete defect." + "reason": "Qwen3-VL-Moe loading bug and CPU/CUDA result divergence in PPDocLayoutV3 are unrelated." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Video-input regression in Qwen models and Windows ROCm torch.distributed import/type errors are unrelated." + "reason": "Video-input regression in Qwen/VL models and Windows ROCm torch.distributed import issues are unrelated." }, { - "left": "issue:44060", - "right": "issue:44077", + "left": "issue:43352", + "right": "issue:45250", "accept": false, - "reason": "Incorrect tied-weights warning in Qwen3-Next and PatchTSMixer post_init policy are unrelated model/config issues." + "reason": "A model-specific Flash Attention 2.0 unsupported error is not the same as a generic Flash Attention 2.0 issue title." } ] }, @@ -12805,21 +12819,21 @@ "cluster_id": "cluster-10000-479", "nodes": [ "issue:42886", - "issue:42994", "issue:43065", "issue:43232", "issue:43408", "issue:43475", - "issue:43482", "issue:43540", - "issue:43643", "issue:43646", "issue:43653", "issue:43673", + "issue:43701", "issue:43723", "issue:43906", "issue:43931", "issue:43994", + "issue:44060", + "issue:44077", "issue:44162", "issue:44188", "issue:44297", @@ -12834,7 +12848,6 @@ "issue:44995", "issue:45200", "issue:45229", - "issue:45310", "issue:45397", "issue:45412", "issue:45563", @@ -12854,20 +12867,21 @@ "issue:46726" ], "soft_pairs": [ + "issue:44060|issue:44077", "issue:45412|issue:46123", - "issue:44368|issue:44977", "issue:43653|issue:44162", - "issue:44995|issue:45200", + "issue:44368|issue:44977", "issue:44877|issue:46170", + "issue:44995|issue:45200", "issue:43475|issue:43540", "issue:42886|issue:45229", "issue:44368|issue:45563", "issue:43931|issue:44368", "issue:43065|issue:43408", - "issue:44995|issue:45588", "issue:45397|issue:45656", - "issue:43723|issue:46710", + "issue:44995|issue:45588", "issue:45987|issue:46082", + "issue:43723|issue:46710", "issue:44188|issue:46202", "issue:43906|issue:46202", "issue:44483|issue:46202", @@ -12879,234 +12893,219 @@ "issue:44297|issue:44623", "issue:43994|issue:46726", "issue:43646|issue:44162", - "issue:44871|issue:46132", "issue:44749|issue:45656", + "issue:44871|issue:46132", "issue:43232|issue:43673", - "issue:44749|issue:45310", - "issue:42994|issue:43482", - "issue:44623|issue:45656", - "issue:43475|issue:43643" + "issue:43701|issue:44749" ], - "prepared_review_unit_hash": "2819099ce69046b127248b9693db0a3b0c81c5de03aefb0017e46dd9adaa6633", + "prepared_review_unit_hash": "ad5ee9157271483980af2ee70f48fff82238388b826dfd5d20d956caf81bcef3", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37238, - "estimated_input_tokens": 9310, - "estimated_eval_tokens": 18876 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36117, + "estimated_input_tokens": 9030, + "estimated_eval_tokens": 18316 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated closed issues spanning tokenizers, generation, model loading, multimodal processors, and backend-specific bugs. I do not see a true duplicate cluster here, so all soft edges should be rejected.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "The cluster is heterogeneous: tokenizer/serialization, generation/cache, model-loading, vision, and deepspeed issues are all mixed together. None of the soft pairs look like the same underlying bug or change, so they should not be merged.", + "confidence": 0.82, + "canonical_issue_reason": "issue:44297 is the most concrete, self-contained serialization bug in the set, and it is the closest to another save_pretrained-style report.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:44297 has the clearest failure mode and the most focused code path, making it the best representative issue here.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45412", - "right": "issue:46123", + "left": "issue:44060", + "right": "issue:44077", "accept": false, - "reason": "RT-DETR memory cleanup and MaskGenerationPipeline batch-finalization are different bugs in different code paths." + "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init policy are unrelated subsystems and failure modes." }, { - "left": "issue:44368", - "right": "issue:44977", + "left": "issue:45412", + "right": "issue:46123", "accept": false, - "reason": "A tie_word_embeddings warning and a flash-attention generation failure are unrelated symptoms and subsystems." + "reason": "RT-DETR memory retention and MaskGenerationPipeline partial-batch dropping are different runtime behaviors." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are distinct tokenizer vs model issues." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated tokenizer/model bugs." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "A stale indexer-cache crash on second forward and mm_token_type_ids defaulting are separate model behaviors." + "reason": "A tie_word_embeddings warning for Qwen3.5 is not the same bug as flash-attention generation failure." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to the Windows ROCm torch.distributed import failure." + "reason": "Strict config loading for granite_speech and Windows ROCm import errors have no shared code path." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale indexer cache crash and a Gemma4 mm_token_type_ids default issue are different defects." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Missing fpn_position_embeddings in SAM3 video output and Qwen3OmniMoe video-processing errors are different bugs." + "reason": "Missing fpn_position_embeddings in SAM3 and Qwen3OmniMoe video input validation are distinct bugs." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are not the same underlying problem." + "reason": "Tokenizer offline cache loading and multi-GPU CUDA OOM are unrelated problems." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "A config warning about tie_word_embeddings does not match the paged generate stale num_return_sequences warning." + "reason": "A model config warning and a paged-generate stale warning for num_return_sequences are not the same issue." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and a tie_word_embeddings warning are unrelated failure modes." + "reason": "Qwen3-VL weight-shape mismatch is unrelated to the Qwen3.5 tie-word-embeddings warning." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and a Sam3Tracker model-type warning are different SAM3 issues." - }, - { - "left": "issue:44995", - "right": "issue:45588", - "accept": false, - "reason": "Stale indexer-cache crashes and flash_attention.py crashing on s_aux=None are separate code-path bugs." + "reason": "Dummy Conv2d in Sam3PixelDecoder and SAM3 model-type mismatch warning are different SAM3 issues." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Gemma-4 ZeRO3 loading and deepspeed double optimizer stepping are not the same defect." + "reason": "gemma-4 from_pretrained/zero3 loading and double optimizer-step behavior under deepspeed are separate bugs." }, { - "left": "issue:43723", - "right": "issue:46710", + "left": "issue:44995", + "right": "issue:45588", "accept": false, - "reason": "Tokenizer loading in v5 and a DeepSeek accuracy regression are different classes of problems." + "reason": "GlmMoeDsa stale cache crash and flash_attention s_aux=None AttributeError affect different code paths." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "StaticCache return-type mismatch and LlamaConfig head_dim validation are unrelated." + "reason": "StaticCache return-type contract and LlamaConfig head_dim validation are unrelated API changes." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading in v5 and an accuracy regression after upgrade are too broad and not the same defect." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "Attention-kernel divergence under torch.compile and Gemma pipeline roundtrip failure are distinct issues." + "reason": "Attention-kernel divergence under torch.compile and Gemma pipeline roundtrip on \u2581 are different issues." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of an older issue is not the same bug as the Gemma roundtrip failure." + "reason": "An isolated reproduction of another issue and a Gemma roundtrip failure are not clearly the same bug." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "Chat-completions rejection and Gemma token roundtrip failure are different API problems." + "reason": "chat/completions request rejection and Gemma pipeline roundtrip failure involve different APIs." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Attention branching under torch.compile and cache lazy initialization causing recompilation are different generation bugs." + "reason": "torch.compile attention branching and chunked-prefill cache lazy init are separate generation/runtime problems." }, { "left": "issue:45584", "right": "issue:46170", "accept": false, - "reason": "Whisper empty-transcription failure and Windows ROCm import/type errors are unrelated." + "reason": "Whisper empty transcription failure and Windows ROCm import dependency issues are unrelated." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "An MPS histogram kernel type error and missing chat template in a processor are unrelated." + "reason": "MPS histogram on Int in moe code and missing chat template in a processor are unrelated errors." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "Sam3PixelDecoder dummy-layer handling and mask pipeline final-batch handling do not share the same root cause." + "reason": "Sam3PixelDecoder dummy layer and partial-batch result dropping are different code paths." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError and a slow-data-filter regression are unrelated." + "reason": "MLukeTokenizer AttributeError and a performance slowdown after upgrade are not the same underlying bug." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer save_pretrained class-metadata mismatch and processor.save_pretrained missing files are different save-path bugs." + "reason": "Tokenizer save_pretrained mismatch and processor.save_pretrained missing files are similar on surface but different serialization failures." }, { "left": "issue:43994", "right": "issue:46726", "accept": false, - "reason": "Siglip2 nonsensical outputs and missing vision-backend processor recognition are different multimodal issues." + "reason": "SigLIP2 nonsensical outputs and AutoProcessor backend-missing detection are different vision-stack issues." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Generic custom-model initialization breakage and ESM2 breakage are not the same concrete defect." - }, - { - "left": "issue:44871", - "right": "issue:46132", - "accept": false, - "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated." + "reason": "Custom model initialization breakage in v5 and ESM2 breakage are broad but distinct regressions." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "A post-upgrade data filtering slowdown and double optimizer stepping under deepspeed are different symptoms." - }, - { - "left": "issue:43232", - "right": "issue:43673", - "accept": false, - "reason": "sync_gpus-related generation kwargs updates and chunked_prefill cache absence are separate generation bugs." - }, - { - "left": "issue:44749", - "right": "issue:45310", - "accept": false, - "reason": "The filtering slowdown report and Qwen3.5 Moe from_pretrained failure do not describe the same bug." + "reason": "Training slowdown after upgrade and a deepspeed double-step bug do not share a concrete code path." }, { - "left": "issue:42994", - "right": "issue:43482", + "left": "issue:44871", + "right": "issue:46132", "accept": false, - "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are opposite-direction, distinct issues." + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated." }, { - "left": "issue:44623", - "right": "issue:45656", + "left": "issue:43232", + "right": "issue:43673", "accept": false, - "reason": "Processor save_pretrained missing files and deepspeed double optimizer stepping are unrelated." + "reason": "sync_gpus generation kwargs update and chunked_prefill cache-missing regression are different generation bugs." }, { - "left": "issue:43475", - "right": "issue:43643", + "left": "issue:43701", + "right": "issue:44749", "accept": false, - "reason": "SAM3 vision output missing fields and AutoConfig trust_remote_code missing fields are different object-shape bugs." + "reason": "resume_from_checkpoint key mismatch and a filtering slowdown after upgrade are unrelated." } ] }, @@ -13137,12 +13136,13 @@ "issue:41628", "issue:42371", "issue:42673", - "issue:42831", "issue:42994", "issue:43232", "issue:43278", "issue:43475", + "issue:43482", "issue:43525", + "issue:43643", "issue:43704", "issue:43723", "issue:43784", @@ -13154,7 +13154,6 @@ "issue:44222", "issue:44351", "issue:44623", - "issue:44792", "issue:44869", "issue:44933", "issue:44945", @@ -13171,6 +13170,7 @@ "issue:45397", "issue:45406", "issue:45478", + "issue:45656", "issue:45753", "issue:46202", "issue:46326", @@ -13179,13 +13179,16 @@ "issue:46726" ], "soft_pairs": [ + "issue:42994|issue:43482", + "issue:44623|issue:45656", + "issue:43475|issue:43643", "issue:45375|issue:45478", - "issue:41628|issue:44933", "issue:28282|issue:30990", + "issue:41628|issue:44933", "issue:44869|issue:45356", "issue:42371|issue:45092", - "issue:45310|issue:45375", "issue:44964|issue:45137", + "issue:45310|issue:45375", "issue:43986|issue:46726", "issue:42994|issue:43525", "issue:44038|issue:44623", @@ -13199,231 +13202,235 @@ "issue:43873|issue:44945", "issue:44945|issue:45753", "issue:41628|issue:44351", - "issue:45081|issue:45103", "issue:46202|issue:46701", + "issue:45081|issue:45103", "issue:42994|issue:43824", "issue:42673|issue:43704", "issue:36010|issue:41628", "issue:43278|issue:45137", "issue:44038|issue:45310", - "issue:44222|issue:45310", - "issue:43232|issue:44792", - "issue:42831|issue:44222" + "issue:44222|issue:45310" ], - "prepared_review_unit_hash": "cec5c1da85215f1f5f9e6fb4fe3d7325e81b96a71d7427c5bb8012e212dc4bf5", + "prepared_review_unit_hash": "343973e152ad19dece89e723bbf5d999649f7029d393734f14efe0b5fa5e5ad4", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35897, - "estimated_input_tokens": 8975, - "estimated_eval_tokens": 18206 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36791, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are false positives from broad lexical similarity. The only clear duplicate cluster is the Qwen3.5 Moe from_pretrained regression: issue:45375 provides the concrete root cause and the other two reports are repeats of the same load failure.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45375 is the best canonical issue: it identifies the specific config/strict-loading root cause (missing deepstack_visual_indexes) behind the broader Qwen3.5 Moe from_pretrained failures reported in issue:45310 and issue:45478.", + "summary": "Mostly heterogeneous issue set, but there are two clear duplicate pockets: the Qwen3ForCausalLM multi-thread VRAM leak and the Qwen3.5 MoE from_pretrained/loading regression. The Qwen3.5 config-field report is the best canonical issue.", + "confidence": 0.63, + "canonical_issue_reason": "issue:45375 is the most specific/root-cause report in the Qwen3.5 MoE loading-error cluster; the other matching reports are more generic symptom variants.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45375 is the most actionable representative of the cluster because it states the concrete bug mechanism, while the related reports are broader duplicates.", + "best_issue_reason": "Most representative item overall: it captures the clearest diagnosis in the strongest duplicate cluster in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized save failure vs GGUF loading failure; related theme but different operations and failure modes." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "processor.save_pretrained file omission vs DeepSpeed optimizer double-step; unrelated code paths." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM3 missing attribute vs trust_remote_code returning missing fields; both are attribute errors but not the same bug." + }, { "left": "issue:45375", "right": "issue:45478", "accept": true, - "reason": "Same Qwen3.5 Moe from_pretrained regression; 45478 is essentially a re-report of the same failure described more specifically in 45375." + "reason": "Same Qwen3.5 MoE from_pretrained regression; one report is the specific config-field cause, the other is the generic failure symptom." }, { - "left": "issue:41628", - "right": "issue:44933", + "left": "issue:28282", + "right": "issue:30990", "accept": false, - "reason": "Both are import-related, but they concern different missing symbols and different code paths, not the same underlying bug." + "reason": "PyTorch-missing ImportError vs sentence-transformers loading hang; different root causes." }, { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:41628", + "right": "issue:44933", "accept": false, - "reason": "Different symptoms: PyTorch missing ImportError vs sentence-transformers loading hang." + "reason": "Different missing-import symbols (AutoImageProcessor vs image_utils export); same area but not the same defect." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Both involve tokenizer/decoding behavior, but one is a timestamp decode crash and the other is a Kimi tokenizer regression with codec handling; not the same bug." + "reason": "Whisper timestamp decode crash vs tokenizer codec regression; unrelated." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "Completely different problems: TF32 API settings warning versus InternVL2 meta-initialization compatibility." - }, - { - "left": "issue:45310", - "right": "issue:45375", - "accept": true, - "reason": "Same underlying Qwen3.5 Moe from_pretrained failure; 45375 appears to pinpoint the root cause that explains the broader error in 45310." + "reason": "TF32 API warning vs InternVL2 meta-init incompatibility; different subsystems." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Different failure modes and code paths: model loading failure versus DeepSpeed ZeRO3 deque IndexError." + "reason": "Phi-4 multimodal loading failure vs DeepSpeed ZeRO3 deque pop; unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": true, + "reason": "Same Qwen3.5 MoE from_pretrained failure; the detailed config-field report and the generic loading error describe the same regression." }, { "left": "issue:43986", "right": "issue:46726", "accept": false, - "reason": "Both are AutoProcessor/backend-related, but one is about missing torchvision for video models and the other is an unrecognized image processor error; not the same bug." + "reason": "Both involve AutoProcessor/backend handling, but one is video-model/no-torchvision and the other is image-processor recognition; too broad to merge." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "One is quantized model saving failure, the other is a Llama4Config attribute error; unrelated." + "reason": "Quantized model save bug vs Llama4 config missing pad_token_id; different issues." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Different subsystems and failures: Qwen3-VL-Moe from_pretrained issue versus missing processor save_pretrained files." + "reason": "Qwen3-VL-Moe transformers v5 bug vs processor.save_pretrained missing files; unrelated." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "No meaningful overlap beyond generic bug-report wording; the actual model failures differ." + "reason": "Placeholder 'bug report' provides no concrete linkage to the gemma-4 ZeRO3 from_pretrained error." }, { "left": "issue:39692", "right": "issue:46726", "accept": false, - "reason": "Both mention processors/vision, but one is a SigLIP2 doc example bug and the other is missing vision-backend handling in AutoProcessor; not the same underlying issue." + "reason": "SigLIP2 doc mistakes vs missing-vision-backend AutoProcessor error; different problems." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "Both are vision-model issues, but one is a missing attribute in SAM 3 video output and the other is a qwen2.5-vl position-id bug." + "reason": "SAM3 attribute error vs qwen2.5-vl video position-id bug; different models and code paths." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "Tokenizer-loading trouble versus continue_final_message behavior for mistral_common tokenizers; related area, different bug." + "reason": "AutoTokenizer loading issue in v5 vs mistral_common continue_final_message behavior; different tokenizer bugs." }, { "left": "issue:43784", "right": "issue:46399", "accept": false, - "reason": "Different import/runtime failures: NameError during sentence-transformers import versus Qwen3.5 kernel registration error." + "reason": "sentence-transformers import NameError vs Qwen3.5 kernel registration failure; unrelated." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Both touch generation/attention behavior, but one is a post-sync_gpus kwargs update issue and the other is a torch.compile attention-kernel divergence." + "reason": "_update_model_kwargs_for_generation after sync_gpus vs attention-kernel divergence under torch.compile; different generation bugs." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Different Gemma-related failures: processor missing _tokenizer versus tokenizer roundtrip failure on \u2581." + "reason": "Gemma4Processor missing _tokenizer vs Gemma pipeline roundtrip on \u2581; related model family but distinct failures." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading problem and pipeline-parallelism output corruption are distinct issues." + "reason": "Quantization offloading issue vs pipeline-parallelism wrong output; different runtime behaviors." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Incorrect output under pipeline parallelism versus an infinite loop on a specific image; not the same code-path bug." + "reason": "Pipeline-parallelism incorrect output vs Qwen3_5 infinite loop on a specific image; distinct bugs." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are missing-import issues, but they refer to different objects and different regressions." + "reason": "Different missing imports (AutoImageProcessor vs HybridCache); same symptom class but not the same defect." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:46202", + "right": "issue:46701", "accept": false, - "reason": "Tokenizer regex patch crash versus auto-docstring annotation parsing crash; unrelated." + "reason": "Gemma pipeline token roundtrip vs DiffusionGemma image-text-to-text pipeline failure; related family but different path." }, { - "left": "issue:46202", - "right": "issue:46701", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "Gemma tokenizer roundtrip failure and DiffusionGemma image-text-to-text pipeline failure are not the same underlying bug." + "reason": "Mistral regex tokenizer crash vs auto_docstring future-annotations crash; unrelated utility bugs." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving and Qwen2_5_VL import failure are unrelated." + "reason": "Quantized save failure vs missing Qwen2.5-VL export; different import/load problems." }, { "left": "issue:42673", "right": "issue:43704", - "accept": false, - "reason": "Same title text, but the later issue is not clearly the same underlying bug from the provided summaries alone; do not merge on title similarity only." + "accept": true, + "reason": "Exact duplicate Qwen3ForCausalLM VRAM leak when used across multiple dataloader threads." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they concern different symbols and different regressions." + "reason": "GenerationMixin import error vs AutoImageProcessor import error; different missing symbols." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "dtype drift between training/eval and a ZeRO3 deque IndexError are unrelated." + "reason": "Embedding dtype regression in evaluate vs ZeRO3 deque underflow; unrelated." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both involve Qwen Moe models, but the model families and reported failures differ enough that this is not safely the same bug." + "reason": "Different Qwen model families (Qwen3-VL-Moe vs Qwen3.5 MoE); similar shape but not the same bug." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained failure versus Qwen3.5 Moe from_pretrained error; different code paths." - }, - { - "left": "issue:43232", - "right": "issue:44792", - "accept": false, - "reason": "Generation-kwargs handling and a janus image-generation test failure are distinct." - }, - { - "left": "issue:42831", - "right": "issue:44222", - "accept": false, - "reason": "Different FP8-related issues: accuracy regression versus save_pretrained failure." + "reason": "FP8 save_pretrained issue vs Qwen3.5 from_pretrained error; different code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge pairs. The Qwen3.5 Moe cluster around 45310/45375/45478 is a reasonable canonical choice, and the false pairs are not overstated." + "feedback": "The clustering summary is grounded in the packet, and the soft-edge calls are generally conservative. The two accepted pockets are supported by the titles: the exact duplicate Qwen3ForCausalLM VRAM leak pair, and the Qwen3.5 MoE from_pretrained regression cluster where 45375 is a more specific config-field report. The rejected pairs mostly involve clearly different operations or subsystems, so I don\u2019t see overstatement here." }, "evaluator_used": true, "retried": false, @@ -13450,12 +13457,13 @@ "issue:33290", "issue:34689", "issue:35141", - "issue:36331", "issue:42175", "issue:42222", "issue:42673", + "issue:42831", "issue:42898", "issue:43116", + "issue:43232", "issue:43295", "issue:43377", "issue:43479", @@ -13466,24 +13474,23 @@ "issue:43874", "issue:43976", "issue:44016", - "issue:44060", "issue:44079", "issue:44190", + "issue:44222", "issue:44242", "issue:44496", "issue:44568", "issue:44623", "issue:44734", "issue:44779", + "issue:44792", "issue:44869", "issue:44933", "issue:45092", "issue:45103", - "issue:45125", "issue:45200", "issue:45310", "issue:45356", - "issue:45405", "issue:45561", "issue:45600", "issue:45712", @@ -13494,14 +13501,16 @@ "issue:46693" ], "soft_pairs": [ + "issue:43232|issue:44792", + "issue:42831|issue:44222", + "issue:30333|issue:33290", "issue:43504|issue:45561", "issue:42175|issue:44933", "issue:45561|issue:45600", - "issue:30333|issue:33290", "issue:44496|issue:45561", "issue:43976|issue:44933", - "issue:42175|issue:45712", "issue:30990|issue:35141", + "issue:42175|issue:45712", "issue:42898|issue:44568", "issue:46077|issue:46123", "issue:43377|issue:44242", @@ -13510,8 +13519,8 @@ "issue:42673|issue:46693", "issue:44016|issue:46231", "issue:42898|issue:45812", - "issue:42222|issue:45200", "issue:43723|issue:44568", + "issue:42222|issue:45200", "issue:43874|issue:45092", "issue:42898|issue:44779", "issue:43479|issue:44079", @@ -13521,225 +13530,216 @@ "issue:30333|issue:34689", "issue:42898|issue:45356", "issue:42898|issue:43295", - "issue:43116|issue:44190", - "issue:45405|issue:45712", - "issue:44060|issue:45125", - "issue:30333|issue:36331" + "issue:43116|issue:44190" ], - "prepared_review_unit_hash": "4ecad5d5ce7b74c1cee8bb80c6d89d716d80090e59cfc4da9b9acf84720ca699", + "prepared_review_unit_hash": "e1a4149fbe6d645710d689d1bb895aefdc395395a9f16ea037884cafed739f62", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37327, - "estimated_input_tokens": 9332, - "estimated_eval_tokens": 18920 + "soft_pair_count": 30, + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No convincing duplicate cluster here: the issues span unrelated subsystems and failure modes, so none of the soft edges should be merged.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue for the cluster; the items are heterogeneous. If one representative is needed, issue:45812 is the strongest standalone bug report.", + "summary": "This cluster is highly heterogeneous: the issues span unrelated bugs in tokenization, model loading, training, packaging, tests, and inference. None of the soft-edge pairs look like the same underlying defect, so this is not a true duplicate cluster.", + "confidence": 0.93, + "canonical_issue_reason": "No obvious duplicate centerpiece exists; if one issue must represent the cluster, issue 46693 is the most actionable open regression with a concrete performance problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the best representative issue: it describes a broad, concrete, user-visible silent regression with clear model scope.", + "best_issue_reason": "Issue 46693 is the strongest standalone bug report here: it is open, specific, and describes a reproducible performance regression. The rest are mostly unrelated closed issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Different failures: generation kwargs handling vs a janus image-generation test regression." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both mention FP8, but one is an accuracy issue and the other is save_pretrained/moe serialization." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLFlow reporting on failed jobs is unrelated to a deepspeed adafactor OOM." + }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Unrelated: model preset loading with legacy fields vs a pytest-xdist file race in testing utils." + "reason": "Model preset loading via a legacy field is unrelated to a pytest-xdist file race." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Different problems: backend extras packaging vs a missing import in image_utils." + "reason": "Packaging/backend install behavior is unrelated to a missing import symbol in image_utils." }, { "left": "issue:45561", "right": "issue:45600", "accept": false, - "reason": "Both touch tests/CI, but one is an xdist capture-file race and the other is stale auto_mappings config references." - }, - { - "left": "issue:30333", - "right": "issue:33290", - "accept": false, - "reason": "MLFlow reporting failure is unrelated to an OOM from Adafactor in DeepSpeed." + "reason": "A test-file race is not the same bug as stale auto_mappings references to removed configs." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Model config recognition bug vs testing utility race; no shared code-path." + "reason": "Unrecognized model/config loading failure is unrelated to a testing-utils race condition." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version compatibility issue vs a missing import; different causes and fixes." + "reason": "Python version compatibility is a different issue from a nonexistent import in image_utils." }, { - "left": "issue:42175", - "right": "issue:45712", + "left": "issue:30990", + "right": "issue:35141", "accept": false, - "reason": "Packaging/backend dependency issue vs leftover dummy classes leaking into repo checks." + "reason": "Sentence-transformers loading hangs vs output embedding reinitialization during resize are different code paths." }, { - "left": "issue:30990", - "right": "issue:35141", + "left": "issue:42175", + "right": "issue:45712", "accept": false, - "reason": "Loading hang for sentence-transformers is unrelated to embedding resizing reinitialization." + "reason": "Install extra backend packaging differs from dummy class cleanup and repo-check failures." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Tokenizer cleanup behavior change vs add_special_tokens not adding BOS/EOS tokens; different tokenizer paths." + "reason": "Tokenization cleanup behavior in v5 is unrelated to BOS/EOS tokens not being added." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Attention-mask semantics for Gemma 4 are unrelated to the partial-batch pipeline result-dropping bug." + "reason": "Gemma4 attention-mask construction and partial-batch pipeline result dropping are unrelated." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder vs gating of load-balancing loss by output_router_logits." + "reason": "MIMI padding-mask output mismatch is unrelated to missing load-balancing loss for router logits." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing files from processor.save_pretrained is not the same as Qwen3.5 Moe from_pretrained failure." + "reason": "Missing processor files on save is a different bug from Qwen3.5 MoE from_pretrained loading failure." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Config default initialization bug vs Whisper word-timestamp decode crash; no overlap." + "reason": "Default multimodal config initialization is unrelated to Whisper word-timestamp decoding crash." }, { "left": "issue:42673", "right": "issue:46693", "accept": false, - "reason": "VRAM leak in multi-threaded dataloading is distinct from an FA2 long-context performance regression." + "reason": "VRAM leak in multi-threaded dataloading is distinct from a long-context FA2 performance regression." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and a docs parameter-name mismatch are unrelated documentation issues." + "reason": "A notebook syntax error is not the same as a documentation parameter-name mismatch." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer-related regressions, but they affect different behaviors and code paths." + "reason": "Tokenizer cleanup behavior changes are unrelated to Granite token-id regression." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:43723", + "right": "issue:44568", "accept": false, - "reason": "Vitpose model breakage is unrelated to Gemma 4's mm_token_type_ids defaulting issue." + "reason": "Tokenizer loading in v5 is not the same defect as add_special_tokens failing to add BOS/EOS." }, { - "left": "issue:43723", - "right": "issue:44568", + "left": "issue:42222", + "right": "issue:45200", "accept": false, - "reason": "Tokenizer loading failure vs special-token insertion regression; different failure surfaces." + "reason": "Broken vitpose model files and Gemma4 token-type defaults concern different models and code paths." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image-patch method in GLM46V vs legacy remote-code incompatibility in InternVL2; not the same bug." + "reason": "A missing image-processor method is unrelated to old InternVL2 checkpoints failing meta initialization." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Different tokenizer regressions: cleanup-space behavior vs Deepseek codec handling." + "reason": "Tokenization cleanup behavior is unrelated to Deepseek tokenizer codec handling." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Default multimodal config initialization is unrelated to ModelOutput key assignment when values were previously None." + "reason": "Default config initialization and ModelOutput key assignment are separate bugs." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Import-time NameError in sentence-transformers vs missing Qwen2.5-VL class export; different modules and causes." + "reason": "Import-time NameError in sentence-transformers is unrelated to a missing Qwen2.5-VL export." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Serving KV-cache tensor indexing crash is unrelated to the auto_docstring annotations bug." + "reason": "KV-cache continuation indexing in serve has nothing to do with auto_docstring annotations handling." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta is unrelated to MLFlow reporting after training failure." + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to MLFlow reporting after failed training jobs." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "Training-job reporting problem vs model loading regression for Llama 3.2 Vision; no shared code path." + "reason": "Failed-job MLFlow reporting is unrelated to a Llama 3.2 vision model-loading regression." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Tokenizer cleanup behavior change is not the same as Kimi-K2.5 codec regression and warning mismatch." + "reason": "Tokenization-space cleanup changes are unrelated to Kimi-K2.5 codec handling regression." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "Generic tokenizer cleanup change vs custom code breaking on processor.tokenizer/image handling; different regressions." + "reason": "Tokenization cleanup behavior is unrelated to processor.tokenizer/image-token passing regression." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Example script returning empty multi-label results is unrelated to loading a local dataset in the no-trainer image classification script." - }, - { - "left": "issue:45405", - "right": "issue:45712", - "accept": false, - "reason": "Dependency version bump issue is unrelated to dummy class cleanup and repo-check failures." - }, - { - "left": "issue:44060", - "right": "issue:45125", - "accept": false, - "reason": "Incorrect tied-weights warning is a different tensor-sharing bug than missing tensor-parallel plan metadata." - }, - { - "left": "issue:30333", - "right": "issue:36331", - "accept": false, - "reason": "MLFlow reporting on failure is unrelated to the CustomTrainer compute_loss signature mismatch." + "reason": "Empty multi-label classification outputs in the example script are not the same as local dataset loading failure." } ] }, @@ -13767,16 +13767,17 @@ "issue:28282", "issue:30333", "issue:35141", + "issue:36331", "issue:41628", "issue:42757", "issue:42796", "issue:42831", "issue:43122", "issue:43232", - "issue:43475", "issue:43479", "issue:43644", "issue:43723", + "issue:44060", "issue:44075", "issue:44190", "issue:44206", @@ -13784,7 +13785,6 @@ "issue:44263", "issue:44295", "issue:44297", - "issue:44360", "issue:44373", "issue:44448", "issue:44484", @@ -13796,8 +13796,10 @@ "issue:45071", "issue:45092", "issue:45103", + "issue:45125", "issue:45245", "issue:45310", + "issue:45405", "issue:45656", "issue:45712", "issue:45834", @@ -13805,16 +13807,17 @@ "issue:46132", "issue:46209", "issue:46231", - "issue:46305", "issue:46421", "issue:46424", - "issue:46431", "issue:46665", "issue:46710" ], "soft_pairs": [ + "issue:45405|issue:45712", + "issue:44060|issue:45125", "issue:44295|issue:45656", "issue:30333|issue:35141", + "issue:30333|issue:36331", "issue:22355|issue:28282", "issue:44933|issue:45712", "issue:44075|issue:46231", @@ -13832,221 +13835,230 @@ "issue:44373|issue:44485", "issue:46132|issue:46710", "issue:44297|issue:45310", - "issue:43723|issue:44977", "issue:44263|issue:45092", + "issue:43723|issue:44977", "issue:44222|issue:44484", "issue:44190|issue:46082", "issue:43232|issue:46424", "issue:44206|issue:45103", "issue:43122|issue:46209", "issue:43232|issue:45834", - "issue:44734|issue:46421", - "issue:43475|issue:46305", - "issue:46209|issue:46431", - "issue:44263|issue:44360" + "issue:44734|issue:46421" ], - "prepared_review_unit_hash": "bbe042f977e697ed2e0bb117de4b0b4a932f146bc2fb1bdb2b425e3257745623", + "prepared_review_unit_hash": "95f36a4ddca1d29977f44bfb891f7bb7f2f37f120ab5ddfddd058ab9f1c3a3ef", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36711, - "estimated_input_tokens": 9178, - "estimated_eval_tokens": 18612 + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is not a true duplicate cluster: the items span unrelated import errors, tokenizer regressions, config/initialization bugs, FP8 issues, generation bugs, and docs nits. All soft-pair proposals are false positives and should be kept separate.", + "summary": "No clear duplicate cluster here: the items span many unrelated Transformers issues, so all soft edges should be rejected.", "confidence": 0.86, - "canonical_issue_reason": "issue:43122 is the broadest regression report in the set and the best representative of the small tokenizer/version-behavior subset; there is no stronger single canonical issue because the cluster is heterogeneous.", + "canonical_issue_reason": "No single canonical issue; the cluster is heterogeneous (imports, tokenizer serialization, training, generation, config, FP8/cache, and docs bugs) rather than one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the most general user-facing issue here and the closest thing to an anchor among otherwise unrelated reports.", + "best_issue_reason": "Issue 44297 is the most concrete and self-contained bug report, but it only covers a narrow tokenizer save_pretrained mismatch and does not represent the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Dependency version bump vs leftover dummy-class cleanup; different problems." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both involve Qwen MoE, but one is a tied-weights warning and the other is missing tensor-parallel metadata." + }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Registered-buffer read error vs. deepspeed optimizer being stepped twice; unrelated bugs." + "reason": "Position_ids buffer read error vs optimizer step duplication under deepspeed; unrelated." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "MLflow reporting on failed training jobs vs. embedding reinitialization after resize; different code paths." + "reason": "MLflow job reporting failure vs embedding reinitialization after resize; different code paths." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting bug vs custom trainer loss signature mismatch; not the same issue." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing `transformers.onnx` module vs. missing PyTorch for `AutoModel`; both import-related but not the same bug." + "reason": "Both are import errors, but for different modules and causes." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Nonexistent `image_utils` import vs. leftover dummy classes leaking into `dir(transformers)`; different issues." + "reason": "Nonexistent image_utils import vs dummy class cleanup; different fixes." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Ignored SGD optimizer arguments vs. a docstring naming mismatch; code bug vs. documentation bug." + "reason": "Optimizer arguments ignored vs docstring parameter-name mismatch." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 `save_pretrained` issue in MoE models vs. tokenizer class metadata mismatch; unrelated save/load problems." + "reason": "FP8 save_pretrained bug vs tokenizer save_pretrained metadata mismatch." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Model output differs between v4 and v5 vs. `AttentionInterface.register` changing function behavior; not the same defect." + "reason": "Generation output regression vs attention registration behavior change; not the same defect." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FineGrainedFP8 accuracy regression vs. missing files from `processor.save_pretrained`; different functionality." + "reason": "FP8 accuracy issue vs processor.save_pretrained missing files." }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "Tokenizer BPE marker handling vs. `AutoTokenizer.from_pretrained` loading failure; both tokenizer-related but distinct bugs." + "reason": "Tokenizer BPE marker handling vs AutoTokenizer loading failure in v5." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer config mismatch vs. optimizer step being called twice; no meaningful overlap." + "reason": "Tokenizer config mismatch vs optimizer step behavior; unrelated." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Default multimodal config initialization vs. non-persistent buffers being filled with junk; different initialization bugs." + "reason": "Phi4 multimodal config defaults vs non-persistent buffer junk; different underlying bugs." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-count limit error vs. attention registry behavior change; unrelated." + "reason": "Category limit error vs attention registration behavior." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "FP8 MoE serialization bug vs. processor serialization missing files; both mention save/load, but not the same problem." + "reason": "FP8 save_pretrained issue vs processor save_pretrained missing files." }, { "left": "issue:43479", "right": "issue:46665", "accept": false, - "reason": "Config defaults when passed `None` vs. inconsistent `from_pretrained` behavior inside Python classes; separate causes." + "reason": "Config default initialization bug vs from_pretrained behavior inside Python classes." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Missing `AutoImageProcessor` import vs. missing `is_offline_mode` import; different symbols and failures." + "reason": "Different import failures from different packages/modules." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "`PretrainedConfig` type checking regression vs. `LlamaConfig` head_dim validation; different config issues." + "reason": "PretrainedConfig type-checking regression vs LlamaConfig head_dim validation." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Wrong docstring for `position_ids` vs. GLM-5 RoPE implementation bug; documentation vs. model math." + "reason": "Docstring typo vs RoPE implementation discussion; not the same bug." }, { "left": "issue:46132", "right": "issue:46710", "accept": false, - "reason": "Behavior change in registered attention functions vs. accuracy regression after upgrade; not the same code-path bug." + "reason": "Attention registration behavior vs model accuracy regression; too broad to merge." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer save metadata mismatch vs. Qwen3.5 MoE `from_pretrained` error; unrelated." + "reason": "Tokenizer save_pretrained metadata bug vs Qwen3.5 MoE from_pretrained error." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:44263", + "right": "issue:45092", "accept": false, - "reason": "`AutoTokenizer.from_pretrained` issue vs. flash-attention generation bug in Qwen3.5; different subsystems." + "reason": "torch.split return values in an indexer vs remote-code/meta-init incompatibility." }, { - "left": "issue:44263", - "right": "issue:45092", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "`torch.split` return-value handling in one indexer vs. InternVL2 remote-code/meta-init incompatibility; not the same bug." + "reason": "Tokenizer loading issue in v5 vs Qwen3.5 flash-attention generation bug." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "FP8 MoE save bug vs. `max_shard_size` default question; unrelated save_pretrained topics." + "reason": "FP8 save_pretrained bug vs max_shard_size question." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading in an example script vs. `LlamaConfig` head_dim validation; completely different." + "reason": "Example script dataset loading issue vs LlamaConfig head_dim validation." }, { "left": "issue:43232", "right": "issue:46424", "accept": false, - "reason": "Generation kwarg update after `sync_gpus` vs. cache length handling for quantized cache; different cache/generation behaviors." + "reason": "Generation kwargs update after sync_gpus vs cache max_cache_length behavior." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Feature-extractor center-argument crash vs. auto-docstring crash under future annotations; unrelated." + "reason": "Feature extractor center-arg regression vs auto-docstring AttributeError." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenizer output changed across versions vs. FP8 inference failing on B200; unrelated regression types." + "reason": "Tokenizer output drift across versions vs FP8 inference failure on B200." }, { - "left": "issue:44373", - "right": "issue:44485", - "accept": false, - "reason": "Docstring typo vs. RoPE implementation discussion; not the same artifact." - }, - { - "left": "issue:46209", - "right": "issue:46431", + "left": "issue:43232", + "right": "issue:45834", "accept": false, - "reason": "FP8 inference failure on B200 vs. missing chat template error; distinct model/runtime issues." + "reason": "Generation kwargs update bug vs Kosmos2.5 OCR index error." }, { - "left": "issue:44263", - "right": "issue:44360", + "left": "issue:44734", + "right": "issue:46421", "accept": false, - "reason": "`torch.split` return handling vs. missing ReLU in the DSA indexer; same broad area but different concrete defects." + "reason": "Serve KV-cache indexing crash vs lazy cache initialization causing recompilation." } ] }, @@ -14077,17 +14089,15 @@ "issue:38175", "issue:39401", "issue:41628", - "issue:42371", "issue:42757", - "issue:42796", "issue:42831", "issue:42886", "issue:42907", "issue:42994", "issue:43066", - "issue:43122", "issue:43232", "issue:43408", + "issue:43475", "issue:43644", "issue:43646", "issue:43824", @@ -14097,12 +14107,12 @@ "issue:44263", "issue:44279", "issue:44295", + "issue:44360", "issue:44373", "issue:44483", "issue:44568", "issue:44623", "issue:44829", - "issue:44945", "issue:45092", "issue:45278", "issue:45310", @@ -14113,25 +14123,30 @@ "issue:45753", "issue:46002", "issue:46050", + "issue:46209", "issue:46305", "issue:46328", "issue:46369", "issue:46402", + "issue:46431", "issue:46655" ], "soft_pairs": [ + "issue:43475|issue:46305", + "issue:46209|issue:46431", "issue:46305|issue:46655", - "issue:30990|issue:38175", - "issue:45092|issue:45357", + "issue:44263|issue:44360", "issue:44295|issue:44623", + "issue:30990|issue:38175", "issue:43644|issue:46305", "issue:44261|issue:45092", + "issue:45092|issue:45357", "issue:44263|issue:44373", "issue:45381|issue:46305", "issue:42907|issue:46050", "issue:44279|issue:46328", - "issue:44829|issue:46002", "issue:42831|issue:45310", + "issue:44829|issue:46002", "issue:22355|issue:27957", "issue:44483|issue:45406", "issue:41628|issue:43824", @@ -14144,229 +14159,226 @@ "issue:27957|issue:30990", "issue:42886|issue:43066", "issue:42886|issue:42994", - "issue:43232|issue:45593", "issue:44188|issue:44373", + "issue:43232|issue:45593", "issue:43408|issue:46369", "issue:44261|issue:44373", - "issue:43646|issue:46002", - "issue:42796|issue:43122", - "issue:42371|issue:44263", - "issue:42907|issue:44945" + "issue:43646|issue:46002" ], - "prepared_review_unit_hash": "814962ce0a753e7e195fcb30024fffad99c388c30dbae920628c97c2017ff1bf", + "prepared_review_unit_hash": "dd1816e8c7cee2b7b2bf961a90fc05f0abbbe35c893822db821607f24d56e65e", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36758, - "estimated_input_tokens": 9190, - "estimated_eval_tokens": 18636 + "serialized_chars": 36817, + "estimated_input_tokens": 9205, + "estimated_eval_tokens": 18666 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is broad and noisy: it mixes unrelated tokenizer, import, generation, quantization, model-init, and processor bugs. None of the soft-similarity pairs are convincing duplicates of the same concrete code-path issue.", - "confidence": 0.74, - "canonical_issue_reason": null, + "summary": "All candidate soft pairs are rejects: they only share broad themes like tokenizers, loading, quantization, or SAM3, but they describe different concrete bugs and code paths. No single issue stands out as a canonical representative for this mixed set.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue exists here: the items are a mixed bag of unrelated bug reports, and the superficially similar pairs still point to different models, features, or failure modes.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "None of the issues is a strong global representative because the set is heterogeneous and the closest-looking pairs are not the same underlying defect.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:46305", - "right": "issue:46655", + "left": "issue:43475", + "right": "issue:46305", "accept": false, - "reason": "Both involve video-related masking/padding, but one is a MoonshineStreaming sliding-window off-by-one and the other is a SmolVLM video processor padding bug." + "reason": "Different SAM3 vision-encoder attribute error vs MoonshineStreaming mask off-by-one; unrelated code paths." }, { - "left": "issue:30990", - "right": "issue:38175", + "left": "issue:46209", + "right": "issue:46431", "accept": false, - "reason": "One is a loading hang for Sentence Transformers and the other is zero probabilities in a SigLIP2 model; different symptoms and likely different code paths." + "reason": "FP8 B200 inference failure is unrelated to processor.save_pretrained missing files." }, { - "left": "issue:45092", - "right": "issue:45357", + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both involve video processing, but they are different bugs in different models and functions." + }, + { + "left": "issue:44263", + "right": "issue:44360", "accept": false, - "reason": "Both touch model save/load compatibility, but they describe different failures in different models: meta-init incompatibility vs incorrect visual encoder keys on save_pretrained." + "reason": "Different DSA indexer problems: torch.split return values vs missing ReLU." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "A position_ids buffer read error and missing files from processor.save_pretrained are separate serialization problems, not the same bug." + "reason": "Registering position_ids as a buffer is a different issue from processor.save_pretrained missing files." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and SigLIP2 zero probabilities are different failures." }, { "left": "issue:43644", "right": "issue:46305", "accept": false, - "reason": "Non-persistent buffer junk in Transformers 5.0.0 is unrelated to a sliding-window attention mask off-by-one." + "reason": "Non-persistent buffer junk in Transformers 5.0 is unrelated to MoonshineStreaming masking." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm is a precision/config issue, not the same as the InternVL meta-initialization compatibility bug." + "reason": "MLA RMS eps precision issue is unrelated to InternVL2 meta-initialization incompatibility." }, { - "left": "issue:44263", - "right": "issue:44373", + "left": "issue:45092", + "right": "issue:45357", "accept": false, - "reason": "torch.split return handling in GlmMoeDsaIndexer is unrelated to a wrong docstring for position_ids." + "reason": "Different models and bugs: InternVL2 meta init vs Qwen3.5 save_pretrained key handling." }, { - "left": "issue:45381", - "right": "issue:46305", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "Both are video/vision-adjacent, but one is a Qwen2.5-VL position_ids regression and the other is a MoonshineStreaming mask off-by-one." + "reason": "Indexer torch.split behavior is unrelated to a wrong docstring for position_ids." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both concern quantized model persistence, but one is about failing to save dequantized Ministral/Devstral models and the other is a sinq save_pretrained/from_pretrained breakage." + "reason": "Both concern quantized save/load, but they affect different quantization flows and concrete failure modes." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "A dependency issue and a SequenceFeatureExtractor.pad performance regression are not the same underlying bug." + "reason": "Generic dependency issue is not the same as SequenceFeatureExtractor.pad inefficiency." }, { - "left": "issue:44829", - "right": "issue:46002", + "left": "issue:42831", + "right": "issue:45310", "accept": false, - "reason": "Flash Attention 3 training degeneration and NaNs after checkout/init are both training problems, but they are not the same failure mode or code path." + "reason": "FineGrainedFP8 accuracy issue and Qwen3.5 Moe from_pretrained error are different bugs." }, { - "left": "issue:42831", - "right": "issue:45310", + "left": "issue:44829", + "right": "issue:46002", "accept": false, - "reason": "FineGrainedFP8 accuracy issues and a Qwen3.5 MoE from_pretrained error are unrelated." + "reason": "Flash Attention 3 degenerate training is not the same as NaNs after checkout/init." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "A missing transformers.onnx module and an XLMRoberta Flash Attention 2 issue are completely different problems." + "reason": "onnx import missing and XLMRoberta Flash Attention 2 are unrelated." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "The chat/completions request rejection and the Gemma4Processor _tokenizer crash affect different serving paths." + "reason": "Chat completions request rejection differs from Gemma4Processor missing _tokenizer." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different export breakages." + "reason": "Both are import errors, but for different symbols and likely different causes." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FineGrainedFP8 accuracy regression and Qwen3-VL-Moe import/use failure are not the same bug." + "reason": "FineGrainedFP8 accuracy and Qwen3-VL-Moe are not the same underlying defect." }, { "left": "issue:43408", "right": "issue:46402", "accept": false, - "reason": "A sam3_video vs sam3_tracker warning and missing Sam3Model text encoder weights are separate model-loading issues." + "reason": "Wrong model-type warning and missing text encoder weights are distinct SAM3 issues." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Missing transformers.onnx and tokenizer adds an extra space after an added token do not describe the same defect." + "reason": "Missing transformers.onnx module and extra whitespace after added token are different tokenizer/import problems." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "FineGrainedFP8 accuracy drift and a Qwen3_5 infinite loop on a specific image are unrelated failures." + "reason": "FP8 accuracy issue is unrelated to a Qwen3_5 infinite loop on a specific image." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer behavior bugs, but one is wrong offset_mapping for Qwen3 and the other is add_special_tokens failing to add BOS/EOS for mdeberta-v3-base." + "reason": "Wrong offset_mapping and missing BOS/EOS insertion are separate tokenizer bugs." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "The is_offline_mode import error may be one symptom class, but 'many import errors after update' is broader and not the same concrete bug." + "reason": "A specific is_offline_mode import failure is only one of many unrelated import regressions." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and a Sentence Transformers loading hang are unrelated." + "reason": "Flash Attention 2 on XLMRoberta and Sentence Transformers loading hang are unrelated." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading and wrong tokenizer decoder type in v5 are distinct tokenizer issues." + "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer problems." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Offline cache loading failure and quantized model saving failure are different bugs." + "reason": "Offline tokenizer cache loading and quantized model saving failure are unrelated." }, { - "left": "issue:43232", - "right": "issue:45593", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Generation state handling after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." + "reason": "torch.compile attention-kernel divergence is unrelated to a docstring issue." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:43232", + "right": "issue:45593", "accept": false, - "reason": "A torch.compile attention-kernel branching bug and a wrong docstring for position_ids are not duplicates." + "reason": "Generation kwargs update after sync_gpus and D-FINE auxiliary-loss behavior are different code paths." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "Both involve sam3 models, but a model-type warning and memory overflow in sam3-litetext are different issues." + "reason": "SAM3 model-type warning and SAM3-litetext memory overflow are different failures." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm and a wrong position_ids docstring are unrelated." + "reason": "MLA layernorm precision bug is unrelated to the position_ids docstring." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage and NaNs in classification heads upon checkout+init are both init-related, but not the same code-path bug." - }, - { - "left": "issue:42796", - "right": "issue:43122", - "accept": false, - "reason": "A Ministral tokenizer BPE-marker bug and a cross-version tokenization difference report are related only at a very high level." - }, - { - "left": "issue:42371", - "right": "issue:44263", - "accept": false, - "reason": "TF32 API guidance and torch.split behavior in GlmMoeDsaIndexer are unrelated." + "reason": "Custom model initialization breakage and NaNs in classification heads are not the same bug." } ] }, @@ -14391,18 +14403,18 @@ "cluster_id": "cluster-10000-479", "nodes": [ "issue:22355", - "issue:29127", "issue:29159", "issue:30990", "issue:33357", "issue:36296", "issue:39401", "issue:41628", + "issue:42371", "issue:42757", + "issue:42796", "issue:42907", "issue:43116", "issue:43122", - "issue:43278", "issue:43408", "issue:43452", "issue:43475", @@ -14416,6 +14428,7 @@ "issue:44060", "issue:44117", "issue:44222", + "issue:44263", "issue:44295", "issue:44297", "issue:44351", @@ -14439,15 +14452,18 @@ "issue:46614" ], "soft_pairs": [ + "issue:42796|issue:43122", + "issue:42371|issue:44263", + "issue:42907|issue:44945", "issue:30990|issue:33357", - "issue:43756|issue:44060", "issue:43576|issue:43784", + "issue:43756|issue:44060", "issue:39401|issue:44488", - "issue:43976|issue:44945", "issue:42907|issue:44964", + "issue:43976|issue:44945", "issue:41628|issue:45278", - "issue:43784|issue:45600", "issue:44295|issue:44297", + "issue:43784|issue:45600", "issue:42907|issue:43937", "issue:44117|issue:44704", "issue:44625|issue:44829", @@ -14455,8 +14471,8 @@ "issue:39401|issue:44521", "issue:44704|issue:45290", "issue:43408|issue:43475", - "issue:43122|issue:46326", "issue:43122|issue:44521", + "issue:43122|issue:46326", "issue:42907|issue:43452", "issue:22355|issue:29159", "issue:44038|issue:44297", @@ -14464,234 +14480,238 @@ "issue:39401|issue:46614", "issue:45216|issue:45478", "issue:44038|issue:44222", - "issue:43824|issue:44351", "issue:44038|issue:44295", + "issue:43824|issue:44351", "issue:43122|issue:44779", "issue:36296|issue:43116", - "issue:42757|issue:44351", - "issue:43278|issue:44945", - "issue:22355|issue:29127" + "issue:42757|issue:44351" ], - "prepared_review_unit_hash": "e0fd92454c9afd1f975ffb70eb0ede29dbad4ac08782f9e7479030316fbfd585", + "prepared_review_unit_hash": "0979ba55038e42903adc8151920e6ca61cdb2660ff6815b777106a32cb926892", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 32, - "serialized_chars": 36447, - "estimated_input_tokens": 9112, - "estimated_eval_tokens": 18480 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37218, + "estimated_input_tokens": 9305, + "estimated_eval_tokens": 18866 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft pairs are mostly broad-similarity false positives across unrelated bugs, models, or APIs. None of the candidates looks like the same underlying change with enough specificity to merge.", - "confidence": 0.84, - "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous. If forced to pick a representative, issue:44038 is the most concrete model-specific regression, but it is not a true duplicate hub.", + "summary": "This cluster is loosely related by v5 regressions and import/tokenizer breakages, but most pairs are not true duplicates. The broadest representative is the v5 upgrade/import-regression issue.", + "confidence": 0.39, + "canonical_issue_reason": "issue:45278 is the broadest umbrella here: it describes multiple import failures after upgrading to 5.5.0, which matches the cluster\u2019s strongest recurring theme better than the more specific model- or tokenizer-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44038 is the most self-contained, specific bug report in the set (clear model/version scope), but the overall cluster is too mixed for a strong global best.", + "best_issue_reason": "issue:45278 is the best single issue to represent this cluster because it is the most general and central regression report among a set of mostly version-related import/tokenizer problems.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:42796", + "right": "issue:43122", "accept": false, - "reason": "Different model/platform failures; only vague loading symptoms overlap." + "reason": "Both are tokenizer-related, but one is a model-specific BPE-marker bug and the other is a broader v4\u2192v5 tokenization regression; not the same underlying defect." }, { - "left": "issue:43756", - "right": "issue:44060", + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "Unrelated: TF32 settings vs. a DSA indexer ReLU/math bug." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Different failures in different code paths: dequantized save/load vs pipeline-parallel output correctness." + }, + { + "left": "issue:30990", + "right": "issue:33357", "accept": false, - "reason": "Same broad transformer area, but different models and defects (RoPE layers vs weight tying)." + "reason": "Both mention loading, but they concern different models and failure modes (stuck load vs Mac bus error)." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "CLI env command vs sentence-transformers import NameError; unrelated code paths." + "reason": "Both are version-5 breakages, but one is a broken CLI command and the other is a missing import during sentence-transformers import; not the same bug." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Different models and failure modes (offset mapping vs model load)." + "reason": "Completely different model-architecture bugs: RoPE layer count vs tied-weights warning." }, { - "left": "issue:43976", - "right": "issue:44945", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Python/version compatibility issue vs pipeline-parallel output bug; unrelated." + "reason": "Different tokenizer/model issues: wrong offset mappings vs failure to load a specific model." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Save/dequantize bug vs model-load bug; different paths." + "reason": "Save/dequantization bug is unrelated to a multimodal model-loading failure." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:44945", "accept": false, - "reason": "Broad import-error theme only; one missing symbol, one many import failures." + "reason": "A Python-version compatibility issue is not the same as a pipeline-parallelism output bug." }, { - "left": "issue:43784", - "right": "issue:45600", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Sentence-transformers import NameError vs SAM3 CI mapping reference; unrelated." + "reason": "Both are import-related, but one is a missing symbol and the other is a broad set of upgrade import errors; too different to merge." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Both involve save/pretrained serialization, but one is buffer access, the other tokenizer config metadata." + "reason": "Unrelated: buffer access/readback bug vs tokenizer save_pretrained metadata mismatch." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Different subsystems and failures: missing nn import vs stale auto_mappings references." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Save/export bug vs generation config validation bug; unrelated." + "reason": "Different model families and error types: save_pretrained dequantization vs invalid GenerationConfig." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "Both touch from_pretrained plumbing, but tokenizer mapping None vs processor cached_file kwargs are different bugs." + "reason": "Both involve from_pretrained APIs, but one is tokenizer mapping None handling and the other is AutoProcessor kwargs propagation; not the same bug." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Configuration propagation vs flash-attention training degeneration; unrelated." + "reason": "Different issues: config propagation for Qwen3.5 vs degenerate training with flash_attention_3." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "Completely different model discussions." + "reason": "DSA math/activation discussion is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset mapping vs multimodal chat masking; different issues." + "reason": "Both are tokenizer-adjacent, but one is offset_mapping correctness and the other is assistant mask generation for multimodal chat templates." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Different APIs: AutoProcessor kwargs passthrough vs chat template tool-call crash." + "reason": "Unrelated API bugs: cached_file kwargs forwarding vs chat-template crashing on tool-call messages." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Same SAM3 family, but config-type warning and missing output attribute are distinct bugs." + "reason": "Same model family, but different concrete bugs: model-type warning vs missing output attribute." }, { "left": "issue:43122", - "right": "issue:46326", + "right": "issue:44521", "accept": false, - "reason": "Tokenizer regression vs continue_final_message behavior; different code paths." + "reason": "Both concern tokenization behavior, but one is a general regression between versions and the other is multimodal assistant-mask generation." }, { "left": "issue:43122", - "right": "issue:44521", + "right": "issue:46326", "accept": false, - "reason": "Tokenization regression vs multimodal assistant masks; not the same bug." + "reason": "Different tokenizer regressions in different functions/models; not the same underlying bug." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Different load/save mechanisms (gguf load vs dequantized save)." + "reason": "Unrelated failures: dequantized save and gguf loading behavior are different code paths." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Different submodules: ONNX import vs slow/fast tokenizer inconsistency." + "reason": "Different topics: missing transformers.onnx module vs slow/fast tokenizer inconsistency." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe load bug vs tokenizer config save mismatch; unrelated." + "reason": "Different bugs: Qwen3-VL-Moe model issue vs tokenizer_config save metadata mismatch." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example-script classification bug vs mm_token_type_ids handling; different features." + "reason": "Example script classification output bug vs Gemma 4 multimodal token-type-id requirement; unrelated." }, { "left": "issue:39401", "right": "issue:46614", "accept": false, - "reason": "Different tokenizer families and bugs; only both are tokenizer-related." + "reason": "Both are tokenizer-related, but they concern different tokenizers and distinct state bugs." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Qwen3.5 save_pretrained checkpoint issue vs from_pretrained error; related family but not the same concrete bug." + "reason": "Related only at a high level (Qwen3.5 and v5.4.0), but one is save_pretrained checkpoint correctness and the other is from_pretrained loading; not enough evidence of the same concrete bug." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Both mention MoE, but one is Qwen3-VL load and the other FP8 save_pretrained." + "reason": "Different MoE/model-save problems; no shared concrete failure path." }, { - "left": "issue:43824", - "right": "issue:44351", + "left": "issue:44038", + "right": "issue:44295", "accept": false, - "reason": "Different missing imports from different packages/symbols." + "reason": "Unrelated: Qwen3-VL-Moe loading bug vs position_ids buffer access bug." }, { - "left": "issue:44038", - "right": "issue:44295", + "left": "issue:43824", + "right": "issue:44351", "accept": false, - "reason": "Tokenizer/model load bug vs position_ids buffer access bug." + "reason": "Both are import errors, but for different missing exports and likely different root causes." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but different models and symptoms." + "reason": "Both are tokenizer regressions, but they affect different tokenizers/models and are not the same bug." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bug vs example-script classification issue." + "reason": "Completely unrelated: tensor-parallel training vs example-script classification results." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Different missing imports from different libraries." - }, - { - "left": "issue:43278", - "right": "issue:44945", - "accept": false, - "reason": "Training/eval dtype mismatch vs pipeline-parallel output bug; unrelated." - }, - { - "left": "issue:22355", - "right": "issue:29127", - "accept": false, - "reason": "ONNX import failure vs LayoutLMv3 error-message clarity; unrelated." + "reason": "Different missing imports from different packages; no shared underlying issue." } ] }, @@ -14715,21 +14735,21 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ + "issue:22355", "issue:28348", + "issue:29127", "issue:38617", "issue:42175", "issue:42222", "issue:42371", "issue:42673", - "issue:42796", "issue:42898", - "issue:43064", "issue:43066", "issue:43116", + "issue:43278", "issue:43441", "issue:43493", "issue:43611", - "issue:43673", "issue:43825", "issue:43827", "issue:43874", @@ -14746,6 +14766,7 @@ "issue:44779", "issue:44936", "issue:44938", + "issue:44945", "issue:44995", "issue:45072", "issue:45083", @@ -14753,7 +14774,6 @@ "issue:45357", "issue:45446", "issue:45478", - "issue:45563", "issue:45753", "issue:46002", "issue:46231", @@ -14764,6 +14784,8 @@ "issue:46710" ], "soft_pairs": [ + "issue:43278|issue:44945", + "issue:22355|issue:29127", "issue:42175|issue:44938", "issue:43493|issue:45753", "issue:44188|issue:46328", @@ -14771,224 +14793,215 @@ "issue:38617|issue:45278", "issue:42898|issue:46614", "issue:43611|issue:46002", - "issue:43066|issue:44779", "issue:43116|issue:43493", "issue:44355|issue:46231", + "issue:43066|issue:44779", "issue:43976|issue:45446", "issue:44038|issue:45478", - "issue:42673|issue:45072", "issue:44625|issue:46002", "issue:42371|issue:43874", + "issue:42673|issue:45072", "issue:45357|issue:45478", "issue:45083|issue:45446", "issue:43116|issue:43825", - "issue:44075|issue:44936", "issue:42222|issue:43441", + "issue:44075|issue:44936", "issue:43992|issue:44704", - "issue:28348|issue:46640", "issue:44995|issue:46697", + "issue:28348|issue:46640", "issue:43066|issue:46710", - "issue:43827|issue:45446", - "issue:43064|issue:43673", - "issue:43937|issue:45563", - "issue:42673|issue:42796" + "issue:43827|issue:45446" ], - "prepared_review_unit_hash": "1c45d5eeb2076f0fba5156933fab72880a21a72b855222bf8948feb8c2175d77", + "prepared_review_unit_hash": "c5d9d1d27830db28ecfb03f1f6cc7a559abbad2a2416c44fa46cb0892c6701dd", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35185, - "estimated_input_tokens": 8797, - "estimated_eval_tokens": 17850 + "soft_pair_count": 27, + "serialized_chars": 34822, + "estimated_input_tokens": 8706, + "estimated_eval_tokens": 17668 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a loose cluster of Transformers v5 regressions and model-specific bugs rather than true duplicates. The soft links are title-level similarities, not the same concrete issue. No PRs were provided. The broadest umbrella issue is 45278, but the cluster is heterogeneous.", - "confidence": 0.84, - "canonical_issue_reason": "issue:45278 is the broadest umbrella report for post-v5 import/load regressions and can serve as the closest canonical issue, though it is still too generic to subsume the rest cleanly.", + "summary": "Mostly unrelated closed issues with superficial title similarity around Transformers v5 regressions; all soft edges should be rejected. No PRs are present.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across tokenization, model loading, training, vision, packaging, and docs; the soft matches are mostly broad v5-era similarity rather than the same underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most suitable single issue for triage because it covers a broad class of v5 breakages and is more general than the model-specific reports.", + "best_issue_reason": "issue:45278 is the best umbrella representative because it is broad, upgrade-related, and closest to the cluster\u2019s common theme of v5 breakage/import errors, even though it is not a true duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Both mention training/evaluation behavior, but one is a dtype mismatch between train and eval while the other is incorrect output under pipeline parallelism; different code paths and symptoms." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Importing transformers.onnx failing is unrelated to LayoutLMv3 error-message clarity for bounding boxes; different subsystem and bug type." + }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Different failures: backend dependency mismatch for pip extra vs Python 3.14 import/load breakage." + "reason": "Missing TensorFlow in a torch extra install is a packaging/backend issue, while Python 3.14 load failure is an interpreter-compatibility issue; not the same defect." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "Unrelated model bugs: SigLIP2 implementation discrepancy vs Qwen3_5 image-specific infinite loop." + "reason": "SigLIP2 HF/JAX numerical discrepancy and Qwen3_5 infinite-loop-on-image are different models and different failure modes." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "Different areas: attention-kernel divergence under torch.compile vs a SequenceFeatureExtractor padding performance issue." + "reason": "Attention-kernel divergence under torch.compile is unrelated to a performance issue in SequenceFeatureExtractor.pad." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GLM-5 generation config validation error is not the same as a RoPE implementation discussion." + "reason": "GenerationConfig invalid for GLM-5 and RoPE implementation discussion are distinct model issues; not the same concrete bug." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both mention import problems, but one is a specific missing symbol and the other is a broad post-upgrade import-error roundup." + "reason": "A missing layer_type_validation import is a specific symbol breakage, while 45278 is a broad multi-import regression after upgrade; similar theme but not the same underlying issue." }, { "left": "issue:42898", "right": "issue:46614", "accept": false, - "reason": "Different tokenizer/language-path bugs: v5 tokenization cleanup behavior vs stale phonemizer backend state." + "reason": "Tokenizer cleanup behavior change and stale phonemizer backend after language switching are unrelated text-processing vs phonemizer state bugs." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Base-model-prefix loading regression is unrelated to NaNs in classification heads after checkout/init." - }, - { - "left": "issue:43066", - "right": "issue:44779", - "accept": false, - "reason": "Both are tokenizer-related, but one is a wrong decoder type in v5 and the other is Deepseek tokenization output corruption." + "reason": "base_model_prefix loading breakage and NaNs after checkout/init affect different code paths and have different root causes." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example-script multi-label output bug is unrelated to SigLIP2 implementation mismatch." + "reason": "Multi-label example script returning empty results is unrelated to a SigLIP2 implementation discrepancy." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors are not the same as a documentation parameter-name mismatch." + "reason": "Compiled Python file errors and a documentation parameter-name typo are not the same bug." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but one is decoder-type selection and the other is incorrect DeepSeek tokenization results; too different to merge." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python-version compatibility issue is distinct from an incorrect torch version check in flex_attention." + "reason": "Python version compatibility and a wrong PyTorch version check for AuxRequest are separate environment-check problems." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Related to Qwen3.5/Moe, but these describe different from_pretrained failures and are not the same code-path bug." - }, - { - "left": "issue:42673", - "right": "issue:45072", - "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel CI." + "reason": "Qwen3-VL-Moe and Qwen3.5 Moe from_pretrained errors are related by model family, but the concrete failure paths are different." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Qwen3.5 num_labels propagation bug is separate from NaNs in classification heads." + "reason": "num_labels propagation in Qwen3.5 config is unrelated to NaNs in classification heads." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 API warning is unrelated to a missing image-patch-count method in GLM46VImageProcessorFast." + "reason": "TF32 behavior settings and a missing image-patch-count method in GLM46VImageProcessorFast are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "A VRAM leak in multi-threaded dataloader use and dtype mismatches in bfloat16 inference are different bugs." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5/Moe, but one is save_pretrained key regression and the other is from_pretrained load error." + "reason": "Incorrect visual encoder keys on save_pretrained and from_pretrained failure are both Qwen3.5-related, but they are opposite sides of different serialization/loading bugs." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Feature-extraction length helper behavior is unrelated to the flex_attention version-check bug." + "reason": "Feature-extractor length calculation in qwen3_omni_moe is unrelated to a PyTorch version gate for AuxRequest import." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Classification example output bug is unrelated to a pipeline() error message about translation support." + "reason": "Example-script multi-label classification failure and pipeline() translation messaging after v5 removals are different issues." }, { - "left": "issue:44075", - "right": "issue:44936", + "left": "issue:42222", + "right": "issue:43441", "accept": false, - "reason": "Optimizer SGD arguments not being used is not the same as trainer.evaluate() failing after trainer.train()." + "reason": "Vitpose model breakage and Ministral-3 flash-attention failure are unrelated model-specific regressions." }, { - "left": "issue:42222", - "right": "issue:43441", + "left": "issue:44075", + "right": "issue:44936", "accept": false, - "reason": "Vitpose model breakage and Ministral FlashAttention failures are different model-specific issues." + "reason": "Ignored SGD optimizer args and trainer.evaluate() failing after train are distinct training-loop bugs." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens weight loading is unrelated to AutoProcessor kwargs forwarding." + "reason": "UMT5Encoder missing embed_tokens.weight during loading and AutoProcessor not forwarding kwargs to cached_file are different loading/processor paths." }, { - "left": "issue:28348", - "right": "issue:46640", + "left": "issue:44995", + "right": "issue:46697", "accept": false, - "reason": "Both mention Flash Attention 2, but one is GPT2 support and the other is a generic flash-attention issue with no shared concrete code path." + "reason": "Stale indexer cache crash on second forward pass and Swin2SR over-padding are unrelated model/runtime vs image-processing bugs." }, { - "left": "issue:44995", - "right": "issue:46697", + "left": "issue:28348", + "right": "issue:46640", "accept": false, - "reason": "Stale indexer cache crash in GlmMoeDsa is unrelated to Swin2SR image over-padding." + "reason": "FlashAttention 2 support request for GPT2LMHeadModel and a vague flash-attention issue are too underspecified and not the same concrete bug." }, { "left": "issue:43066", "right": "issue:46710", "accept": false, - "reason": "Tokenizer decoder-type regression is unrelated to DeepSeek accuracy regression after upgrade." + "reason": "Wrong tokenizer decoder type and DeepSeek-R1-Distill-Llama-8B accuracy regression are both v5 regressions, but they involve different models and failure modes." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs referencing pipeline() and a flex_attention version check are different changes, not the same bug." - }, - { - "left": "issue:43064", - "right": "issue:43673", - "accept": false, - "reason": "Trainer optimizer-state bug under v5/FSDP2/PEFT is unrelated to missing GenerationMixin cache in chunked prefill." - }, - { - "left": "issue:43937", - "right": "issue:45563", - "accept": false, - "reason": "GenerationConfig validation failure and a stale warning for num_return_sequences are different generation issues." - }, - { - "left": "issue:42673", - "right": "issue:42796", - "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to a tokenizer BPE-marker handling bug." + "reason": "Stale docs referencing pipeline() and a PyTorch version check bug are unrelated documentation vs runtime-import issues." } ] }, @@ -15020,6 +15033,8 @@ "issue:42371", "issue:42617", "issue:42673", + "issue:42796", + "issue:43064", "issue:43122", "issue:43408", "issue:43450", @@ -15033,8 +15048,6 @@ "issue:44380", "issue:44410", "issue:44466", - "issue:44484", - "issue:44496", "issue:44779", "issue:44829", "issue:44877", @@ -15049,6 +15062,7 @@ "issue:45405", "issue:45446", "issue:45561", + "issue:45563", "issue:45588", "issue:46002", "issue:46077", @@ -15058,249 +15072,248 @@ "issue:46328", "issue:46512", "issue:46614", - "issue:46617", "issue:46655" ], "soft_pairs": [ + "issue:43064|issue:43673", + "issue:43937|issue:45563", "issue:33453|issue:39401", + "issue:42673|issue:42796", "issue:44380|issue:46077", "issue:45356|issue:46614", - "issue:42673|issue:46655", "issue:43673|issue:46328", + "issue:42673|issue:46655", "issue:44291|issue:45071", - "issue:45405|issue:45446", "issue:45071|issue:46614", - "issue:44117|issue:45588", + "issue:45405|issue:45446", "issue:44466|issue:46002", + "issue:44117|issue:45588", "issue:44410|issue:45310", "issue:39401|issue:44779", "issue:43937|issue:44977", "issue:45071|issue:46512", "issue:45341|issue:45561", "issue:45341|issue:46328", - "issue:43450|issue:43986", "issue:44829|issue:45216", + "issue:43450|issue:43986", "issue:44117|issue:45081", "issue:38617|issue:41628", "issue:42617|issue:45341", "issue:43122|issue:46298", "issue:45362|issue:46328", "issue:43408|issue:43742", - "issue:42222|issue:46178", "issue:42371|issue:44261", - "issue:44877|issue:46231", - "issue:44484|issue:45310", - "issue:42371|issue:46617", - "issue:43937|issue:44496" + "issue:42222|issue:46178", + "issue:44877|issue:46231" ], - "prepared_review_unit_hash": "f898e275d89eb0f25c1a45b95cd7af6cd42e05e987ad13057088bcc9a4df9e91", + "prepared_review_unit_hash": "8dfa9f4c947dd0ff97e5471775fa53d4981daa0d7f641a4df88f216fd1ad2eef", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36708, - "estimated_input_tokens": 9177, - "estimated_eval_tokens": 18610 + "serialized_chars": 36721, + "estimated_input_tokens": 9181, + "estimated_eval_tokens": 18618 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose cluster of unrelated Transformers issues. None of the soft pairs look like the same underlying bug/change, and there are no PRs to compare.", - "confidence": 0.67, - "canonical_issue_reason": "issue:43122 is the broadest and clearest version-regression report in the set: it describes tokenizer behavior changing across releases with a reproducible before/after comparison, so it is the best umbrella candidate among these issues.", + "summary": "This cluster is a loose mix of tokenizer regressions, model-loading failures, attention/backend bugs, and test-utils issues; the soft links mostly reflect broad topical similarity rather than true duplicates.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43122 is the broadest umbrella for the tokenizer-regression subset: it explicitly compares v4.57.3 vs v5 behavior and can cover several related tokenizer reports better than the more model-specific issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the strongest standalone duplicate target because it is a general tokenizer regression claim rather than a model-specific one, making it the most plausible anchor for related reports.", + "best_issue_reason": "issue:44779 is the most concrete standalone bug report here: it names a specific model, a version regression, and a user-visible incorrect tokenization result, making it the strongest actionable issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Different failures: optimizer state placement in FSDP/PEFT vs generation cache missing." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Unrelated generation bugs: invalid GenerationConfig vs a stale warning for num_return_sequences." + }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both are tokenizer-related, but one is about loading regressions while the other is about offset_mapping correctness for Qwen3; different failure modes and code paths." + "reason": "Both are tokenizer-related, but one is loading regression and the other is offset_mapping correctness for Qwen3." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in multi-threaded dataloading is unrelated to tokenizer BPE marker handling." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "Both involve attention behavior, but they concern different models and different bugs: GPT-2 scaling config vs Gemma4 bidirectional mask construction." + "reason": "Both involve attention behavior, but they are different models/backends and different bugs." }, { "left": "issue:45356", "right": "issue:46614", "accept": false, - "reason": "Both are tokenizer/processor bugs, but one is a Kimi tokenizer codec regression and the other is a Wav2Vec2 phonemizer backend state bug; not the same underlying issue." - }, - { - "left": "issue:42673", - "right": "issue:46655", - "accept": false, - "reason": "VRAM leakage in multithreaded generation is unrelated to incorrect video padding in a processor." + "reason": "Tokenizer codec/fix_mistral_regex regression is unrelated to stale phonemizer backend state." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Chunked-prefill cache handling and padding performance in SequenceFeatureExtractor are different subsystems and different bugs." + "reason": "Generation cache missing and SequenceFeatureExtractor.pad inefficiency are unrelated components." }, { - "left": "issue:44291", - "right": "issue:45071", + "left": "issue:42673", + "right": "issue:46655", "accept": false, - "reason": "Unexpected constructor argument with init_empty_weights is a loading/API incompatibility, while the other is a PreTrainedConfig type-checking regression." + "reason": "VRAM leak in dataloader threads is unrelated to video processor padding calculation." }, { - "left": "issue:45405", - "right": "issue:45446", + "left": "issue:44291", + "right": "issue:45071", "accept": false, - "reason": "A dependency version bump issue is not the same as a PyTorch version-check bug in flex_attention." + "reason": "Both are v5-loading/type issues, but one is an unexpected init arg and the other is PretrainedConfig type checking." }, { "left": "issue:45071", "right": "issue:46614", "accept": false, - "reason": "PreTrainedConfig type checking and stale phonemizer language state are unrelated." + "reason": "PretrainedConfig type-checking regression is unrelated to phonemizer backend persistence." }, { - "left": "issue:44117", - "right": "issue:45588", + "left": "issue:45405", + "right": "issue:45446", "accept": false, - "reason": "A tokenizer mapping returning None is unrelated to a flash-attention crash on sink-less models." + "reason": "PEFT version bump problem is unrelated to a PyTorch version gate in flex_attention." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "Serialization of tied weights depending on device is a save/load issue; NaNs in classification heads during checkout/init is a separate numerical bug." + "reason": "Serialization of tied weights is unrelated to NaNs in classification heads." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES fallback bug is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Both mention Qwen3.5-family models, but one is missing projections in a layer definition and the other is a from_pretrained loading error; not the same concrete defect." + "reason": "Different Qwen3-related issues: missing projections in one model versus a from_pretrained error in another." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer correctness regressions, but they are model-specific and the descriptions do not indicate the same tokenizer or the same root cause." + "reason": "Both are tokenizer regressions, but they affect different models and different symptoms." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GenerationConfig validation failure and flash-attention generation issues are different bugs." + "reason": "GenerationConfig validation and flash-attention generation failure are different code paths." }, { "left": "issue:45071", "right": "issue:46512", "accept": false, - "reason": "Type-checking breakage in PreTrainedConfig is unrelated to from_config(dtype=...) leaving config.dtype stale." + "reason": "PretrainedConfig type checking and stale config.dtype after from_config are separate bugs." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "The shared testing_utils area is too broad; one is an unspecified small bug and the other is a specific xdist race on captured_info.txt." + "reason": "Generic testing_utils bug and xdist race on captured_info.txt are not clearly the same defect." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "Testing utils and SequenceFeatureExtractor padding are unrelated." + "reason": "Testing utils and SequenceFeatureExtractor.pad inefficiency are unrelated." }, { - "left": "issue:43450", - "right": "issue:43986", + "left": "issue:44829", + "right": "issue:45216", "accept": false, - "reason": "Incorrect batched video processor output shape and a torchvision-missing crash in AutoProcessor are different video-processing failures." + "reason": "FlashAttention training degeneration is unrelated to Qwen3.5 checkpoint serialization." }, { - "left": "issue:44829", - "right": "issue:45216", + "left": "issue:43450", + "right": "issue:43986", "accept": false, - "reason": "Degenerate training with flash_attention_3 is not the same as incorrect checkpoint serialization in save_pretrained." + "reason": "Video processor shape bug and AutoProcessor crash without torchvision are different problems." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping assumptions and the fix_mistral_regex backend_tokenizer crash are separate tokenizer-loading issues with different failure points." + "reason": "Mapping-name None handling and Mistral regex patch crashing are unrelated tokenizer-loading failures." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different code paths." + "reason": "Both are import errors, but they concern different missing exports and different APIs." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "A script execution problem is not the same as a testing_utils bug." + "reason": "3d_parallel.py runtime failure is unrelated to a testing_utils bug." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenizer regression vs noisy logging message; no shared underlying bug." + "reason": "Tokenizer regression is unrelated to a noisy logging message." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen3.5 chat crashes and a padding performance issue are unrelated." + "reason": "Qwen3.5 chat crash is unrelated to SequenceFeatureExtractor.pad performance." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A model-type mismatch warning is unrelated to a key error when loading a different model." - }, - { - "left": "issue:42222", - "right": "issue:46178", - "accept": false, - "reason": "Vitpose model breakage and an unsupported backend mapping error are different issues." + "reason": "SAM3 model-type warning is unrelated to a MobileLLM loading key error." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 API guidance and a missing rms_norm_eps precision bug are unrelated." - }, - { - "left": "issue:44877", - "right": "issue:46231", - "accept": false, - "reason": "A strict config loading failure is not the same as a documentation parameter-name mismatch." - }, - { - "left": "issue:44484", - "right": "issue:45310", - "accept": false, - "reason": "Default shard size behavior in save_pretrained and a Qwen3.5 from_pretrained error do not share a concrete bug path." + "reason": "TF32 API guidance is unrelated to MLA rms_norm_eps precision behavior." }, { - "left": "issue:42371", - "right": "issue:46617", + "left": "issue:42222", + "right": "issue:46178", "accept": false, - "reason": "TF32 settings documentation/runtime behavior is unrelated to cache loading failures for trust_remote_code models." + "reason": "vitpose breakage and tensorflow_text backend mapping are unrelated." }, { - "left": "issue:43937", - "right": "issue:44496", + "left": "issue:44877", + "right": "issue:46231", "accept": false, - "reason": "An invalid GenerationConfig and an unrecognized-model config.json error are different loading/validation problems." + "reason": "Strict config loading for granite_speech is unrelated to a documentation parameter-name mismatch." } ] }, @@ -15330,6 +15343,7 @@ "issue:41628", "issue:42175", "issue:42222", + "issue:42371", "issue:43012", "issue:43066", "issue:43097", @@ -15337,15 +15351,14 @@ "issue:43352", "issue:43441", "issue:43452", - "issue:43531", "issue:43577", "issue:43673", "issue:43874", "issue:43931", + "issue:43937", "issue:43986", "issue:44038", "issue:44075", - "issue:44117", "issue:44220", "issue:44261", "issue:44263", @@ -15355,40 +15368,42 @@ "issue:44484", "issue:44485", "issue:44493", + "issue:44496", "issue:44829", "issue:44933", - "issue:44987", "issue:45003", "issue:45071", "issue:45278", + "issue:45310", "issue:45341", - "issue:45440", "issue:45464", "issue:45478", "issue:45542", "issue:45656", - "issue:45753", "issue:46172", "issue:46209", "issue:46231", - "issue:46369" + "issue:46617" ], "soft_pairs": [ + "issue:44484|issue:45310", + "issue:42371|issue:46617", + "issue:43937|issue:44496", "issue:39401|issue:46209", "issue:44410|issue:45478", "issue:44933|issue:45278", "issue:38617|issue:44220", "issue:44075|issue:45656", "issue:44263|issue:44485", - "issue:43097|issue:44493", "issue:43066|issue:43122", - "issue:42222|issue:45542", + "issue:43097|issue:44493", "issue:43673|issue:45341", + "issue:42222|issue:45542", "issue:41628|issue:45071", - "issue:42175|issue:42222", "issue:43874|issue:44261", "issue:45071|issue:45542", "issue:43931|issue:44410", + "issue:42175|issue:42222", "issue:44483|issue:45464", "issue:44484|issue:45478", "issue:44360|issue:46172", @@ -15400,240 +15415,230 @@ "issue:43097|issue:43452", "issue:45003|issue:45341", "issue:43986|issue:45341", - "issue:44933|issue:46231", "issue:43874|issue:44263", - "issue:44261|issue:44485", - "issue:45753|issue:46369", - "issue:44117|issue:44987", - "issue:43531|issue:45341", - "issue:45440|issue:46172" + "issue:44933|issue:46231", + "issue:44261|issue:44485" ], - "prepared_review_unit_hash": "0a5c128d215018c56333c8c58042599606671a49770234c5e3c7b7cc1ab22b18", + "prepared_review_unit_hash": "8d1a542a68d7018b1c4bbc65484b490e501b17848a470e9eaf4d27b37c11bd39", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37005, - "estimated_input_tokens": 9252, - "estimated_eval_tokens": 18760 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36184, + "estimated_input_tokens": 9046, + "estimated_eval_tokens": 18348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pairs look like true duplicates under a strict same-bug/same-code-path standard, so all candidate merges are rejected. The set is mostly a mix of model-specific regressions, unrelated import/dependency issues, and different tokenizer or generation bugs.", - "confidence": 0.89, - "canonical_issue_reason": "issue:44117 is the clearest actionable issue: it names a concrete failure mode in from_pretrained, points to a likely single code-path assumption, and is broadly useful for fixing tokenizer loading.", + "summary": "This is a loose similarity cluster, not a true duplicate set: it mixes unrelated model bugs, tokenizer regressions, backend/install issues, and API failures. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.79, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44117 is the strongest representative issue because it is specific, reproducible, and likely to lead to a targeted fix rather than a broad or ambiguous cleanup.", + "best_issue_reason": "issue:45278 is the broadest representative of the cluster\u2019s recurring upgrade/import-breakage theme, even though the cluster itself is heterogeneous.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Different problems: save_pretrained shard sizing vs Qwen3.5 MoE loading failure." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 settings warning and trust_remote_code cache FileNotFoundError are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GLM generation config validation vs Olmo model type/config loading error." + }, { "left": "issue:39401", "right": "issue:46209", "accept": false, - "reason": "Different bugs: tokenizer offset_mapping vs FP8 inference on B200; unrelated layers of the stack." + "reason": "Tokenizer offset mapping bug and FP8 inference on B200 are different code paths." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Both Qwen3-family, but one is missing layer projections and the other is a from_pretrained error; different code paths." + "reason": "Both mention Qwen-family models, but the reported failures are distinct loading/projection bugs." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "One missing image_utils export, the other broad import regressions; no evidence they are the same symbol breakage." + "reason": "A missing/nonexistent import and broad upgrade import errors are not the same concrete bug." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for configuration_utils vs audio feature extraction issue; unrelated failures." + "reason": "Configuration import error and fbank feature extraction issue are unrelated." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "Optimizer hyperparameter handling vs double-step with DeepSpeed; different trainer/optimizer bugs." + "reason": "Optimizer args ignored vs double optimizer.step under deepspeed are different defects." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "GlmMoeDsaIndexer split semantics vs GLM-5 RoPE implementation; distinct components." - }, - { - "left": "issue:43097", - "right": "issue:44493", - "accept": false, - "reason": "Removed tie_embeddings flag vs unexpected position_id keys; unrelated regressions." + "reason": "DSA indexer split-return issue and GLM-5 RoPE implementation bug are different." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Tokenizer decoder-type mismatch vs tokenization output drift; too broad to be the same concrete defect." + "reason": "Both are tokenizer-related v5 regressions, but not the same underlying failure." }, { - "left": "issue:42222", - "right": "issue:45542", + "left": "issue:43097", + "right": "issue:44493", "accept": false, - "reason": "VitPose model breakage vs missing TensorFlow backend dependency; unrelated." + "reason": "Removed tie_embeddings API and unexpected position-id keys are unrelated regressions." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache/chunked_prefill bug vs testing_utils issue; unrelated." + "reason": "Generation cache/chunked_prefill issue and a testing_utils bug are unrelated." }, { - "left": "issue:41628", - "right": "issue:45071", + "left": "issue:42222", + "right": "issue:45542", "accept": false, - "reason": "Missing AutoImageProcessor export vs PreTrainedConfig type checking; different failures." + "reason": "Vitpose model breakage and TF/tensorboard backend packaging error do not share a code path." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:41628", + "right": "issue:45071", "accept": false, - "reason": "Backend dependency packaging vs VitPose model loading; unrelated." + "reason": "AutoImageProcessor import failure and PreTrainedConfig type-checking breakage are different issues." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Image patch count helper vs RMS norm epsilon config; different multimodal submodules." + "reason": "Missing image patch helper and missing rms_norm_eps in MLA are unrelated model bugs." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Type-checking regression vs TensorFlow backend installation issue; unrelated." + "reason": "Config type checking and TF backend dependency handling are different regressions." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch vs Qwen3Next missing projections; different model architectures and defects." + "reason": "Qwen3-VL weight-shape mismatch and qwen3next projection omission are distinct model-loading issues." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging and vitpose breakage are not the same bug." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Generic chat/completions rejection vs Qwen3.5 streaming failure; different endpoint behaviors and likely different fixes." + "reason": "Both concern chat/completions, but one is API acceptance and the other is streaming inference failure." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "save_pretrained shard-size question vs Qwen3.5 MoE from_pretrained error; unrelated." + "reason": "save_pretrained shard sizing and Qwen3.5 MoE from_pretrained error are unrelated." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "DSA indexer ReLU omission vs DeepSeek-V4 transpose/cast issue; different code paths." + "reason": "Missing ReLU in the DSA indexer and deepseek-v4 comb.transpose bug are different implementation errors." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "GlmMoeDsaIndexer split outputs vs DeepSeek-V4 dtype/transpose bug; unrelated." + "reason": "torch.split return handling and deepseek-v4 tensor transpose are unrelated." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both dtype-related, but compile-time warning vs model-loading cast bug; not the same defect." + "reason": "bfloat16 precision warning and BLIP2 dtype propagation bug are different symptoms and code paths." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Flash Attention 2 support on different models; separate model-specific enablement issues." + "reason": "XLMRoberta Flash Attention 2 support and NemotronHForCausalLM Flash Attention 2 support are model-specific, separate requests." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Different Qwen3-VL family failures with different symptoms and loaders; not the same bug." + "reason": "Qwen3-VL shape mismatch and Qwen3-VL-Moe loading bug are not the same failure." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "FlashAttention on Ministral-3 vs FA3 degenerate training in sequence classification; different code paths." + "reason": "FlashAttention load failure in Ministral-3 and flash_attention_3 training degeneration are different concrete problems." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Embedding-tying removal vs gguf_file loading breakage; unrelated." + "reason": "Tie-embeddings API removal and gguf_file/from_pretrained behavior are unrelated." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "sys.modules access bug vs testing_utils issue; unrelated." + "reason": "sys.modules access bug and testing_utils bug are unrelated." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "AutoProcessor video-model/torchvision crash vs testing utility bug; unrelated." - }, - { - "left": "issue:44933", - "right": "issue:46231", - "accept": false, - "reason": "Image_utils export issue vs training doc parameter typo; unrelated." + "reason": "AutoProcessor/video-model torchvision dependency crash and testing_utils bug are unrelated." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Image processor helper missing vs DSA indexer split issue; unrelated." - }, - { - "left": "issue:44261", - "right": "issue:44485", - "accept": false, - "reason": "Both GLM-5-ish, but layernorm eps config vs RoPE implementation; separate bugs." - }, - { - "left": "issue:45753", - "right": "issue:46369", - "accept": false, - "reason": "Qwen3_5 image infinite loop vs sam3-litetext memory overflow; unrelated." - }, - { - "left": "issue:44117", - "right": "issue:44987", - "accept": false, - "reason": "Generic from_pretrained/tokenizer mapping bug vs specific model load failure; not the same concrete path." + "reason": "These are separate GLM-family bugs affecting different helpers." }, { - "left": "issue:43531", - "right": "issue:45341", + "left": "issue:44933", + "right": "issue:46231", "accept": false, - "reason": "Qwen3-MoE sliding_window bug vs testing_utils typo; unrelated." + "reason": "Nonexistent image_utils import and a doc mismatch in set_training() are not duplicates." }, { - "left": "issue:45440", - "right": "issue:46172", + "left": "issue:44261", + "right": "issue:44485", "accept": false, - "reason": "DeepSeekV3MoE divergence vs DeepSeek-V4 dtype/transpose issue; different models and defects." + "reason": "Missing rms_norm_eps precision issue and GLM-5 RoPE implementation issue are distinct." } ] }, @@ -15661,17 +15666,15 @@ "issue:42757", "issue:42890", "issue:42907", - "issue:43012", "issue:43064", - "issue:43066", "issue:43097", "issue:43208", "issue:43502", + "issue:43531", "issue:43550", "issue:43673", "issue:43825", "issue:43873", - "issue:43906", "issue:43986", "issue:43992", "issue:44038", @@ -15679,7 +15682,6 @@ "issue:44117", "issue:44315", "issue:44351", - "issue:44368", "issue:44380", "issue:44410", "issue:44484", @@ -15687,9 +15689,12 @@ "issue:44704", "issue:44936", "issue:44938", + "issue:44987", "issue:45103", + "issue:45341", "issue:45372", "issue:45375", + "issue:45440", "issue:45588", "issue:45712", "issue:45753", @@ -15699,6 +15704,7 @@ "issue:46178", "issue:46298", "issue:46365", + "issue:46369", "issue:46421", "issue:46655", "issue:46664", @@ -15706,19 +15712,23 @@ "issue:46710" ], "soft_pairs": [ + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43531|issue:45341", "issue:43673|issue:46123", + "issue:45440|issue:46172", "issue:43097|issue:45588", "issue:38617|issue:42757", "issue:43986|issue:45372", - "issue:43064|issue:44315", "issue:44410|issue:45375", + "issue:43064|issue:44315", "issue:43097|issue:46050", "issue:44038|issue:44410", "issue:46655|issue:46664", "issue:43986|issue:44484", "issue:38617|issue:44351", - "issue:46172|issue:46710", "issue:42890|issue:46298", + "issue:46172|issue:46710", "issue:43097|issue:43873", "issue:44062|issue:44117", "issue:44117|issue:46701", @@ -15732,226 +15742,222 @@ "issue:42907|issue:44655", "issue:42907|issue:45753", "issue:43208|issue:43550", - "issue:45712|issue:46298", - "issue:43012|issue:43550", - "issue:43012|issue:44368", - "issue:43992|issue:44484", - "issue:43066|issue:43906" + "issue:45712|issue:46298" ], - "prepared_review_unit_hash": "7266618240254368e3a2a7ccc94e4705228fe95dece5889312063cbb7aae4edd", + "prepared_review_unit_hash": "c8b44b7b6b692f29c06cb852952be38d76c784a8378dda4266b3590a90c64b74", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36348, - "estimated_input_tokens": 9087, - "estimated_eval_tokens": 18430 + "serialized_chars": 36763, + "estimated_input_tokens": 9191, + "estimated_eval_tokens": 18638 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft pairs are overwhelmingly false positives: they share broad Transformers themes but describe different bugs, different code paths, or unrelated regressions. No duplicate merge is justified from the evidence provided.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Mostly unrelated Transformers bug reports with topical overlap only; none of the soft pairs look like true duplicates or mergeable single-change PRs.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43097 is the best umbrella representative: it is a concrete, high-signal v5 regression in a core code path and broader than the more model-specific or incidental reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43502 is the strongest standalone bug report here: it has a concrete, reproducible symptom (network calls despite local_files_only=True) and a clear impact surface, making it the best representative issue if one must be picked.", + "best_issue_reason": "issue:43097 is the strongest standalone issue to keep: clear impact, core compatibility regression, and enough context to guide a fix without being overly narrow.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different bugs: one is a Qwen3_5 image-loop issue, the other is a Trainer import failure tied to datasets." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Both involve loading/tokenizer paths, but the failures are distinct: mapping lookup returning None vs a model load break in v5.1.0." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Unrelated: a Qwen3-MoE sliding_window bug versus a testing_utils bug." + }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Different failures: chunked_prefill cache initialization vs final partial-batch handling in MaskGenerationPipeline." + "reason": "Both touch generation/pipeline behavior, but one is chunked_prefill cache handling and the other is final-batch flag logic in MaskGenerationPipeline." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Different DeepSeek-related code paths and symptoms; not the same underlying change or bug." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "Unrelated problems: removed tie_embeddings API vs flash_attention crash on s_aux=None." + "reason": "Core issues differ: a removed config attribute in v5 versus a flash-attention crash on sink-less models." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different missing symbols in different modules and likely different dependency breakages." + "reason": "Both are ImportErrors, but for different symbols from different modules; no shared root cause shown." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Different loading paths and causes: missing torchvision for AutoProcessor video loading vs mistral_common ReasoningEffort import failure." + "reason": "Different dependency/loading failures: missing torchvision for video models versus missing ReasoningEffort from mistral_common." }, { - "left": "issue:43064", - "right": "issue:44315", + "left": "issue:44410", + "right": "issue:45375", "accept": false, - "reason": "Different training bugs: optimizer state corruption under FSDP2/PEFT vs Liger Kernel not applied with model_init." + "reason": "Both are Qwen-family config/model issues, but they concern different models and different missing fields/projections." }, { - "left": "issue:44410", - "right": "issue:45375", + "left": "issue:43064", + "right": "issue:44315", "accept": false, - "reason": "Different model/config defects: missing projections in qwen3next vs a config field dropped by @strict." + "reason": "Trainer/FSDP optimizer-state corruption is unrelated to Liger Kernel application in model_init." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "Different save/load concerns: removed encoder-decoder tie flag vs quantized model broken after save_pretrained/from_pretrained." + "reason": "A v5 config/API removal issue is not the same as quantized save/load breakage." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Different Qwen-family issues with no shared concrete code-path or symptom." + "reason": "Related family name, but the concrete failures are different Qwen variants and different missing components." }, { "left": "issue:46655", "right": "issue:46664", "accept": false, - "reason": "Both touch SmolVLM video processing, but one is padding math and the other is num_frames semantics; not the same bug." + "reason": "Both are SmolVLM video-processor bugs, but one is padding math and the other is num_frames semantics." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "Different areas: AutoProcessor video loading crash vs save_pretrained shard-size default behavior." + "reason": "Different areas: missing torchvision during processor loading versus save_pretrained shard-size behavior." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are missing-import reports, but for different exported names and unrelated modules." + "reason": "Both are import-name errors, but they involve different exported symbols and likely different fixes." }, { - "left": "issue:46172", - "right": "issue:46710", + "left": "issue:42890", + "right": "issue:46298", "accept": false, - "reason": "One is a specific DeepSeek-V4 tensor operation bug; the other is a broader accuracy regression after upgrading Transformers." + "reason": "A test seeding problem is unrelated to a noisy logging message." }, { - "left": "issue:42890", - "right": "issue:46298", + "left": "issue:46172", + "right": "issue:46710", "accept": false, - "reason": "Test flakiness from missing set_seed is unrelated to a noisy logging message." + "reason": "Different DeepSeek issues: a specific tensor operation bug versus an accuracy regression after upgrade." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Removed config flag vs quantization/offloading behavior; not the same underlying issue." + "reason": "One is about an API/config removal; the other is an offloading/quantization malfunction." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Tokenizer constructor TypeError and TOKENIZER_MAPPING_NAMES returning None are distinct bugs." + "reason": "Tokenizer constructor argument handling is unrelated to TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:44117", "right": "issue:46701", "accept": false, - "reason": "Unrelated: tokenizer mapping assumption vs DiffusionGemma pipeline failure." + "reason": "Tokenizer mapping fallback and an image-text-to-text pipeline failure are separate code paths." }, { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "Trainer.evaluate after train and importing Trainer being blocked by datasets are different problems." + "reason": "Trainer.evaluate failing after train is not the same as failing to import Trainer because datasets is installed." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Different from_pretrained issues: unwanted network access vs missing embed_tokens.weight." + "reason": "Local-files-only network leakage and a missing embed_tokens.weight during UMT5 loading are different defects." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor kwargs forwarding bug is unrelated to auto_docstring crashing under future annotations." + "reason": "AutoProcessor kwargs forwarding and an auto_docstring annotation crash are unrelated." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "GPT2 attention scaling being ignored is distinct from chunked-prefill cache recompilation." + "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to cache lazy initialization and recompilation." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import failure and unsupported tensorflow_text backend are different failure modes." + "reason": "Python 3.14 import/load failure is not the same as an unsupported tensorflow_text backend mapping." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Different subsystems: video-model loading dependency issue vs tokenizer mapping None handling." + "reason": "Different failure modes and layers: video dependency loading versus tokenizer mapping resolution." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline error messaging and saving Pipeline objects are unrelated issues." + "reason": "Pipeline task-support messaging and saving Pipeline objects are separate problems." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Dequantized model save failure is unrelated to Pipeline.save_pretrained failure." + "reason": "Dequantized model save failure is different from inability to save Pipeline objects." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Different model-specific bugs: dequantized save failure vs Qwen3_5 image-induced infinite loop." + "reason": "Quantized save/load behavior and a Qwen3_5 image infinite loop are unrelated." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs and Bamba torch.compile/SDPA failure are unrelated." + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure do not share a concrete code path." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Leftover dummy classes leaking into dir(transformers) is unrelated to a noisy logging message." - }, - { - "left": "issue:43012", - "right": "issue:43550", - "accept": false, - "reason": "A bfloat16 compile warning is not the same as a model failure under torch.compile with SDPA." - }, - { - "left": "issue:43012", - "right": "issue:44368", - "accept": false, - "reason": "Different warnings with different causes: PyTorch float32 precision warning vs tie_word_embeddings config warning." - }, - { - "left": "issue:43992", - "right": "issue:44484", - "accept": false, - "reason": "Missing weight load in from_pretrained is unrelated to the max_shard_size default question." - }, - { - "left": "issue:43066", - "right": "issue:43906", - "accept": false, - "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same bug." + "reason": "Dummy-class cleanup/check_repo issues are unrelated to a noisy logging message." } ] }, @@ -15975,7 +15981,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ - "issue:36010", "issue:36683", "issue:38617", "issue:42175", @@ -15984,54 +15989,58 @@ "issue:42994", "issue:43012", "issue:43064", + "issue:43066", "issue:43097", "issue:43208", "issue:43299", "issue:43352", - "issue:43475", "issue:43502", "issue:43526", + "issue:43550", "issue:43582", - "issue:43611", "issue:43618", "issue:43756", "issue:43761", "issue:43824", "issue:43867", - "issue:43937", + "issue:43906", "issue:43976", + "issue:43992", "issue:44038", "issue:44077", "issue:44162", "issue:44222", "issue:44230", + "issue:44368", "issue:44380", "issue:44483", + "issue:44484", "issue:44485", "issue:44514", "issue:44625", "issue:44655", "issue:44683", - "issue:45276", "issue:45478", "issue:45479", "issue:45584", "issue:45600", - "issue:46077", "issue:46132", "issue:46172", "issue:46298", - "issue:46365", - "issue:46369" + "issue:46365" ], "soft_pairs": [ + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:43992|issue:44484", + "issue:43066|issue:43906", "issue:43352|issue:43976", - "issue:43352|issue:44380", "issue:43097|issue:46298", + "issue:43352|issue:44380", "issue:44162|issue:46172", - "issue:45584|issue:46365", "issue:44380|issue:46132", "issue:44077|issue:46365", + "issue:45584|issue:46365", "issue:44222|issue:45478", "issue:36683|issue:43582", "issue:44380|issue:45479", @@ -16042,248 +16051,230 @@ "issue:43012|issue:44683", "issue:43208|issue:43526", "issue:45600|issue:46298", - "issue:43618|issue:44380", "issue:44483|issue:44514", + "issue:43618|issue:44380", "issue:43618|issue:43761", "issue:44038|issue:44230", "issue:38617|issue:43502", "issue:43756|issue:44485", "issue:43299|issue:43867", "issue:42994|issue:44655", - "issue:44625|issue:46365", - "issue:36683|issue:43475", - "issue:45276|issue:46077", - "issue:36010|issue:42757", - "issue:43352|issue:43937", - "issue:43611|issue:44655", - "issue:44483|issue:46369" + "issue:44625|issue:46365" ], - "prepared_review_unit_hash": "d864858c6882eede911ffa533a80cec9cacd22d2ece46bdb627937378bdfc038", + "prepared_review_unit_hash": "f0fe2aca0d8d784aeb5a96374e4bd4699a278d508d6c7179cf3a636666459c5f", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36906, - "estimated_input_tokens": 9227, - "estimated_eval_tokens": 18710 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35774, + "estimated_input_tokens": 8944, + "estimated_eval_tokens": 18144 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated bug reports, with one plausible duplicate pair around a CLIP v5 output regression. The other soft links look like false positives across different models, APIs, and subsystems.", - "confidence": 0.64, - "canonical_issue_reason": "issue:43761 is the clearest representative of the only convincing duplicate pair: a CLIP v5 regression where optional outputs are no longer populated.", + "summary": "Mostly a mixed bag of unrelated Transformers issues; the only strong duplicate-like pair is the CLIP v5 output regression pair. No PRs are present.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43761 is the cleanest representative of the CLIP v5 regression cluster: it states the concrete forward-pass symptom, includes the explicit regression tag, and is broader than the single attentions-only symptom in issue:43618.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43761 is the best representative in this cluster because it is the most concrete ticket in the only likely duplicate pair; the rest are unrelated.", + "best_issue_reason": "issue:43761 is the best overall anchor among the cluster candidates because it has the clearest concrete reproduction of the shared CLIP output regression and is the most mergeable as a single bug report.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43352", - "right": "issue:43976", + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Different failures: a bfloat16 precision warning vs a torch.compile/SDPA model error." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Unrelated symptoms: precision warning vs a tie_word_embeddings warning during LoRA fine-tuning." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Both involve saving, but one is a missing weight during load/encode and the other is max shard size behavior in save_pretrained()." + }, + { + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Different problems: a FlashAttention support issue for Nemotron vs a Python-version compatibility/import issue." + "reason": "Both are loading/import related, but they point to different symbols and different underlying code paths." }, { "left": "issue:43352", - "right": "issue:44380", + "right": "issue:43976", "accept": false, - "reason": "Both mention attention backends, but one is model support and the other is GPT2 attention scaling being ignored; not the same bug." + "reason": "Different problems: Flash Attention support for a model vs a Python version compatibility break." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "Unrelated: removed config API in v5 vs a noisy logging message." + "reason": "API removal/regression vs a generic noisy logging message; not the same bug." }, { - "left": "issue:44162", - "right": "issue:46172", + "left": "issue:43352", + "right": "issue:44380", "accept": false, - "reason": "Different model/code paths: ESM2 training breakage vs a DeepSeek v4 tensor operation bug." + "reason": "Different attention-related issues: model support gap vs GPT2 attention scaling being ignored." }, { - "left": "issue:45584", - "right": "issue:46365", + "left": "issue:44162", + "right": "issue:46172", "accept": false, - "reason": "Whisper generation bug vs an import-path issue for Trainer; no shared underlying failure." + "reason": "Different model families and failure modes; no shared code-path is evident." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "Both involve attention, but one is a config/backend regression and the other is an API registration behavior change." + "reason": "Config handling in attention backends vs AttentionInterface.register changing callback behavior." }, { "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "PatchTSMixer post_init validation and Trainer import failure are unrelated." + "reason": "PatchTSMixer post_init validation is unrelated to Trainer import conflicts with datasets." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper generation after align_special_tokens is unrelated to Trainer import resolution." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both touch MoE/FP8, but one is save_pretrained and the other is from_pretrained loading; different bugs." + "reason": "Both mention MoE/save/load, but one is FP8 save_pretrained and the other is a from_pretrained load regression." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "AttributeError in Gemma3Config vs an Apple Silicon allocator TypeError; unrelated." + "reason": "Gemma config attribute error and Apple Silicon allocator TypeError are unrelated." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "GPT2 attention scaling/backend regression is unrelated to the single-label classification zero-loss bug." + "reason": "Attention backend config bug vs degenerate zero-loss classification behavior." }, { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "TensorFlow backend packaging and a Trainer import conflict are different issues." + "reason": "Package-install backend selection issue vs importing Trainer being blocked by datasets." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "Both are warning-related, but one is local-path import warnings and the other is a bfloat16 precision warning." + "reason": "Local-path trailing slash warning vs bfloat16/float32 precision warning." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state corruption vs attention scaling being ignored are not the same failure." + "reason": "Optimizer state mismatch under FSDP2/PEFT is not the same as attention scaling being ignored." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but for different missing symbols in different packages." + "reason": "Both are ImportErrors, but for different missing symbols and different dependency surfaces." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "Precision warning during compilation vs compiled flex_attention failure on newer torch; different code paths." + "reason": "A runtime precision warning is not the same as a torch >= 2.9 compiled flex_attention failure." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs and BeitImageProcessor label reduction are unrelated." + "reason": "xLSTM training blockers and BeitImageProcessorFast label reduction are unrelated." }, { "left": "issue:45600", "right": "issue:46298", "accept": false, - "reason": "Broken auto_mappings CI vs noisy logging; not the same underlying problem." + "reason": "auto_mappings CI breakage is unrelated to a generic logging complaint." }, { - "left": "issue:43618", - "right": "issue:44380", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "CLIP output-field regression vs GPT2 attention scaling/config issue; same broad area, different bug." + "reason": "Chat completions API rejection and batched processor padding crash are different request paths." }, { - "left": "issue:44483", - "right": "issue:44514", + "left": "issue:43618", + "right": "issue:44380", "accept": false, - "reason": "Chat-completions request rejection and a batched processor crash are different failures." + "reason": "Both involve model internals, but CLIP output fields and GPT2 attention scaling are distinct bugs." }, { "left": "issue:43618", "right": "issue:43761", "accept": true, - "reason": "Both describe the same CLIP v5 regression where optional outputs are no longer populated; one manifests as missing attentions, the other as missing hidden states." + "reason": "Same underlying CLIP v5 regression: optional outputs are no longer propagated in forward results, affecting both attentions and hidden_states." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "One is a loading bug for Qwen3-VL-Moe, the other is fp8 support/inference work; not the same change." + "reason": "A Qwen3-VL-Moe regression and an fp8 inference support request are not the same issue." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "Layer-type import regression and local_files_only network requests are unrelated." + "reason": "Different import/dependency failures: a missing internal symbol vs unexpected API requests under local_files_only." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Both mention RoPE, but they concern different models and different implementation problems." + "reason": "RoPE layer-dropping in Smollm3 and GLM-5 RoPE implementation discussion are not the same concrete bug." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Two separate model-loading regressions with different symptoms and causes." + "reason": "Qwen3VL-Moe loading regression and a generic sorted-state_dict load error are not clearly the same failure." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving and pipeline save_pretrained are different save-path bugs." + "reason": "Saving quantized models and saving Pipeline objects fail in different save_pretrained paths." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Qwen3.5 label propagation and Trainer import failure are unrelated." - }, - { - "left": "issue:36683", - "right": "issue:43475", - "accept": false, - "reason": "Different model families and different missing attributes." - }, - { - "left": "issue:45276", - "right": "issue:46077", - "accept": false, - "reason": "Both are Gemma4 issues, but one is embedding resize propagation and the other is attention-mask construction." - }, - { - "left": "issue:36010", - "right": "issue:42757", - "accept": false, - "reason": "Both are import errors, but they are for different symbols from different libraries." - }, - { - "left": "issue:43352", - "right": "issue:43937", - "accept": false, - "reason": "FlashAttention support for Nemotron is unrelated to GLM-5 generation config validation." - }, - { - "left": "issue:43611", - "right": "issue:44655", - "accept": false, - "reason": "Base-model-prefix load regression and Pipeline.save_pretrained failure are different issues." - }, - { - "left": "issue:44483", - "right": "issue:46369", - "accept": false, - "reason": "Chat API request parsing and SAM3 LiteText memory overflow are unrelated." + "reason": "Qwen3.5 num_labels propagation and Trainer import failures are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The CLIP pair (issue:43618 / issue:43761) is a plausible duplicate match based on the titles and shared regression theme, and the rest of the soft edges are reasonably rejected as unrelated. No overstatement beyond the evidence shown." + "feedback": "Grounded overall: the packet supports issue:43761 as the best canonical CLIP regression anchor, and the single accepted soft pair (issue:43618 vs issue:43761) is a conservative duplicate-like call. The remaining soft-edge rejections are appropriately cautious and do not overclaim shared bugs. No PR pairs are present." }, "evaluator_used": true, "retried": false, @@ -16306,54 +16297,60 @@ "nodes": [ "issue:28282", "issue:33453", + "issue:36010", "issue:36296", "issue:36683", "issue:41669", + "issue:42757", "issue:42890", - "issue:42898", "issue:43010", "issue:43064", - "issue:43502", + "issue:43352", + "issue:43475", "issue:43525", "issue:43575", + "issue:43611", "issue:43618", "issue:43867", "issue:43901", + "issue:43937", "issue:43976", "issue:43992", - "issue:44062", "issue:44075", "issue:44077", - "issue:44112", "issue:44246", "issue:44355", - "issue:44360", + "issue:44483", "issue:44492", "issue:44625", "issue:44655", "issue:44855", "issue:44857", "issue:44938", - "issue:44964", "issue:45071", "issue:45216", + "issue:45276", "issue:45278", "issue:45310", "issue:45357", "issue:45375", - "issue:45440", "issue:45478", "issue:45561", - "issue:45600", "issue:45850", "issue:46077", "issue:46082", "issue:46202", "issue:46285", "issue:46365", - "issue:46664" + "issue:46369" ], "soft_pairs": [ + "issue:45276|issue:46077", + "issue:36683|issue:43475", + "issue:36010|issue:42757", + "issue:43352|issue:43937", + "issue:44483|issue:46369", + "issue:43611|issue:44655", "issue:43901|issue:44655", "issue:36683|issue:44077", "issue:43867|issue:45216", @@ -16364,8 +16361,8 @@ "issue:36683|issue:43525", "issue:36683|issue:44625", "issue:43992|issue:45310", - "issue:43064|issue:44492", "issue:43010|issue:45278", + "issue:43064|issue:44492", "issue:44075|issue:46365", "issue:44355|issue:44855", "issue:44938|issue:45071", @@ -16382,262 +16379,235 @@ "issue:43992|issue:45478", "issue:28282|issue:33453", "issue:43618|issue:46285", - "issue:43618|issue:44857", - "issue:44360|issue:45440", - "issue:42890|issue:44964", - "issue:42890|issue:44112", - "issue:42890|issue:45600", - "issue:42898|issue:46664", - "issue:43502|issue:44062" + "issue:43618|issue:44857" ], - "prepared_review_unit_hash": "be6facda0a5f54a73e6e21ffd0b22e7d46065604b7b324a791e7364219a0dae3", + "prepared_review_unit_hash": "71dfd700f43ef09f444a2b97ca67b85df7c73f204f8db7ecd15ca2aeb42c7c45", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 37752, - "estimated_input_tokens": 9438, - "estimated_eval_tokens": 19132 + "serialized_chars": 37786, + "estimated_input_tokens": 9447, + "estimated_eval_tokens": 19150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most soft pairs are thematically related but not the same underlying bug. The only strong duplicate-like match is the import-time slowdown pair (41669/44246); the rest span distinct models, configs, Python-version issues, and training/loading regressions.", - "confidence": 0.68, - "canonical_issue_reason": "issue:41669 is the most concrete diagnosis of the import slowdown, identifying the likely cause ('import *' usage in models) and giving a specific actionable fix.", + "summary": "The cluster is heterogeneous: it mixes unrelated import/load/save/config regressions across different models and subsystems. None of the soft pairs look like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set, covering import regressions after upgrading; if one issue must represent this mixed cluster, it is the closest fit, though the cluster is not a clean duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the best representative issue overall: it is specific, diagnostic, and likely serves as the most useful duplicate target for the broader 'import transformers takes long' complaint.", + "best_issue_reason": "issue:45278 is the best single representative because it is the most general import-regression report and overlaps with several other import-time complaints more than the model-specific issues do.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Same model family, but different bugs: token embedding/output embedding resize vs causal attention mask construction." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different models and different missing attributes; not the same failure path." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different packages and likely different fixes." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support rejection and invalid GenerationConfig are distinct validation/code-path issues." + }, + { + "left": "issue:44483", + "right": "issue:46369", + "accept": false, + "reason": "One is a chat/completions API rejection; the other is an import conflict around Trainer and datasets." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Loading models with base_model_prefix and saving Pipeline objects are separate code paths." + }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs mentioning return_all_scores and saving Pipeline objects are different problems." + "reason": "Docs mentioning return_all_scores do not describe the same bug as Pipeline save_pretrained failure." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Different model/config bugs: missing vocab_size vs patchtsmixer post_init handling." + "reason": "Missing Gemma3 config attribute and ignored SGD optimizer args are unrelated." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Both involve loading/saving, but one is state_dict ordering failure and the other is incorrect saved checkpoint contents for Qwen3.5." + "reason": "Sorted state_dict load failure and incorrect save_pretrained output are different regressions." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are Python-version incompatibilities, but they affect different versions and report different failures." + "reason": "Both are version/platform compatibility reports, but the actual breakage and environment constraints differ." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Num_labels propagation and bidirectional-attention mask construction are unrelated bugs." + "reason": "Config propagation for num_labels is unrelated to attention mask generation." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Different Llama/Gemma configuration behaviors: causal mask generation vs head_dim validation." + "reason": "Bidirectional attention mask logic and head_dim validation are separate issues." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIPOutput attentions assignment and Qwen3.5 save_pretrained regression are unrelated." + "reason": "CLIP attentions being unset is unrelated to Qwen3.5 checkpoint saving." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Different missing-attribute regressions on different configs/models." + "reason": "Different config attributes missing on different model families; not a shared root cause." }, { "left": "issue:36683", "right": "issue:44625", "accept": false, - "reason": "Missing vocab_size on Gemma3Config is not the same as num_labels propagation in Qwen3.5." + "reason": "Both are config-related, but one is a missing Gemma3 attribute and the other is num_labels propagation." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Different model-loading bugs affecting different architectures and failure modes." + "reason": "Missing embed_tokens.weight and Qwen3.5 from_pretrained errors are not clearly the same bug." }, { - "left": "issue:43064", - "right": "issue:44492", + "left": "issue:43010", + "right": "issue:45278", "accept": false, - "reason": "Optimizer-state corruption and a cache-strategy typo are unrelated." + "reason": "Cache update no_grad behavior and broad import failures after upgrade are unrelated." }, { - "left": "issue:43010", - "right": "issue:45278", + "left": "issue:43064", + "right": "issue:44492", "accept": false, - "reason": "No-grad decoration for cache updates is unrelated to broad import errors after upgrade." + "reason": "Distributed optimizer-state corruption and a typo in cache strategies do not share the same code path." }, { "left": "issue:44075", "right": "issue:46365", "accept": false, - "reason": "SGD arguments being ignored is unrelated to Trainer import failing unless datasets is uninstalled." + "reason": "Ignored SGD args and Trainer import failure are different subsystems and fixes." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled-Python-file errors and a torch.jit.script parsing/IndentationError are different import-time problems." + "reason": "Compiled Python import errors and a torch.jit parser/IndentationError are not the same underlying defect." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are distinct issues." + "reason": "Python version incompatibility and PretrainedConfig type-checking regression are distinct issues." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "State_dict sorting load failure is unrelated to a missing config field being dropped by strict handling." + "reason": "Load failure from state_dict ordering and silent dropping of a config field are different problems." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "Loading failure from sorted state_dict and Gemma pipeline roundtrip on \u2581 are not the same bug." + "reason": "State_dict load errors and Gemma pipeline roundtrip failure are not the same bug." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "A config attribute error and a Trainer FSDP/PEFT optimizer-state bug are different code paths." + "reason": "Gemma3 config attribute loss and FSDP2/PEFT optimizer-state corruption are unrelated." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Seed-related flaky tests and an xdist race on captured_info.txt are not the same failure." + "reason": "Test flakiness from missing set_seed and xdist race on captured_info.txt have different causes." }, { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Both describe the same import-time slowdown; 41669 identifies the likely root cause ('import *' in models) behind the generic 'import transformers takes long' complaint." + "accept": false, + "reason": "Import-* slowdown and intermittent import slowness are performance complaints, but not the same concrete regression." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Cache update no-grad issue is not the same as a broad complaint that v5 breaks too many models." + "reason": "Cache update decoration and a broad 'breaks too many models' complaint are too different and too vague to merge." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug and TP load OOM on Qwen2-57B are related in topic only, not the same concrete defect." + "reason": "Tensor-parallel training bug and tp-related OOM are different failure modes." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "State_dict sorting load error and incorrect visual-encoder keys on save_pretrained are different save/load bugs." - }, - { - "left": "issue:43010", - "right": "issue:43976", - "accept": false, - "reason": "Cache update decoration and Python-version import compatibility are unrelated." - }, - { - "left": "issue:43618", - "right": "issue:44077", - "accept": false, - "reason": "CLIP attentions regression and patchtsmixer post_init restrictions are unrelated." - }, - { - "left": "issue:43992", - "right": "issue:45478", - "accept": false, - "reason": "UMT5 missing weight and Qwen3.5 MoE from_pretrained error are different model-loading failures." - }, - { - "left": "issue:28282", - "right": "issue:33453", - "accept": false, - "reason": "Missing PyTorch ImportError and tokenizer-loading regression are unrelated." + "reason": "Sorted state_dict loading and incorrect visual encoder keys on save are separate checkpoint bugs." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "CLIPOutput attentions and CLIP text-model structural inconsistency are different CLIP regressions." + "reason": "Both concern CLIP model structure, but one is attentions assignment and the other is .text_model structure inconsistency." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "CLIP attentions assignment and LwDetrImageLoss AMP crash are unrelated." - }, - { - "left": "issue:44360", - "right": "issue:45440", - "accept": false, - "reason": "DSA indexer activation choice and DeepseekV3MoE divergence are different subsystems and bugs." - }, - { - "left": "issue:42890", - "right": "issue:44964", - "accept": false, - "reason": "Seed flakiness in SAM HQ tests is unrelated to Qwen multimodal TP OOM." - }, - { - "left": "issue:42890", - "right": "issue:44112", - "accept": false, - "reason": "Lack of set_seed in tests and a stale device-override CI failure are not the same issue." - }, - { - "left": "issue:42890", - "right": "issue:45600", - "accept": false, - "reason": "Test seeding flakiness and auto_mappings referencing removed configs are unrelated." - }, - { - "left": "issue:42898", - "right": "issue:46664", - "accept": false, - "reason": "Tokenization cleanup behavior change and SmolVLM num_frames semantics are different regressions." + "reason": "CLIPOutput attentions and an AMP/CUDA loss crash are unrelated." }, { - "left": "issue:43502", - "right": "issue:44062", + "left": "issue:28282", + "right": "issue:33453", "accept": false, - "reason": "local_files_only network leakage and AddedToken keyword duplication are unrelated." + "reason": "PyTorch missing import and tokenizer loading regression are different failure classes." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is conservative and consistent with the packet. The only accepted soft pair, issue:41669 / issue:44246, is a plausible duplicate-like match from the titles alone, and the remaining verdicts are appropriately cautious rather than overstated." - }, - "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 }, @@ -16664,21 +16634,21 @@ "issue:38617", "issue:42548", "issue:42757", - "issue:43010", + "issue:42890", + "issue:42898", + "issue:43010", "issue:43065", "issue:43232", "issue:43299", - "issue:43475", "issue:43502", - "issue:43526", "issue:43576", "issue:43723", "issue:43824", "issue:43827", "issue:43874", - "issue:43901", + "issue:44062", + "issue:44112", "issue:44222", - "issue:44230", "issue:44315", "issue:44351", "issue:44355", @@ -16689,26 +16659,32 @@ "issue:44655", "issue:44734", "issue:44749", - "issue:44908", "issue:44961", + "issue:44964", "issue:45071", "issue:45125", "issue:45375", + "issue:45440", "issue:45468", "issue:45478", + "issue:45600", "issue:46132", - "issue:46143", - "issue:46144", "issue:46617", "issue:46664", "issue:46726" ], "soft_pairs": [ + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:42898|issue:46664", + "issue:43502|issue:44062", + "issue:44734|issue:46617", "issue:43723|issue:44556", "issue:44222|issue:45125", - "issue:44734|issue:46617", - "issue:43874|issue:46617", "issue:44387|issue:44556", + "issue:43874|issue:46617", "issue:43824|issue:46617", "issue:43010|issue:43065", "issue:43502|issue:44485", @@ -16738,288 +16714,282 @@ "issue:44961|issue:45468", "issue:44351|issue:46617", "issue:44485|issue:44961", - "issue:38617|issue:43576", - "issue:44908|issue:44961", - "issue:36296|issue:43901", - "issue:46143|issue:46144", - "issue:44222|issue:44230", - "issue:36296|issue:43526", - "issue:43475|issue:46617" + "issue:38617|issue:43576" ], - "prepared_review_unit_hash": "22b05ad71cf180232d27df72f94162799553673ac65b951333883cc480df10b6", + "prepared_review_unit_hash": "6ca9abc0b6ebab76143d67070eed7600aa8976a194f602bfdf53feb1683085f0", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39455, - "estimated_input_tokens": 9864, - "estimated_eval_tokens": 19984 + "serialized_chars": 39507, + "estimated_input_tokens": 9877, + "estimated_eval_tokens": 20010 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it contains several v5 regression/loading reports and many unrelated subsystem issues. I would center the cluster on the broad checkpoint reload regression, but none of the soft-edge pairs are strong enough to merge as duplicates.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44556 is the most broadly representative regression report here: upgrading from v4.57 to v5.2/v5.3 breaks checkpoint reload across models, which fits the general v5 breakage theme better than the narrower or more specialized issues.", + "summary": "The set is highly heterogeneous: most items are unrelated model-loading, import, CI, docs, or performance bugs. None of the soft pairs look like safe duplicates under a strict reading.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43576 is the broadest, most representative v5 regression in the set and a reasonable cluster anchor; it is user-facing and version-wide rather than tied to one model or test case.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44556 is the strongest single issue in the set because it describes a clear, user-facing upgrade regression with broad impact and good potential to anchor related reports.", + "best_issue_reason": "issue:43576 best fits as the cluster representative because it captures the generic v5 breakage theme seen across several nearby issues without being overly model-specific.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43723", - "right": "issue:44556", + "left": "issue:44360", + "right": "issue:45440", "accept": false, - "reason": "Both involve loading regressions, but one is tokenizer loading and the other is checkpoint/model reload after version upgrade; different code paths." + "reason": "Different bugs: a DSA indexer/ReLU discussion versus a DeepSeekV3 implementation divergence." }, { - "left": "issue:44222", - "right": "issue:45125", + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated symptoms and subsystems: SamHQ test seeding failures versus Phi-4 multimodal model loading." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are test/CI issues, but they target different models and failure causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "SamHQ integration flakiness is unrelated to removed Sam3LiteText configs in auto_mappings." + }, + { + "left": "issue:42898", + "right": "issue:46664", "accept": false, - "reason": "Both mention Qwen3.5/MoE/fp8-related problems, but one is save_pretrained for FP8 and the other is tensor-parallel plan support; not the same bug." + "reason": "One is a tokenizer cleanup behavior change; the other is a video processor num_frames semantic issue." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Different code paths: local_files_only network access versus AddedToken keyword duplication." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "KV-cache continuation crash in serving is unrelated to FileNotFoundError when loading trust_remote_code models from cache." + "reason": "Different loading/runtime failures: KV-cache continuation indexing in serve versus cache loading FileNotFoundError." }, { - "left": "issue:43874", - "right": "issue:46617", + "left": "issue:43723", + "right": "issue:44556", "accept": false, - "reason": "Image-patch counting in a GLM46V image processor is a different failure than cache-based trust_remote_code loading." + "reason": "Both involve loading, but one is AutoTokenizer loading in v5 and the other is checkpoint reload compatibility after upgrade." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained behavior and missing _tp_plan for tensor parallelism are separate issues." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "One is int4 quantization memory/OOM behavior; the other is checkpoint reload failure after upgrading versions." + "reason": "Different regressions: int4 memory growth/OOM versus checkpoint reload breakage after version upgrade." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Missing get_number_of_image_patches is unrelated to trust_remote_code cache loading." }, { "left": "issue:43824", "right": "issue:46617", "accept": false, - "reason": "ImportError for a specific Qwen2.5-VL class is not the same as failing to load a trust_remote_code model from cache." + "reason": "ImportError for a specific Qwen VL class is not the same bug as cache FileNotFoundError." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Decorating cache/layer update methods with no_grad is unrelated to the dummy Conv2d in Sam3PixelDecoder." + "reason": "No shared underlying change: no_grad decoration for cache updates versus a dummy Conv2d in Sam3PixelDecoder." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "local_files_only causing network access is unrelated to GLM-5 RoPE implementation details." + "reason": "local_files_only network requests and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "Broken transformers env command and AttentionInterface.register behavior change are different features with different fixes." + "reason": "A broken CLI env command is not the same as AttentionInterface.register changing function behavior." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Same model family, but tensor-parallel plan missing is not clearly the same as a from_pretrained loading error." + "reason": "Closely themed, but the titles point to different problems: missing _tp_plan versus a broader from_pretrained error." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow reporting on failed training jobs and tokenizer-loading regression are unrelated." + "reason": "MLFlow reporting on failure and tokenizer loading regression are different issues." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and slower data filtering after upgrade are different symptoms in different code paths." + "reason": "Compiled Python file errors and a Chinese-language performance regression report are unrelated." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different symbols and likely different missing exports." + "reason": "Both are import errors, but they concern different missing symbols and likely different fixes." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "One is tensor-parallel support for Qwen3_5Moe; the other is a missing config field silently dropped by strict parsing." + "reason": "Missing _tp_plan for tensor parallelism is unrelated to a missing vision-config field." }, { "left": "issue:46664", "right": "issue:46726", "accept": false, - "reason": "Both are vision/video processor issues, but one is num_frames semantics and the other is backend detection/processor recognition." + "reason": "Video num_frames semantics and missing vision-backend handling are distinct processor issues." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both mention update-style generation code, but cache update no_grad decoration and sync_gpus kwargs handling are different bugs." + "reason": "Cache no_grad decoration and generation kwargs updates after sync_gpus are different code paths." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading getting stuck and documentation confusion in model.forward are unrelated." + "reason": "Sentence-transformers loading hang and model.forward documentation confusion are not the same bug." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Network requests despite local_files_only has nothing to do with the placeholder 'racoon' issue." + "reason": "local_files_only networking and an unrelated racoon placeholder issue are not duplicates." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Local-files-only network leakage and FP8 save_pretrained behavior are unrelated failures." + "reason": "No overlap beyond being in the same library; one is a download flag bug, the other is FP8 save_pretrained." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() after removal is unrelated to the placeholder issue." + "reason": "Docs referencing pipeline() after removals is unrelated to the racoon issue." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLflow reporting failure and tensor-parallel training bug are different concerns." + "reason": "MLFlow status reporting and tensor parallel training are different problems." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are version-related loading regressions, but they affect different model families and are not clearly the same underlying bug." + "reason": "Both are version-related loading regressions, but they affect different model families and failure modes." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file errors and PretrainedConfig type-checking regressions are unrelated." + "reason": "Compiled-file errors do not match a PretrainedConfig type-checking regression." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer-loading regression and Llama 3.2 vision model-loading breakage are different issues." + "reason": "Both are loading regressions, but they concern different models and likely different root causes." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python file errors and saving Pipeline objects are different code paths." + "reason": "Compiled Python file failures and Pipeline save_pretrained problems are unrelated." }, { "left": "issue:27957", "right": "issue:33453", "accept": false, - "reason": "XLMRoberta flash-attention support and tokenizer loading regression are unrelated." + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to tokenizer-loading regression." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing is_offline_mode import from huggingface_hub is unrelated to the transformers env command." + "reason": "Hugging Face Hub offline-mode import failure and the transformers env command are different issues." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env command and PretrainedConfig type-checking regression are different failures." + "reason": "CLI env command breakage and PretrainedConfig type checking are unrelated." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger kernel not applied with model_init and DSA indexer missing ReLU are unrelated." + "reason": "Liger Kernel application via model_init and a DSA indexer/ReLU discussion are different code paths." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion around forward arguments and a missing layer_type_validation import are unrelated." + "reason": "Documentation wording and a missing internal import are not the same bug." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env command and Qwen2.5-VL import failure are unrelated." + "reason": "Broken env command and Qwen2_5_VL import failure do not share an underlying fix." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Placeholder issue and Gemma-4 audio positional encoding bug are unrelated." + "reason": "The placeholder racoon issue is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "Cannot import HybridCache and FileNotFoundError loading from cache are different failures." + "reason": "HybridCache import failure and trust_remote_code cache loading failure are different problems." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to the placeholder issue." + "reason": "GLM-5 RoPE implementation and the racoon placeholder issue are unrelated." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Missing layer_type_validation import and broken env command are unrelated." - }, - { - "left": "issue:44908", - "right": "issue:44961", - "accept": false, - "reason": "inverse_sqrt scheduler kwargs handling is unrelated to the placeholder issue." - }, - { - "left": "issue:36296", - "right": "issue:43901", - "accept": false, - "reason": "Tensor parallel training bug and text-classification docs mismatch are unrelated." - }, - { - "left": "issue:46143", - "right": "issue:46144", - "accept": false, - "reason": "Missing kwargs propagation in RoFormer methods and lack of attention-interface usage are related only at a broad subsystem level, not the same bug." - }, - { - "left": "issue:44222", - "right": "issue:44230", - "accept": false, - "reason": "Both are FP8/MoE-related, but save_pretrained support and inference support are different changes." - }, - { - "left": "issue:36296", - "right": "issue:43526", - "accept": false, - "reason": "Tensor-parallel training bug and BeitImageProcessorFast reduce_labels behavior are unrelated." - }, - { - "left": "issue:43475", - "right": "issue:46617", - "accept": false, - "reason": "SAM 3 Video missing attribute error and trust_remote_code cache loading FileNotFoundError are different bugs." + "reason": "Missing layer_type_validation import and the broken env command are different regressions." } ] }, @@ -17043,6 +17013,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ + "issue:36296", "issue:41628", "issue:41669", "issue:42371", @@ -17050,21 +17021,20 @@ "issue:42890", "issue:43352", "issue:43408", + "issue:43475", "issue:43525", + "issue:43526", "issue:43572", "issue:43576", "issue:43673", - "issue:43827", + "issue:43901", "issue:43976", "issue:44038", - "issue:44077", + "issue:44222", "issue:44230", "issue:44246", "issue:44273", - "issue:44295", - "issue:44297", "issue:44315", - "issue:44351", "issue:44360", "issue:44393", "issue:44485", @@ -17072,27 +17042,33 @@ "issue:44556", "issue:44623", "issue:44704", - "issue:44861", "issue:44908", + "issue:44961", "issue:45003", "issue:45071", "issue:45092", "issue:45125", "issue:45230", "issue:45278", - "issue:45341", "issue:45362", - "issue:45478", "issue:45542", "issue:45753", "issue:45850", "issue:46095", + "issue:46143", + "issue:46144", "issue:46209", "issue:46218", - "issue:46328", + "issue:46617", "issue:46664" ], "soft_pairs": [ + "issue:44908|issue:44961", + "issue:36296|issue:43901", + "issue:46143|issue:46144", + "issue:43475|issue:46617", + "issue:36296|issue:43526", + "issue:44222|issue:44230", "issue:41628|issue:42548", "issue:44315|issue:46209", "issue:44485|issue:46664", @@ -17122,160 +17098,186 @@ "issue:43576|issue:44908", "issue:43976|issue:44393", "issue:44704|issue:44908", - "issue:44908|issue:45071", - "issue:44295|issue:45230", - "issue:44297|issue:45230", - "issue:41669|issue:44351", - "issue:44492|issue:46328", - "issue:41669|issue:45341", - "issue:43976|issue:45850", - "issue:44393|issue:45478", - "issue:43976|issue:46218", - "issue:44077|issue:44861", - "issue:43827|issue:44908" + "issue:44908|issue:45071" ], - "prepared_review_unit_hash": "64fc9c70fab4274b8a4121e7f8423a06e6e65a9ee71875063007468fe1893e90", + "prepared_review_unit_hash": "5a6dc41540fb8a230e42a119abd36942ce4218229f5319de6febb1e1a8c4e48a", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 40, - "serialized_chars": 38992, - "estimated_input_tokens": 9748, - "estimated_eval_tokens": 19752 + "soft_pair_count": 36, + "serialized_chars": 37671, + "estimated_input_tokens": 9418, + "estimated_eval_tokens": 19092 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: a few items are import/lazy-loading/API-regression themed, but most are unrelated model-specific, docs, or config bugs. I would not collapse the soft pairs into duplicates except possibly the lazy-loading/import-slowdown pair, but I\u2019m keeping the triage strict here.", - "confidence": 0.62, - "canonical_issue_reason": "If a single issue must be chosen, issue 44273 is the closest root-cause-style candidate because it describes a core lazy-loading regression that could explain some import/import-time complaints. That said, the cluster is still too mixed to treat it as a clean duplicate set.", + "summary": "This cluster is mostly a loose mix of Transformers v5 import/lazy-loading regressions and unrelated model-specific bugs; none of the soft pairs look like clear duplicates.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45850 is the broadest umbrella complaint in the set about v5 breakage, so it best anchors this mixed cluster despite being vague.", "canonical_pr_reason": null, - "best_issue_reason": "44273 is the most reusable umbrella issue among the set, but it is still not a great canonical because many other items are model-specific breakages or unrelated bugs. It\u2019s the least-bad anchor, not a true duplicate center.", + "best_issue_reason": "Issue 45850 is the most central representative of the repeated 'v5 breaks things' theme and is the most suitable single issue to stand in for the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Completely unrelated: scheduler kwargs bug vs a placeholder issue with no matching code path." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug and a pipeline docs mismatch are unrelated." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "Both are RoFormer-related, but one is kwargs propagation and the other is attention-interface usage; different bugs." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "SAM3 vision encoder attribute error vs trust-remote-code cache FileNotFoundError; no shared failure path." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor parallel training and BEiT image-processor label reduction are separate subsystems." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both mention fp8/moe, but one is save_pretrained serialization and the other is inference support." + }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Both are import errors, but they are different missing symbols with no evidence of the same root cause." + "reason": "Both are import errors, but for different symbols and likely different export issues; too broad to merge." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Different subsystems and failure modes: Liger/model_init vs FP8/B200 inference." + "reason": "Model_init/liger-kernel application and B200 fp8 inference failure are unrelated problems." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "Unrelated models and bugs: RoPE implementation vs video frame-count semantics." + "reason": "RoPE implementation discussion vs video num_frames semantics; different model components." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "One is a specific chat crash, the other is a broad complaint about many models; not the same bug." + "reason": "A specific Qwen3.5 chat crash is not the same concrete bug as a broad 'v5 breaks too many models' complaint." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Cache-missing runtime bug vs a typo in cache strategy wording; not the same issue." + "reason": "Cache-missing during chunked prefill and a typo in cache strategies are different issues." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy-loading/import performance is unrelated to the TensorFlow/tensorboard backend error." + "reason": "Lazy loading regression and TensorFlow backend detection/installation behavior are related only at a high level." }, { "left": "issue:44360", "right": "issue:46664", "accept": false, - "reason": "Completely different topics: DSA indexer activation vs video processor frame handling." + "reason": "DSA indexer/ReLU discussion is unrelated to the SmolVLM video processor bug." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "One is a concrete processor.save_pretrained file bug; the other is too generic to match it." + "reason": "One is a concrete missing-files bug in processor.save_pretrained; the other is an empty generic report." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both mention Qwen3/Moe, but they are different problems: general v5 compatibility vs missing tensor-parallel plan." + "reason": "Same model family, but generic Qwen3-VL-Moe issues and a missing tensor-parallel plan are different code paths." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading/import slowdown is not the same as unsafe sys.modules access, even if they may interact." + "reason": "Both touch loading mechanics, but lazy loading and unsafe sys.modules access are distinct bugs." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Vision bounding-box output bug vs PretrainedConfig type-checking regression; different code paths." + "reason": "Qwen3-VL bounding-box output issues are unrelated to PretrainedConfig type-checking regressions." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Both are Qwen3/VL-related, but one is bbox output and the other is an infinite loop on a specific image." + "reason": "Same model family, but bbox hallucinations and an infinite loop on a specific image are not the same bug." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-star cleanup affecting import cost is not the same as a missing GenerationMixin cache during chunked prefill." + "reason": "Import-star performance regression vs generation cache missing during chunked prefill; different failure modes." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "CLI env command breakage and checkpoint reload compatibility are separate failures." + "reason": "v5 env-command breakage and old-checkpoint reload incompatibility are separate regressions." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "A missing config attribute is a specific model bug, not the same as a broad 'v5 breaks many models' complaint." + "reason": "Missing pad_token_id in Llama4Config is a specific config bug, not the same as a broad v5 complaint." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both concern old checkpoints under v5, but one is generic reload compatibility and the other is remote-code meta-init incompatibility." + "reason": "Both are upgrade/reload compatibility issues, but they target different checkpoint types and initialization paths." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "StableLm pad-token config regression is not the same as a generic breakage report." + "reason": "StableLmConfig pad_token_idx regression is a specific missing field, not the same issue as the broad v5 breakage report." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Many import errors after upgrade is broader than the generic 'many models break' report and not the same defect." + "reason": "Import errors after upgrade are related in theme, but this is not the same concrete bug as the generic v5 breakage complaint." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A cache-strategy typo is unrelated to model breakage after v5." + "reason": "A cache-strategy typo is not the same as a general 'v5 breaks models' report." }, { "left": "issue:42371", @@ -17287,121 +17289,43 @@ "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import performance cleanup is not the same as a specific module import failure for Qwen3_5ForCausalLM." + "reason": "Import-star slowdown and a specific module import failure do not look like the same root cause." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "sam3_video/sam3_tracker warning is unrelated to Liger application in model_init." + "reason": "SAM3 model-type warning and Liger Kernel model_init behavior are unrelated." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 inference support for Qwen models and DeepSeek weight conversion are different requests." + "reason": "FP8 inference support for Qwen models and a DeepSeek weight conversion script are different requests." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test seeding flakiness is unrelated to scheduler kwargs propagation." + "reason": "Test flakiness from missing set_seed and an inverse_sqrt scheduler kwargs bug are unrelated." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "These both point at import/lazy-loading pain, but one is vague performance and the other is a specific lazy-loading bug; too little evidence to merge." + "reason": "Import slowness and lazy-loading failure may be adjacent, but the evidence is too weak to call them duplicates." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention support and Qwen3-VL bbox hallucination are unrelated." + "reason": "Nemotron Flash Attention support and Qwen3-VL bounding-box errors are unrelated." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "CLI env command breakage and scheduler kwargs bug are unrelated." - }, - { - "left": "issue:43976", - "right": "issue:44393", - "accept": false, - "reason": "Python-version compatibility and Qwen3-VL bbox output are unrelated." - }, - { - "left": "issue:44704", - "right": "issue:44908", - "accept": false, - "reason": "AutoProcessor kwargs forwarding and inverse_sqrt scheduler kwargs are different APIs and failures." - }, - { - "left": "issue:44908", - "right": "issue:45071", - "accept": false, - "reason": "Scheduler kwargs bug and config type-checking regression are unrelated." - }, - { - "left": "issue:44295", - "right": "issue:45230", - "accept": false, - "reason": "Position_ids buffer handling is not the same as an unspecified generic bug report." - }, - { - "left": "issue:44297", - "right": "issue:45230", - "accept": false, - "reason": "Tokenizer save_pretrained mismatch is a concrete bug; the other issue is too generic to align." - }, - { - "left": "issue:41669", - "right": "issue:44351", - "accept": false, - "reason": "Both are import-related, but one is import-time performance and the other is a missing top-level export." - }, - { - "left": "issue:44492", - "right": "issue:46328", - "accept": false, - "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance are unrelated." - }, - { - "left": "issue:41669", - "right": "issue:45341", - "accept": false, - "reason": "Import-star cleanup and a testing_utils bug do not describe the same problem." - }, - { - "left": "issue:43976", - "right": "issue:45850", - "accept": false, - "reason": "Python compatibility regression is not the same as a generic statement that v5 breaks many models." - }, - { - "left": "issue:44393", - "right": "issue:45478", - "accept": false, - "reason": "Different Qwen model families and different symptoms: bbox output vs from_pretrained error." - }, - { - "left": "issue:43976", - "right": "issue:46218", - "accept": false, - "reason": "Python version compatibility is unrelated to a missing Qwen3_5 module import." - }, - { - "left": "issue:44077", - "right": "issue:44861", - "accept": false, - "reason": "Optional post_init policy and tied-weights-key AttributeError are unrelated." - }, - { - "left": "issue:43827", - "right": "issue:44908", - "accept": false, - "reason": "Docs still mentioning pipeline() and scheduler kwargs propagation are unrelated." + "reason": "Broken env command and inverse_sqrt scheduler kwargs are different features with no shared code path." } ] }, @@ -17426,12 +17350,10 @@ "cluster_id": "cluster-10000-479", "nodes": [ "issue:28348", - "issue:30333", "issue:31515", "issue:31795", "issue:33453", "issue:34689", - "issue:38617", "issue:41669", "issue:42371", "issue:42913", @@ -17440,41 +17362,53 @@ "issue:43352", "issue:43502", "issue:43519", - "issue:43525", - "issue:43572", "issue:43576", "issue:43606", - "issue:43618", "issue:43704", - "issue:44230", + "issue:43827", + "issue:43976", + "issue:44077", "issue:44261", "issue:44263", "issue:44273", + "issue:44295", + "issue:44297", "issue:44315", "issue:44351", "issue:44360", + "issue:44393", "issue:44485", "issue:44492", - "issue:44623", "issue:44655", "issue:44829", "issue:44861", "issue:44908", - "issue:44938", "issue:44998", "issue:45230", "issue:45250", - "issue:45290", "issue:45310", + "issue:45341", "issue:45468", + "issue:45478", "issue:45850", "issue:46018", "issue:46095", "issue:46172", "issue:46218", + "issue:46328", "issue:46655" ], "soft_pairs": [ + "issue:44295|issue:45230", + "issue:44297|issue:45230", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:43976|issue:46218", + "issue:44077|issue:44861", + "issue:43827|issue:44908", "issue:31795|issue:33453", "issue:43576|issue:44273", "issue:44315|issue:44829", @@ -17488,8 +17422,8 @@ "issue:43519|issue:44908", "issue:41669|issue:43704", "issue:31795|issue:34689", - "issue:46018|issue:46172", "issue:31515|issue:31795", + "issue:46018|issue:46172", "issue:43704|issue:44485", "issue:44485|issue:45468", "issue:43502|issue:43519", @@ -17505,292 +17439,282 @@ "issue:42913|issue:43066", "issue:28348|issue:45250", "issue:44908|issue:44998", - "issue:44492|issue:44998", - "issue:43618|issue:44998", - "issue:44998|issue:45850", - "issue:44938|issue:44998", - "issue:44861|issue:45290", - "issue:31515|issue:38617", - "issue:43525|issue:44861", - "issue:43572|issue:44861", - "issue:28348|issue:43352", - "issue:44230|issue:44623", - "issue:30333|issue:31515" + "issue:44492|issue:44998" ], - "prepared_review_unit_hash": "34452a9d2a7220c0f24e56d1966891719b8c9119a3dfeb2ee7611a3407d161a2", + "prepared_review_unit_hash": "e7f15fb5f187def6f11fec4b59f96b7f7725ef12c589ed42475fb1e2bee0caaf", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39116, - "estimated_input_tokens": 9779, - "estimated_eval_tokens": 19814 + "serialized_chars": 39145, + "estimated_input_tokens": 9787, + "estimated_eval_tokens": 19830 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items span unrelated tokenizer, loading, processor, scheduler, import, and model-support issues. None of the soft pairs look like the same underlying bug/change, so no duplicate merge candidate stands out.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue; the set is not a coherent duplicate cluster.", + "summary": "This cluster is mostly a loose set of unrelated Transformers bug reports. None of the soft pairs appear to describe the same concrete bug/change, so I would reject all candidate duplicate edges. The strongest standalone issue is the specific Qwen3.5 MoE `from_pretrained` regression.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 45310 is the clearest concrete regression report in the set: a specific `from_pretrained` failure for Qwen3.5 MoE on Transformers >=5.4.0, with a narrowly defined code path and likely reproducible behavior.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue cleanly represents the cluster because the items cover many unrelated problems and feature requests.", + "best_issue_reason": "Issue 45310 is the most triage-ready issue here because it is specific, actionable, and version-scoped; the other candidates are either too generic, documentation-only, or too loosely related.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "`44295` is a specific `position_ids` buffer bug; `45230` is a placeholder-level generic report, with no evidence they share the same root cause." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save metadata mismatch is a different failure mode from the generic `45230` report." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "One is about removing `import *` for startup/perf; the other is an import error for `HybridCache`. Related area, but not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "A cache-strategy typo and a `SequenceFeatureExtractor.pad` performance issue are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Model import cleanup versus a testing utility bug; no shared underlying defect." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Both mention broad v5 breakage, but they are different reports with no concrete shared code path." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different models and different failures: Qwen3-VL bounding-box output versus Qwen3.5 MoE loading." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python version compatibility and a missing Qwen3_5 module are different problems." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "`post_init` validation and `_get_tied_weight_keys` crashing on a list are separate code paths." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing `pipeline()` and `inverse_sqrt` scheduler kwargs being ignored are unrelated." + }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Different symptoms and scopes: documentation confusion in forward args vs a tokenizer-loading regression." + "reason": "Documentation confusion around forward args versus a tokenizer loading regression are not the same bug." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Unrelated areas: broken env command vs lazy loading behavior." + "reason": "A broken `transformers env` command and lazy loading not working are different issues." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Different code paths: model_init/Liger Kernel application vs flash_attention_3 training degeneration." + "reason": "`model_init` not applying Liger Kernel and degenerate training with flash-attn-3 are different failures." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer ReLU issue and Gemma4Audio positional encoding bug are distinct model components." + "reason": "A DSA indexer math discussion and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Both are processor bugs, but they affect different models and different calculations (timestamps vs video padding)." + "reason": "Both are processor bugs, but one is Qwen3-VL timestamp calculation and the other is SmolVLM video padding; not the same root cause." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading failure and inverse_sqrt scheduler kwargs handling are unrelated." + "reason": "Lazy loading versus scheduler kwargs are unrelated." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and _get_tied_weight_keys list handling are different serialization/weight-key bugs." + "reason": "Saving pipeline objects and tied-weight-key handling are different subsystems and failures." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Broken env command vs module import failure for Qwen3_5ForCausalLM are separate issues." + "reason": "CLI env command breakage is not the same as a missing `Qwen3_5ForCausalLM` import." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "The first is too generic; the second is a specific Qwen3.5 MoE from_pretrained error." + "reason": "`45230` is too generic to establish the same underlying bug as the specific Qwen3.5 MoE `from_pretrained` regression." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Timestamp calculation in Qwen3VL processor is unrelated to GLM-5 RoPE implementation." + "reason": "Different models and different implementation areas: timestamp handling versus RoPE implementation." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Processor timestamp math and scheduler kwargs handling are different subsystems." + "reason": "Processor timestamp math and scheduler kwargs are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup/performance and a VRAM leak are different problems." + "reason": "Import cleanup/startup cost and VRAM leakage in dataloader threads are not the same bug." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Forward-doc ambiguity is not the same as a model-loading regression for Llama 3.2 Vision." + "reason": "Forward-doc confusion and a model-loading regression for Llama 3.2 are distinct issues." }, { - "left": "issue:46018", - "right": "issue:46172", + "left": "issue:31515", + "right": "issue:31795", "accept": false, - "reason": "DeepSeek-V4 gating logic and a transpose/dtype bug are unrelated." + "reason": "Slow `from_pretrained` checkpoint loading and confusing forward documentation are unrelated." }, { - "left": "issue:31515", - "right": "issue:31795", + "left": "issue:46018", + "right": "issue:46172", "accept": false, - "reason": "Slow from_pretrained checkpoint loading is not the same as confusing forward documentation." + "reason": "Shared-expert gating and a dtype/transpose bug in DeepSeek-V4 are different concrete problems." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leakage in Qwen3ForCausalLM and GLM-5 RoPE math are unrelated." + "reason": "VRAM leakage and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Both touch positional encoding, but they are different models and distinct implementation bugs." + "reason": "Different model families and different components; no shared bug." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network requests and processor timestamp calculation are unrelated." + "reason": "`local_files_only=True` network leakage and Qwen3-VL timestamp calculation are unrelated." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 behavior settings and a VRAM leak are different concerns." + "reason": "TF32 configuration behavior and VRAM leakage are different topics." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad 'v5 breaks models' complaint is not the same as a specific Qwen import failure." + "reason": "A broad complaint that v5 breaks many models is not the same as a specific missing module import." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp calculation and VRAM leakage are unrelated bugs." + "reason": "Processor timestamp math and dataloader-thread VRAM leakage are unrelated." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is unrelated to Gemma4Audio positional encoding." + "reason": "CPU offload device-mismatch in CI and Gemma-4 audio positional encoding are different bugs." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return handling in DSA indexing is unrelated to audio positional encoding." + "reason": "`torch.split()` return-value handling in a DSA indexer is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "RMS norm epsilon precision and audio positional encoding are different bugs." + "reason": "Layer-norm epsilon precision discussion and Gemma-4 audio positional encoding are not the same issue." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching support for audio models and a weight-conversion-script question are not the same change." + "reason": "Continuous batching audio-model support and DeepSeek-V4 weight conversion are different requests." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "HybridCache import failure and Qwen3_5ForCausalLM import failure are different missing symbols." + "reason": "Missing `HybridCache` import and missing `Qwen3_5ForCausalLM` import are different symbols with no shared bug evidence." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Nemotron Flash Attention 2 support and Qwen import errors are unrelated." + "reason": "Flash Attention 2.0 support for NemotronH is unrelated to a missing Qwen3_5 module import." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both concern tokenizer v5 behavior, but one is a broad v4-to-v5 difference and the other is a specific decoder-type bug." + "reason": "Both concern tokenizer changes in v5, but the reports are not specific enough to establish the same underlying bug." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "Specific GPT2LMHeadModel FA2 support request vs a vague Flash Attention 2.0 issue; too broad to treat as the same bug/change." + "reason": "A GPT2-specific Flash Attention 2.0 support request is not the same as a generic Flash Attention 2.0 issue." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs handling and an unrelated placeholder title do not match." + "reason": "Scheduler kwargs handling and an unrelated placeholder title are not duplicates." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, "reason": "Cache-strategy typo and an unrelated placeholder title are not duplicates." - }, - { - "left": "issue:43618", - "right": "issue:44998", - "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the placeholder title." - }, - { - "left": "issue:44998", - "right": "issue:45850", - "accept": false, - "reason": "A placeholder title is not the same as a broad v5 regression report." - }, - { - "left": "issue:44938", - "right": "issue:44998", - "accept": false, - "reason": "Python 3.14 load failure is unrelated to the placeholder title." - }, - { - "left": "issue:44861", - "right": "issue:45290", - "accept": false, - "reason": "_get_tied_weight_keys AttributeError and chat-template tool-call crashes are different code paths." - }, - { - "left": "issue:31515", - "right": "issue:38617", - "accept": false, - "reason": "Slow checkpoint loading and an import error for layer_type_validation are unrelated." - }, - { - "left": "issue:43525", - "right": "issue:44861", - "accept": false, - "reason": "Missing pad_token_id and tied-weight-key list handling are different attribute/serialization bugs." - }, - { - "left": "issue:43572", - "right": "issue:44861", - "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is unrelated to _get_tied_weight_keys crashing." - }, - { - "left": "issue:28348", - "right": "issue:43352", - "accept": false, - "reason": "Both mention Flash Attention 2.0, but they target different models and likely different implementation work." - }, - { - "left": "issue:44230", - "right": "issue:44623", - "accept": false, - "reason": "FP8 inference support and processor.save_pretrained file emission are unrelated." - }, - { - "left": "issue:30333", - "right": "issue:31515", - "accept": false, - "reason": "MLFlow reporting on failed jobs and slow checkpoint loading are unrelated." } ] }, @@ -17814,23 +17738,46 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-479", "nodes": [ + "issue:28348", + "issue:30333", "issue:31515", "issue:33453", "issue:36296", "issue:37934", + "issue:38617", "issue:42936", "issue:43240", "issue:43298", + "issue:43352", + "issue:43525", + "issue:43572", + "issue:43618", "issue:43668", "issue:43717", "issue:43813", "issue:44031", "issue:44230", + "issue:44623", + "issue:44861", + "issue:44938", + "issue:44998", + "issue:45290", "issue:45310", + "issue:45850", "issue:46018", "issue:46144" ], "soft_pairs": [ + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:31515|issue:38617", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:28348|issue:43352", + "issue:44230|issue:44623", + "issue:30333|issue:31515", "issue:31515|issue:36296", "issue:31515|issue:33453", "issue:44230|issue:45310", @@ -17848,130 +17795,190 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "354387378ae0052f874cd6f0d530db48bfff7ec941db628bf2d1f4f75c0eea04", + "prepared_review_unit_hash": "7f3d1a44c3df2bd9e96cc6bb3717d2a7adbf339299cb6b984c5ebeb1b9182764", "component_budget": { "node_count": 479, "item_count": 479, - "soft_pair_count": 1762, - "serialized_chars": 913668, - "estimated_input_tokens": 228417, - "estimated_eval_tokens": 457090 + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 }, "budget": { - "node_count": 15, - "item_count": 15, - "soft_pair_count": 16, - "serialized_chars": 13075, - "estimated_input_tokens": 3269, - "estimated_eval_tokens": 6794 + "node_count": 28, + "item_count": 28, + "soft_pair_count": 26, + "serialized_chars": 23352, + "estimated_input_tokens": 5838, + "estimated_eval_tokens": 11932 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly heterogeneous false-positive matches. A few are loosely loading/model-related, but none of the soft pairs appear to describe the same underlying bug or fixable change, so no soft edge should be merged.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45310 is the most concrete and actionable report: it names a specific regression, version range, and model family, making it the best representative of the small loading-related subtheme.", + "summary": "This cluster is a grab-bag of unrelated issues: model-specific feature requests, loading/version regressions, tokenizer warnings, and a few training/runtime bugs. None of the soft pairs look like the same underlying bug, so there is no real duplicate hub here.", + "confidence": 0.88, + "canonical_issue_reason": "No true canonical issue emerges because the set is heterogeneous. If forced to nominate a representative, issue 45850 is the broadest umbrella complaint, but it is too vague and not a credible duplicate anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the strongest overall issue candidate because it is specific, reproducible, and easier to anchor than the more generic or orthogonal reports.", + "best_issue_reason": "Issue 45850 is the most general/high-level problem statement in the set, but it still does not match the rest closely enough to serve as a strong canonical duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "Different bugs: CLIP attentions regression vs an unrelated issue titled \"Unemployment\"." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "Unrelated issues; one is nonsensical/irrelevant, the other is a broad model-breakage complaint." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 import/load failure is not the same problem as the unrelated \"Unemployment\" issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Different code paths: tied-weight key handling vs chat template tokenization with tool calls." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading is unrelated to the layer_type_validation ImportError." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a different bug from _get_tied_weight_keys crashing." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLm pad_token_idx regression is unrelated to tied-weight-key handling." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but they are different model-specific support requests and not the same concrete fix." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support for Qwen models is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow job-status reporting failure is unrelated to slow checkpoint loading." + }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowness and tensor-parallel training are different failure modes in different code paths." + "reason": "from_pretrained performance regression and tensor-parallel training bug are different problems." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Both involve loading, but one is slow from_pretrained checkpoint loading while the other is a tokenizer-loading regression." + "reason": "Checkpoint loading slowness is not the same as a tokenizer-loading regression." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Same model family appears, but one is fp8 inference support and the other is a from_pretrained error regression; not the same bug." + "reason": "Both are Qwen3.5-related, but one is fp8 support work and the other is a from_pretrained error in a different context." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching for audio models and fp8 Qwen support are unrelated features/issues." + "reason": "Audio continuous batching support is unrelated to Qwen fp8 inference support." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba/Mamba-2 init_weights behavior and RoFormer attention-interface usage are unrelated model implementation problems." + "reason": "Mamba init_weights behavior and RoFormer attention interface usage are separate model implementations." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared-expert gating are different model bugs." + "reason": "Mask2Former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "A config type-hint issue is unrelated to tokenizer regex warning regressions." + "reason": "A ModernBERT type-hint issue is not the same as the tokenizer regex warning regression." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Both touch loss-related code, but one is an unused ignore_value and the other is kwargs not forwarded to loss calculation." + "reason": "Mask2Former ignore_value handling and kwargs propagation to loss calculation are different training bugs." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs forwarding and RoFormer attention-interface usage affect different components and code paths." + "reason": "kwargs not reaching the loss function is unrelated to RoFormer attention implementation details." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2Former loss masking and tokenizer regex warnings are unrelated." + "reason": "Mask2Former ignore_value logic is unrelated to tokenizer regex warnings." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2Former ignore_value handling is not the same bug as RoFormer attention implementation." + "reason": "Mask2Former loss/ignore_value behavior is not the same as RoFormer attention interface usage." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Unused ignore_value and a typo in quantization-operation naming are unrelated." + "reason": "Mask2Former ignore_value handling is unrelated to the quantization-operation typo." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and ModernBERTConfig type-hint correctness are different concerns." + "reason": "kwargs-to-loss propagation is a runtime bug, not the ModernBERT norm_eps type-hint issue." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "A config annotation bug is unrelated to a typo in a quantization symbol path." + "reason": "ModernBERT type-hint mismatch and a typo in quantization attribute name are unrelated." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Llama4TextL2Norm/RMS norm semantics and ModernBERT norm_eps typing are different issues." + "reason": "A Llama4 norm implementation question is not the same as a ModernBERT type-hint bug." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2Former loss behavior and ModernBERT config typing are unrelated." + "reason": "Mask2Former ignore_value usage and ModernBERT norm_eps typing are unrelated." } ] }, @@ -18004,52 +18011,52 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "b536a998d6c9a688ad543a80241b9cbd960a6b6c6c1e53d468266611e01cfce2", + "prepared_review_unit_hash": "003e4e9864bc2a213041457192b75ef445739dceafb0a721637f2c6df10ff8ba", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2876, - "estimated_input_tokens": 719, - "estimated_eval_tokens": 1694 + "serialized_chars": 2878, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2876, - "estimated_input_tokens": 719, - "estimated_eval_tokens": 1694 + "serialized_chars": 2878, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues are about unrelated topics: token truncation behavior, learning-rate scheduler functionality, and Wav2Vec2 word time offsets. They do not form a duplicate cluster.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "All three items are unrelated feature requests in different areas, so this cluster should not be merged.", + "confidence": 0.99, + "canonical_issue_reason": "None of the issues are duplicates; if one had to be picked as a representative, 10519 is the clearest, self-contained product feature request, but it is not a true canonical duplicate target for the others.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "10519 is the most concrete and well-scoped issue, with a specific tokenizer behavior request; the other two are separate feature asks in unrelated domains.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", - "right": "issue:11307", + "right": "issue:11005", "accept": false, - "reason": "Different domains and requests: tokenizer truncation configuration vs speech model time offsets. No shared bug or fix path." + "reason": "Truncation-direction behavior for tokenizers is unrelated to ReduceLROnPlateau-like scheduler functionality." }, { "left": "issue:10519", - "right": "issue:11005", + "right": "issue:11307", "accept": false, - "reason": "Tokenizer truncation behavior and ReduceLROnPlateau-like scheduler functionality are unrelated features." + "reason": "Tokenizer truncation behavior and Wav2Vec2 word time offsets are different problems in different subsystems." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "A scheduler feature request and Wav2Vec2 word boundary timing are clearly unrelated." + "reason": "Learning-rate scheduling behavior is unrelated to speech model word-boundary timing offsets." } ] }, @@ -18081,7 +18088,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "1e9372e1fb645ee64fcca0822d10299badf59569cb69c9afd9668ea4383b93b4", + "prepared_review_unit_hash": "712cf169da8475617e702327fb998d8f38b3e4d28cdd2ff0d3b42a7626be73d5", "component_budget": { "node_count": 3, "item_count": 3, @@ -18103,24 +18110,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues do not look like the same bug. One is a tokenization discrepancy, one is a specific auth/error-message problem when downloading private models, and one is a generic model-loading failure for bert-base-uncased.", - "confidence": 0.91, - "canonical_issue_reason": "Issue #16618 is the broadest and most central-sounding of the three, with a direct model-loading failure title and the most discussion activity.", + "summary": "These issues do not look like true duplicates: one is about Tapas tokenization parity, and the other two are about model-loading/download failures with different causes and user-facing symptoms.", + "confidence": 0.28, + "canonical_issue_reason": "Issue #16618 is the broadest of the three and sits in the model-loading error area, so it is the best representative if one must be chosen for this noisy cluster.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the best cluster representative because it is the only issue about an actual model load failure; the other two describe unrelated tokenization and login/error-message concerns.", + "best_issue_reason": "#16618 is the most central issue for the cluster because it concerns a common model-load failure path, whereas #13244 is clearly about tokenization behavior and #15980 is specifically about private-model auth messaging.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tokenization output mismatch versus a model-loading failure are different problem classes and do not indicate the same underlying bug." + "reason": "Different problem domains: Tapas tokenization mismatch vs. model loading failure. No indication they share the same underlying bug or fix." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "#15980 is specifically about a private-model download/login error message, while #16618 is a separate generic failure to load bert-base-uncased; the concrete failure modes differ." + "reason": "Both mention model download/load errors, but the concrete scenarios differ: private-model login/error messaging vs. a failure loading 'bert-base-uncased'. Not enough evidence of the same bug." } ] }, @@ -18150,12 +18157,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "217cf50320d2e583e28d14f1f4301e35b504e9015e8333bdce6495a5a62c4f5c", + "prepared_review_unit_hash": "13e72cd1d3a00e2c96a8e8039d7ecc460aac681c1caa3156312c6f73f69b4901", "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 }, @@ -18163,7 +18170,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -18172,8 +18179,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate feature-request issues about adding support for different model/runtime architectures. They share a similar \u201ccommunity contribution\u201d framing, but the underlying changes are unrelated, so they should not be clustered as duplicates.", - "confidence": 0.97, + "summary": "These two issues are both community contribution requests, but they target different features and code paths: Flash Attention 2 support vs GGUF support. They are not duplicates and should stay separate.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -18183,7 +18190,7 @@ "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Both are broad support-request issues, but one asks for Flash Attention 2 support and the other for GGUF support. They target different code paths and different changes, so they are not the same underlying bug/change." + "reason": "Reject: the issues request different capabilities (Flash Attention 2 vs GGUF) for different architectures and underlying implementations. They share only a generic contribution framing, not the same bug or change." } ] }, @@ -18242,184 +18249,184 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "1bbf05884dab2f571f2b38b6e1f25a8b67aac1f087230263e5f2f5598c409679", + "prepared_review_unit_hash": "7d7c835ea1936c504d0b5a71695758e20b2cd20045d8714dd297a088c9206e10", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13511, - "estimated_input_tokens": 3378, - "estimated_eval_tokens": 7012 + "serialized_chars": 13529, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13511, - "estimated_input_tokens": 3378, - "estimated_eval_tokens": 7012 + "serialized_chars": 13529, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly templated i18n translation requests, so many soft similarities are not real duplicates. I\u2019d only merge the exact/near-exact placeholder subgroup and the identical Thai pair; the concrete language-specific requests are distinct issues.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43834 is the cleanest representative of the generic placeholder-based translation request subgroup: it has the exact template title and sits in the middle of the near-identical placeholder issues.", + "summary": "This cluster is noisy and mixes unrelated translation-request issues by language. The only clear duplicates are the placeholder-title translation issues (43834/44365/44370) and the exact Thai duplicate pair (46222/46226).", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43834 is the earliest of the largest near-identical placeholder-title subcluster, so it is the best representative issue for that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "Among the issues here, issue:43834 is the best global representative because it is the most central duplicate candidate in the placeholder cluster and has the strongest linkage to the near-identical template issues.", + "best_issue_reason": "43834 is the cleanest representative of the dominant duplicate-like subcluster; the other language-specific translation requests are separate issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different concrete translation requests; one is Turkish and the other is a generic template issue." + "reason": "Different languages/topics; only generic i18n wording overlaps." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different concrete translation requests; Turkish vs a generic template entry." + "reason": "Different language translation requests, not the same issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish-specific request does not look like the same issue as the generic placeholder translation template." + "reason": "Turkish translation request vs generic placeholder-title issue; not the same underlying issue." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different target languages; Turkish and Romanian are separate translation requests." + "reason": "Different language translation requests; similarity is only broad i18n template wording." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different target languages; Turkish and Spanish are separate translation requests." + "reason": "Turkish vs Spanish translation requests are distinct issues." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish-specific request vs generic template issue; not the same underlying change." + "reason": "Spanish translation request vs generic placeholder-title issue; no clear duplicate evidence." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish-specific request vs generic template issue; not a duplicate." + "reason": "Different language translation requests, not the same underlying issue." }, { "left": "issue:28936", - "right": "issue:38435", + "right": "issue:44370", "accept": false, - "reason": "Different target languages; Spanish and Romanian are separate translation requests." + "reason": "Different language translation requests, not the same underlying issue." }, { "left": "issue:28936", - "right": "issue:44370", + "right": "issue:38435", "accept": false, - "reason": "Spanish-specific request vs generic template issue; not the same issue." + "reason": "Spanish vs Romanian translation requests are distinct." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian-specific request vs generic template issue; different underlying requests." + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian-specific request vs generic template issue; not a duplicate." + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian-specific request does not match the generic placeholder translation template." + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different language and title content; Romanian request is separate from the Thai issue." + "reason": "Romanian translation request vs unrelated Thai-titled issue; no same underlying change." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different language and title content; not the same underlying request." + "reason": "Romanian translation request vs unrelated Thai-titled issue; no same underlying change." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same placeholder translation-request title and near-identical body; looks like a duplicate template issue." + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same generic title and nearly identical body text; likely the same template issue opened twice." + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Generic placeholder issue vs a concrete Thai-language title; not clearly the same underlying request." + "reason": "Placeholder-title translation issue vs different Thai-titled issue; not the same underlying request." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Generic placeholder issue vs a concrete Thai-language title; too different to merge." + "reason": "Placeholder-title translation issue vs different Thai-titled issue; not the same underlying request." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Both are the same generic placeholder translation issue with very similar bodies." + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Generic placeholder issue and a Thai-specific issue are not the same request." + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Generic placeholder issue and a Thai-specific issue are not the same request." + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Generic placeholder issue vs Thai-specific title; not a duplicate." + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Generic placeholder issue vs Thai-specific title; not a duplicate." + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "These are essentially identical Thai-title issues with matching body structure, so they look like the same underlying request." + "reason": "Identical Thai title and likely identical body; clear duplicate pair." } ] }, "evaluator_result": { "accept": true, - "feedback": "The clustering and soft-edge calls are grounded in the packet: the analyst limits merges to the clearly near-identical placeholder subgroup and the identical Thai pair, while keeping the language-specific translation requests separate. The representative/canonical choice is reasonably supported by the high overlap among the placeholder issues, and the verdicts stay conservative rather than over-claiming duplicate status for the concrete-language requests." + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate calls on 43834/44365/44370 are supported by identical placeholder titles plus very high body similarity, and 46222/46226 is a clear exact pair. The language-specific issues are correctly treated as separate rather than collapsed based on broad i18n similarity alone." }, "evaluator_used": true, "retried": false, @@ -18487,7 +18494,6 @@ "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", @@ -18501,274 +18507,268 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "2bdf1fe084442dbb30c400bfab00aa7fded1ba7dfa0dba80b4b378b0cb032b19", + "prepared_review_unit_hash": "ac1e94686d7e52323433fe8a06199721c81c58eed909a3aa61cda15a24be0e9c", "component_budget": { "node_count": 18, "item_count": 18, - "soft_pair_count": 40, - "serialized_chars": 24148, - "estimated_input_tokens": 6037, - "estimated_eval_tokens": 12330 + "soft_pair_count": 39, + "serialized_chars": 23784, + "estimated_input_tokens": 5946, + "estimated_eval_tokens": 12148 }, "budget": { "node_count": 18, "item_count": 18, - "soft_pair_count": 40, - "serialized_chars": 24148, - "estimated_input_tokens": 6037, - "estimated_eval_tokens": 12330 + "soft_pair_count": 39, + "serialized_chars": 23784, + "estimated_input_tokens": 5946, + "estimated_eval_tokens": 12148 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is thematically similar but not duplicate: it groups many distinct model-add/support requests across different model families, tasks, and modalities. No soft pair looks like the same underlying change.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical duplicate stands out; the items are distinct requests. If a hub is needed, issue 43671 is the most central, but it does not subsume the others.", + "summary": "This cluster is a loose collection of model-support request issues, not true duplicates. The soft-similarity links mostly reflect the shared \u201cadd model/support\u201d wording, but each item targets a different model or capability, so all soft edges should be rejected.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 43671 is a representative, well-scoped model-support request and sits near the center of the similarity graph, but it is still a specific Qwen3-TTS request rather than a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43671 is the best representative of the loose cluster because it has the broadest overlap and highest apparent centrality, though it is still not a true duplicate target.", + "best_issue_reason": "Issue 43671 is the best canonical issue only as a representative of the recurring feature-request pattern; it is clear and specific, but still not interchangeable with the other tickets.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config vs DEIMv2: different model families and unrelated requests." + "reason": "EuroBert config support vs DEIMv2 support; different models and different requests." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config vs Qwen3-TTS: different model and different task." + "reason": "EuroBert config support vs Qwen3-TTS support; unrelated model families and tasks." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config vs RF-DETR: unrelated model additions." + "reason": "EuroBert config vs RF-DETR; same broad theme but clearly different model additions." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR vs DEIMv2: different detection models, not the same change." + "reason": "RF-DETR and DEIMv2 are distinct model-add requests, not the same change." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR vs LFM2.5 Audio: different domains and model types." + "reason": "RF-DETR detection model vs LFM2.5 Audio support; different domains and code paths." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Detection model request vs sparse transformer request: unrelated." + "reason": "RF-DETR model add vs weight-sparse transformer request; unrelated concrete changes." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR vs ASR support: different model tasks and code paths." + "reason": "RF-DETR vs Qwen3-ASR; different models and feature requests." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR vs TTS support: different model families and tasks." + "reason": "RF-DETR vs Qwen3-TTS; distinct model-support tickets." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT vs ASR support: different model and modality." + "reason": "NeoBERT vs Qwen3-ASR; different model families and functionality." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs sparse transformer request: not the same underlying change." + "reason": "NeoBERT addition vs sparse-transformer support; not the same underlying change." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT vs audio model support: unrelated requests." + "reason": "NeoBERT vs LFM2.5 Audio; unrelated model requests." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR vs NeoBERT: different model additions entirely." + "reason": "RF-DETR and NeoBERT are separate model-add requests." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT vs ModernVBERT: distinct model requests." + "reason": "NeoBERT vs ModernVBERT; similar naming but different models and requests." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture vs ASR support: different model and task." + "reason": "MiniCPM3 architecture support vs Qwen3-ASR support; different model targets." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 vs sparse transformer request: unrelated." + "reason": "MiniCPM3 architecture vs sparse transformer; different concrete features." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 vs ModernVBERT: different model additions." + "reason": "MiniCPM3 vs ModernVBERT; distinct model support requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 vs Voxtral TTS support: different model families and tasks." + "reason": "MiniCPM3 architecture vs Voxtral-TTS support; different models and modalities." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 vs Qwen3-TTS: unrelated model requests." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model additions." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT vs ASR support: different model and modality." + "reason": "ModernVBERT vs Qwen3-ASR; different model support requests." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT vs TTS support: unrelated requests." + "reason": "ModernVBERT models vs Qwen3-TTS; unrelated requests." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "Specific ModernVBERT request vs generic RoPE note: not the same change." + "reason": "ModernVBERT model support vs BERT-like RoPE support; one is a model request, the other is an architectural feature request." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "Generic BERT-like RoPE issue vs Voxtral TTS support: different scope." + "reason": "BERT-like RoPE support vs Voxtral TTS support; different change types." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "Architecture note vs unrelated issue titled with a person name; no overlap." + "reason": "Architecture feature request vs unrelated person-named issue; no shared underlying bug/change." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS vs Qwen3-ASR: related theme but different submodels and changes." + "reason": "Qwen3-TTS vs Qwen3-ASR; related brand but different modalities and implementation work." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS vs Voxtral TTS: different model support requests." + "reason": "Qwen3-TTS vs Voxtral TTS; both audio-related, but different models and support paths." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "TTS support vs DeepSeek-OCR2: different modalities and code paths." + "reason": "Qwen3-TTS vs DeepSeek-OCR2; different models and tasks." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "TTS support vs sparse transformer request: unrelated." - }, - { - "left": "issue:43671", - "right": "issue:43909", - "accept": false, - "reason": "Qwen3-TTS vs LFM2.5 Audio: different models, not one duplicate." + "reason": "Qwen3-TTS support vs sparse transformer request; different code paths." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "ASR support vs audio model request: different model additions." + "reason": "Qwen3-ASR vs LFM2.5 Audio; both audio-related, but distinct model support issues." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR vs EUPE: different model families and tasks." + "reason": "Qwen3-ASR vs EUPE; different model/product targets." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "ASR support vs sparse transformer request: unrelated." - }, - { - "left": "issue:43837", - "right": "issue:45027", - "accept": false, - "reason": "Qwen3-ASR vs Voxtral TTS: different tasks and models." - }, - { - "left": "issue:43837", - "right": "issue:45177", - "accept": false, - "reason": "ASR support vs DeepSeek-OCR2: unrelated modalities." + "reason": "ASR model support vs sparse transformer request; not the same change." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio vs SAM3-LiteText: different model domains." + "reason": "LFM2.5 Audio vs SAM3-LiteText; different model additions." }, { "left": "issue:43909", "right": "issue:46645", "accept": false, - "reason": "Audio model request vs Granite SWA+sink option: different model families and change types." + "reason": "Audio model support vs Granite SWA+sink option; one is a model request, the other a model option tweak." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Sparse transformer request vs EUPE: unrelated model changes." + "reason": "Sparse transformer model request vs EUPE model request; unrelated additions." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR vs Voxtral-4B-TTS; different audio models and support work." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS support: different model families and tasks." + "reason": "NeoBERT vs Voxtral TTS; different model families and tasks." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR vs DeepSeek-OCR2; different models and modalities." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE vs TIPSv2: distinct vision-language model requests." + "reason": "EUPE by Meta AI vs TIPSv2 by Google DeepMind; separate model requests." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT vs unrelated issue; no shared underlying bug or feature." + "reason": "ModernVBERT model request vs unrelated name-only issue; no shared change." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support vs unrelated issue; not a duplicate." + "reason": "Voxtral TTS support vs unrelated name-only issue; not a duplicate." } ] }, @@ -18798,7 +18798,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "077804748ce75550a60ac7b49fbac40ba22c408aab8656b4bd75341f0114380c", + "prepared_review_unit_hash": "c0727bdfe6f78664f8ddedf6565280f285b6789fbed058b6f47da09605d3b963", "component_budget": { "node_count": 2, "item_count": 2, @@ -18820,18 +18820,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are unrelated: one is a docs request about adding Pruna as an optimization framework, while the other is a Vision RoPE refactor. There is no duplicate relationship or shared underlying change.", - "confidence": 0.99, + "summary": "These two issues are unrelated: one is a documentation/request item about adding Pruna as an optimization framework, and the other is a Vision RoPE refactor. They do not describe the same bug or change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "No duplicate cluster here; if forced to pick a representative, #46443 is the more substantive code-change issue, while #38740 is a docs request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "Different topics and scopes: documentation/framework listing vs. a vision positional-embedding refactor. They do not describe the same bug, feature, or code change." + "reason": "Different topics and outcomes: docs framework addition vs. Vision RoPE refactor. No shared code path, target, or change request." } ] }, @@ -18861,7 +18861,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "87cdc29d6dac62f8fd84c517025baae15901a5737fd6653249ee154d271e11b3", + "prepared_review_unit_hash": "c618097dc32aa465692d8dff4aece3bfb3e21d8afa347f6f846060a9d2f275fd", "component_budget": { "node_count": 2, "item_count": 2, @@ -18883,18 +18883,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is about block decoding, the other about adding DeepSeek v3.2 support. They share no explicit targets, filenames, or code-path evidence.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue: the titles describe unrelated work items rather than the same underlying bug or change.", + "summary": "These issues are unrelated: one asks for generic block decoding behavior, while the other requests DeepSeek v3.2 support.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good representative of the other; there is no duplicate cluster here to consolidate.", + "best_issue_reason": "No true canonical duplicate exists; if one issue must represent the set, 43037 is the more concrete, actionable request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different features with no shared implementation target or concrete bug/change; soft similarity is too weak to treat them as duplicates." + "reason": "Different underlying concerns: block decoding mechanics vs model/version support for DeepSeek v3.2." } ] }, @@ -18924,40 +18924,40 @@ "soft_pairs": [ "issue:41306|issue:41826" ], - "prepared_review_unit_hash": "e10eae019554e68fa63c5947f51201b463c4e23e3814d56bb6c6e5a08e4f2bb4", + "prepared_review_unit_hash": "4e5a026c557c0d44084188830cb16b66ab61ec9101b33a8f5dffd329672c07c4", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, - "estimated_input_tokens": 396, - "estimated_eval_tokens": 1048 + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, - "estimated_input_tokens": 396, - "estimated_eval_tokens": 1048 + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are only loosely related by a shared theme of tiled/memory-efficient layers, but they concern different components and different requests: one is about enabling a specific fused loss flag in `from_pretrained()`, the other is about integrating TiledMLP. They do not look like the same underlying bug or change.", - "confidence": 0.9, - "canonical_issue_reason": "Issue #41826 is the better canonical representative if this cluster is kept together because it describes a broader, self-contained feature request with a clear memory-footprint goal; #41306 is a narrower API-flag request for a different subsystem.", + "summary": "These two issues are only loosely related by the general theme of memory-efficient model components, but they describe different features: one is about exposing a TiledFusedLogitsLoss/Liger FLCE flag in `from_pretrained()`, the other is about integrating TiledMLP for lower memory usage. They do not look like duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue #41306 is the more concrete and actionable request, tied to a specific API flag and code path (`from_pretrained()`); #41826 is broader and more exploratory.", "canonical_pr_reason": null, - "best_issue_reason": "#41826 is the stronger issue overall: it is more general, more actionable as a feature request, and less tied to a specific implementation detail than #41306.", + "best_issue_reason": "#41306 is the best representative issue because it has a clear, specific implementation target, whereas #41826 is a higher-level feature idea.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41306", "right": "issue:41826", "accept": false, - "reason": "Different underlying changes: #41306 is about exposing a TiledFusedLogitsLoss/FLCE flag in `from_pretrained()`, while #41826 is about adding TiledMLP for memory savings. Shared memory/tiling wording is too broad to treat as duplicates." + "reason": "Different requested changes: one is an API flag for TiledFusedLogitsLoss/FLCE in `from_pretrained()`, the other is integrating TiledMLP. Same broad memory-efficiency theme is not enough to treat them as the same underlying issue." } ] }, @@ -18987,7 +18987,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "3e8b45c7e984fe03435af15762a585e5ff8216eb561306cf9d5d272650c72b62", + "prepared_review_unit_hash": "1b5ded1d467fb7607f8d993eb9ef6eddbe693f1d92c460a6c9cd56da09ddbca5", "component_budget": { "node_count": 2, "item_count": 2, @@ -19009,18 +19009,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are related only at a very broad level (both touch Trainer customization), but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. Not a duplicate cluster.", - "confidence": 0.95, - "canonical_issue_reason": "No true canonical duplicate here. If one must be chosen as the representative issue, #43598 is the more concrete and narrowly scoped Trainer behavior request, but it does not subsume #43600.", + "summary": "The two issues are related only at a high level (custom trainers / Trainer internals), but they describe different problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They do not look like duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue fits both: the first targets input-column handling in Trainer, while the second targets FLOPS accounting for custom trainers.", "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the clearer representative because it targets a specific Trainer customization hook (`remove_unused_column`), whereas #43600 is a separate FLOPS-accounting concern.", + "best_issue_reason": "Issue 43598 is the more concrete and narrowly scoped report, but it should not be treated as a duplicate of 43600.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying concerns: `remove_unused_column` customizability vs. FLOPS estimation for custom trainers. Shared Trainer context is too broad to treat as the same bug/change." + "reason": "Same broad Trainer/custom-trainer theme, but the underlying bugs are different code paths and different expected fixes." } ] }, @@ -19050,12 +19050,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "1a794d8c96bef7c840828fda902fa07f877f643c146af4c6ca0fadaa6c2d9247", + "prepared_review_unit_hash": "d6580d97390f98d5547d9ae4e6af3df283d43a48317e1ebedad321b9078b1b19", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1612, + "serialized_chars": 1611, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -19063,7 +19063,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1612, + "serialized_chars": 1611, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -19072,18 +19072,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are not duplicates: one requests Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). They target different model families and different heads.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 43684 is the broader registration request, but it is not the same underlying change as 44405; the topics differ by model family and task head.", + "summary": "These are related only at a very high level: both ask for additional AutoModel support for Qwen variants, but they target different models and different auto classes. They do not look like duplicates of the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more foundational request: it asks for Qwen3-Omni registration in AutoModel and AutoModelForConditionalGeneration, which is a generic model-registration gap rather than a specific task head request.", "canonical_pr_reason": null, - "best_issue_reason": "43684 is the more general issue about model registration, so it is the better cluster anchor if one must be chosen, though the pair should not be merged as duplicates.", + "best_issue_reason": "#43684 is the better canonical issue because it covers the general registration work and is more likely to subsume related follow-up support requests.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying requests: Qwen3-Omni model registration vs Qwen3.5 sequence classification support. They are related only at the broad AutoModel layer, not the same bug/change." + "reason": "Different underlying requests: #43684 is about registering Qwen3-Omni in AutoModel/AutoModelForConditionalGeneration, while #44405 is specifically about AutoModelForSequenceClassification support for Qwen3.5/Qwen3_5Config. Same ecosystem, but not the same bug or change." } ] }, @@ -19120,79 +19120,79 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "c69dda5ad600107fb73bf53de8bcf4b8a587a7a96687c749ae75b0f402b327d7", + "prepared_review_unit_hash": "e67092b1830081399685bd921ec0a35725cf8e6f9692f7caebdc89eb3789c4c8", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4265, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4265, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "43835 looks unrelated. The three \"huggingface\" issues have identical titles and no distinguishing metadata, so they likely duplicate the same report; however, 43859 being much later means timing is not strong evidence. No PRs are present.", - "confidence": 0.76, - "canonical_issue_reason": "Issue 43845 is the best canonical anchor among the three indistinguishable \"huggingface\" issues because it is the earliest and the metadata does not separate them.", + "summary": "Three issues share the same generic 'huggingface' title and appear to be near-duplicates; 'Fett' looks unrelated. Among the huggingface issues, 43845 is the best representative/canonical candidate because it is the earliest of the closely matching set.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 43845 is the earliest and most representative of the near-identical 'huggingface' issues (43845/43846/43859), with slightly more context than 43859 and strong textual overlap with 43846.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the best global issue representative: 43835 appears unrelated, and the remaining three issues share the same title with no differentiating details.", + "best_issue_reason": "43845 is the best global issue candidate because it anchors the cluster of highly similar huggingface reports and is the most plausible single dedup target.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no shared metadata suggesting the same underlying issue." + "reason": "Different titles and no substantive body evidence tying 'Fett' to the huggingface reports." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles with no evidence they report the same bug or change." + "reason": "Different titles and no concrete evidence of the same underlying issue." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "\"Fett\" does not look like the same issue as the \"huggingface\" reports." + "reason": "Different titles and no concrete evidence of the same underlying issue." }, { "left": "issue:43845", "right": "issue:43846", "accept": true, - "reason": "Identical titles and no distinguishing metadata make them look like duplicate reports." + "reason": "Same title, very high textual similarity, and no visible evidence of distinct bugs; these look like duplicate reports." }, { "left": "issue:43845", "right": "issue:43859", "accept": true, - "reason": "Same title and no differentiating details; the later timestamp alone is not enough to separate them." + "reason": "Same generic title and strong similarity suggest the same underlying issue despite the shorter body." }, { "left": "issue:43846", "right": "issue:43859", "accept": true, - "reason": "Identical titles and sparse metadata suggest the same underlying issue despite the time gap." + "reason": "Same generic title and strong similarity make these likely duplicate reports of the same issue." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The grounding is mostly fine for 43835 vs. the three \"huggingface\" issues, but the writeup overstates the timing evidence. 43845 and 43846 are close together, yet 43859 is about 13 hours later, so \"opened within a short window\" and \"similar timing\" are not well supported for all three pairings. The duplicate claim among the three huggingface issues may be plausible from identical titles alone, but it should be stated more conservatively. No PRs are present, so the PR side is fine." + "accept": true, + "feedback": "Grounded in the packet: the three 'huggingface' issues have strong soft-similarity, and 43845 is the earliest among that cluster. The 43835 links are conservatively rejected based on only weak similarity and unrelated title." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -19227,88 +19227,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "9de244f375ad2fd904eb3bd3e34dd065f90f9fdd571f579e2df57ae4f6d7629b", + "prepared_review_unit_hash": "adbd4a222bbdc71b9919b24174fd3e0df6f6ae46d7ef0fd395ee059daede2359", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is low-signal: the items are generic spam/placeholders and one vague docs request, with no shared bug or concrete change to deduplicate.", - "confidence": 0.98, - "canonical_issue_reason": "No reliable canonical issue: the titles are either spam or unrelated generic text, and none describe the same underlying problem.", + "summary": "All items are low-information spam/vague issues with no concrete shared bug or change. The apparent similarity is only from generic titles, so none of the soft pairs should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 44418 is the least spam-like and the only one that resembles a topical request, so it is the least bad representative, but it is still not a strong canonical anchor.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both look like spam placeholders; same superficial title pattern is not evidence of the same underlying issue." + "reason": "Both are generic spam placeholders with no body, targets, or filenames; no concrete shared issue is visible." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both look like spam placeholders; there is no concrete shared bug or feature request." + "reason": "Same generic spam-only pattern; no evidence they describe the same underlying bug or request." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam and the other is a vague documentation suggestion; they do not describe the same problem." + "reason": "One is spam-like and the other is a vague documentation suggestion; too little specificity to treat as the same issue." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus a different vague text title; no shared underlying issue." + "reason": "The titles are unrelated and there is no issue detail tying them to a common bug or task." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Generic documentation request versus spam placeholder; not the same artifact or fix." + "reason": "A vague documentation enhancement is not the same as a spam placeholder; no shared concrete change request." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Generic documentation request versus spam placeholder; no concrete overlap in bug or change." + "reason": "Only broad low-signal similarity; no explicit overlap in problem statement or requested change." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but the cluster signal is too weak to treat them as the same underlying issue." + "reason": "Both are spam placeholders, but there is no evidence they refer to the same underlying report rather than separate junk submissions." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and unrelated vague text title; no shared concrete problem." + "reason": "No shared topic beyond generic similarity; the titles do not indicate the same bug or request." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and unrelated vague text title; no shared concrete problem." + "reason": "No concrete common issue is present; these look like unrelated reports." } ] }, @@ -19338,12 +19338,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "ba9cf472b1d9e6802f559f45366114d26276bd6072083e7423b1250d6b284dec", + "prepared_review_unit_hash": "db1bc3ea61423c719476563c174d36dba7f3f6193a245800423d5b712b8604a2", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -19351,7 +19351,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -19360,24 +19360,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are very likely duplicates: they have identical titles, very similar-sized bodies, and a soft-similarity signal, but the packet does not expose the body text itself, so the match should be treated conservatively. #45145 is the better canonical issue because it was opened first and already has one inbound reference.", - "confidence": 0.8, - "canonical_issue_reason": "#45145 is the best canonical representative because it was created earlier and has an inbound reference; the available evidence suggests a likely duplicate, but not enough to claim definitive sameness from the packet alone.", + "summary": "The two closed issues have identical titles, close creation times, and very similar body sizes, so they look like duplicate reports of the same underlying request. Issue #45145 should be the canonical thread because it was opened first and already has an inbound reference.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #45145 is the better canonical issue because it was created first and already has an inbound reference, making it the more established thread.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest single issue to keep as the representative since it is earlier chronologically and already has slightly more linkage.", + "best_issue_reason": "Issue #45145 is the strongest cluster representative since it predates #45151 and is the more established issue thread.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Accept as a likely duplicate: the titles are identical and the soft-similarity signal is strong, but the evidence is still limited enough that the match should be treated conservatively." + "reason": "Identical titles, close timestamps, and highly similar body lengths with soft-similarity support treating these as duplicate reports of the same issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "The pairing is plausible, but the writeup is a bit too strong for the evidence available. The packet only gives identical titles plus a soft_similarity signal; it does not show the body text, so \"same report\" / \"effectively the same issue\" is stronger than the packet directly supports. The canonical choice of #45145 is fine on chronology and inbound reference, but the confidence should be much lower than 0.99 and the verdict should stay explicitly conservative (e.g. \"likely duplicate\" rather than definitive sameness)." + "feedback": "Reject: the pair looks plausibly duplicate based on identical titles, close timestamps, similar body lengths, and soft-similarity, but the analyst overstates the evidence by claiming the bodies have the \"same wording\" and that #45145 has \"slightly more engagement.\" Those specifics are not directly grounded in the packet. The canonical issue choice is fine, but the cluster summary and best-issue rationale should stay more conservative." }, "evaluator_used": true, "retried": true, @@ -19458,40 +19458,40 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "69c520ce896620f1ef292e502ca6619b8f63da85c00ab1b7544855fb174c9b97", + "prepared_review_unit_hash": "07aab97c34af37d97bb703c8aeeff089d607bff3e48f62da58df57ea0acf8aff", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1673, - "estimated_input_tokens": 419, - "estimated_eval_tokens": 1094 + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1673, - "estimated_input_tokens": 419, - "estimated_eval_tokens": 1094 + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a broad level (4-bit/quantization inference), but they describe different failure modes in different code paths: Gemma4 audio-module uint8 weights hitting torch.finfo() versus BitsAndBytesConfig skip-module handling clearing lm_head exclusion. I would not merge them as duplicates.", - "confidence": 0.92, - "canonical_issue_reason": "Issue #45674 is the more general quantization-config bug and is framed around a reusable inference path rather than a model-specific symptom.", + "summary": "These are two separate 4-bit inference bugs: one in Gemma4 audio modules hitting a uint8/torch.finfo TypeError, and one in BitsAndBytesConfig where llm_int8_skip_modules drops the default lm_head exclusion and triggers an AssertionError. They are related only at the broad feature level, not the same underlying defect.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45674 is the more general, configuration-focused report with a clearer failure mode and broader applicability across 4-bit inference.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the better representative for the cluster because it describes a concrete, user-visible 4-bit inference regression in BitsAndBytesConfig that is broader than the Gemma4-specific torch.finfo() failure.", + "best_issue_reason": "Issue 45674 best represents the cluster because it describes a concrete regression in BitsAndBytesConfig rather than a model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Different underlying bugs and code paths: one is a Gemma4 audio-module dtype/torch.finfo() problem, the other is BitsAndBytesConfig skip-module handling that removes the default lm_head exclusion and triggers an AssertionError." + "reason": "Different root causes and code paths: Gemma4 audio-module dtype handling vs. BitsAndBytesConfig skip-module exclusion logic. They share 4-bit inference context but are not the same bug." } ] }, @@ -19521,40 +19521,40 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "59148f5a03b28f17b3a5d44ab8dd8515a71f028512c967167fe47a2d05a2a192", + "prepared_review_unit_hash": "665531712120e22ef4ee9ac8fc2c898d4b840e3265f95f6a50c44c8d91361548", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1475, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1475, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues look like distinct entries about different people and dates, not the same underlying bug/report. The soft duplicate edge should be rejected.", - "confidence": 0.94, + "summary": "The two closed issues look unrelated: both are bare name/date titles with no body, no shared targets, and no evidence of the same underlying bug or request. The soft similarity is not enough to treat them as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single canonical issue: the titles refer to different names and dates, with no shared context beyond a surname and formatting.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different subjects ('Mirko Privitera' vs 'Ivan privitera') and different dates; this looks like two separate records, not the same issue." + "reason": "Titles refer to different people and different dates; there is no issue body, shared target, or other evidence of the same underlying problem." } ] }, @@ -19584,7 +19584,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "ef10284f0b3f7c0d5e0080ec5b03514421542c87b3ec36687cbcd7f0784afe8b", + "prepared_review_unit_hash": "02540700f82477fae749111438a1843e60d6d625f7645aa410efd2b2c20aedc5", "component_budget": { "node_count": 2, "item_count": 2, @@ -19606,18 +19606,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are closed issues with empty bodies and non-technical, name/date-style titles. They do not describe the same bug or any actionable product problem, so this looks unrelated rather than duplicate clustering.", + "summary": "The two closed issues appear unrelated; both are just person-name/date titles with no body, links, or shared targets, so there is no evidence they describe the same underlying bug or request.", "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a suitable canonical issue: both are effectively placeholder/personal-name entries with no bug description or technical context.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "No evidence of the same underlying bug or request; the titles are just different names/dates and both bodies are empty." + "reason": "No substantive overlap beyond both being empty closed issues; titles point to different people/dates and there is no shared body, target, or code context." } ] }, @@ -19647,12 +19647,12 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "f9fcd93bd073b7b4886cc6eca0d65299ceceacffcf861402180fdd3d2273241d", + "prepared_review_unit_hash": "2b1fe1dd88b7166227b5e43a9b205589eea44af95dcf9bb0c8a4d9c719fec55c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1542, + "serialized_chars": 1541, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -19660,7 +19660,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1542, + "serialized_chars": 1541, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -19669,18 +19669,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are unrelated feature requests: one asks for a new Qwen3.5 token-classification head for value modeling, the other asks for FP8 training support. They do not describe the same underlying change.", + "summary": "The two issues are unrelated: one requests a new Qwen token-classification model class, while the other asks for FP8 training support. They do not describe the same bug or change.", "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 45810 is the more concrete, narrowly scoped request and is easier to use as a representative item, but it is not a duplicate of the FP8 training issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different feature areas and code paths: model-head addition vs FP8 training support. Similarity is superficial, not a duplicate." + "reason": "Different features in different areas: adding a model class for token classification versus training/runtime support for FP8. No shared code path or concrete duplicate behavior is evident." } ] }, @@ -19720,12 +19720,12 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "178ad3c13b677d61b147ccd2542d73beb99df40703b17d0725b2678d814157c8", + "prepared_review_unit_hash": "c2a695b21afc2d5d46a0dafca58f5de3b7b36ad35b276a385a98602fa6d3d348", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -19733,7 +19733,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -19742,69 +19742,69 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster contains two exact-duplicate issue pairs: 45870/45871 and 45872/45873. Issue 45869 is a separate financial data bundle/sample request and should not be merged with the example/scraping-service issues. The canonical/best anchor should be justified by exact-title duplication and chronology, not by claiming one identical title is more generic.", - "confidence": 0.92, - "canonical_issue_reason": "45870 is a reasonable canonical anchor for the exact-duplicate pair 45870/45871 because it is one of the identical-title issues and was created first.", + "summary": "There are two clear duplicate pairs: 45870 \u2194 45871 and 45872 \u2194 45873. Issue 45869 is a different dataset/promo request and should not be grouped with either pair. The two DataForge-themed pairs are superficially similar but describe different scopes, so they should not be merged across pairs.", + "confidence": 0.96, + "canonical_issue_reason": "45872 is the strongest canonical representative for the cluster because it is the more concrete DataForge request and is an exact duplicate of 45873; the other exact-duplicate pair (45870/45871) is a separate, more generic example issue.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is a solid global issue representative because it participates in an exact duplicate pair and there is no evidence that it differs materially from 45871; using it as the anchor is consistent with the cluster structure.", + "best_issue_reason": "45872 is the most actionable issue in the set: it clearly states the DataForge time-series + scraping-service request, while 45869 is unrelated and 45870/45871 are a separate duplicate pair about a generic example.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different concrete requests: a financial data bundle/sample versus a financial time-series example." + "reason": "Different requests: dataset bundle promo/sample vs. a DataForge example." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different concrete requests: a financial data bundle/sample versus a financial time-series example." + "reason": "Different requests: dataset bundle promo/sample vs. a DataForge example." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact duplicate match: identical title and no evidence of differing scope or work." + "reason": "Same title/body pattern and effectively the same example issue; this looks like a direct duplicate." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Same broad financial theme, but different concrete request: time-series example versus a scraping service addition." + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same broad financial theme, but different concrete request: time-series example versus a scraping service addition." + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Overlapping topic, but not the same underlying issue; the scopes are different." + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Overlapping topic, but not the same underlying issue; the scopes are different." + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact duplicate match: identical title and no evidence of differing scope or work." + "reason": "Same title/body pattern and effectively the same DataForge request; this looks like a direct duplicate." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, but the anchor/canonical rationale overstates what is evidenced: 45870 and 45871 have identical titles, so claims that 45870 is the \u201ccleaner, more generic wording\u201d or \u201cmost generic phrasing\u201d are not supported by the packet. The duplicate-pair verdicts themselves are conservative, but the higher-level justification should avoid unsupported distinctions." + "accept": true, + "feedback": "The summary is grounded in the packet: 45870\u219445871 and 45872\u219445873 are the only clearly identical pairs (matching titles/body lengths and jaccard 1.0), while 45869 is a separate dataset bundle request. The soft-edge rejections between the example issues and the DataForge service issues are conservative and consistent with the titles." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -19828,49 +19828,46 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "4886d8e9cdbaeea58cdb0bfba3aa6537353c15e6550e4793ca410ef701b5d3af", + "prepared_review_unit_hash": "b3b9a949e9f8d77070222ac37fd32260fbc11d12e471ccc9c825c74e3188bcc6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are plausibly related around static-cache handling and linear-attention, but the evidence here is limited to titles and a soft-similarity signal. That supports a possible linkage, not a firm conclusion that they are the same bug.", - "confidence": 0.63, - "canonical_issue_reason": "Issue 46439 is the better canonical issue because it states the broader mechanism (`early_initialization` corrupting linear-attention layers), which could encompass the more symptom-specific Qwen3.5 crash report.", + "summary": "Both issues concern `cache_implementation=\"static\"` and linear-attention, but they describe different failure modes: one is `early_initialization` producing wrong shapes and a `RuntimeError`, the other is a Qwen3.5 `KeyError: 'linear_attention'`. I would not merge them as the same underlying bug from the available evidence.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 46439 is the better representative because it names the broader root path (`early_initialization` + linear-attention shape corruption) rather than a model-specific crash symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46439 is the strongest cluster representative since it is more general and mechanism-focused, while 46441 is a narrower model-specific failure symptom.", + "best_issue_reason": "Issue 46439 is the strongest standalone report: it is more general, more diagnostic, and describes the concrete corruption mechanism rather than just a downstream exception.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "The titles suggest they may be related through static cache / linear-attention behavior, but we only have soft similarity and no shared targets or code-level evidence. That is not enough to confirm the same underlying bug." + "reason": "Same broad static-cache/linear-attention area, but different observed failures (`RuntimeError` from shape corruption vs `KeyError` for Qwen3.5); not enough to treat as the same bug." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The cluster linkage is plausible, but the writeup is a bit too strong for the evidence in the packet. We only have titles and a soft-similarity signal; that supports a possible relationship, not that both reports are the same underlying bug. Phrasing like \u201clikely reflect the same underlying bug\u201d and \u201cboth issues point to the same failure mode\u201d overstates what is grounded here. The soft-edge verdict should stay more conservative (e.g., \u201cmay be related\u201d or \u201ccould share a cache/linear-attention root cause\u201d)." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -19894,7 +19891,7 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "64facf85ec321dd252f7326dcc659cccc8e1dff03cf55292e65bbae10d2df505", + "prepared_review_unit_hash": "e336edbe9e1faab0498ee7abb05e6d0c3c963c494c05ebe71f930afcc0267cd5", "component_budget": { "node_count": 2, "item_count": 2, @@ -19916,24 +19913,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are duplicates with identical titles and the same Wav2Vec2CTCTokenizer v5 special-token leak report; issue #46552 is the better canonical thread because it is the open, referenced copy.", + "summary": "These two issues appear to be the same bug report for Wav2Vec2CTCTokenizer v5 treating the word delimiter as a special token, affecting special-tokens masking and token conversion. They differ only by issue number/state, so they should be clustered together.", "confidence": 0.99, - "canonical_issue_reason": "#46552 is the open issue and already has inbound references, so it is the better place to continue discussion and track any fix.", + "canonical_issue_reason": "Issue #46552 is the better canonical because it is open, has inbound references, and appears to be the active report for the same exact problem.", "canonical_pr_reason": null, - "best_issue_reason": "#46552 best represents the active duplicate cluster: same bug report, still open, and already linked from other discussions.", + "best_issue_reason": "Issue #46552 is the strongest representative since it is the active open instance of the identical title/body and is more likely to collect ongoing discussion and resolution.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Same title and same underlying tokenizer bug; this is a true duplicate pair." + "reason": "Exact same title and same reported tokenizer bug; no meaningful difference in scope or content beyond issue metadata." } ] }, "evaluator_result": { "accept": true, - "feedback": "The conclusion is well grounded: both issues have identical titles and identical body lengths, and the open issue #46552 has inbound references, which supports treating it as the canonical thread. The soft-edge duplicate verdict is conservative and consistent with the packet evidence." + "feedback": "The analyst\u2019s clustering is grounded in the packet: both issues have the exact same title and identical body length, and the open issue (#46552) has inbound references while the other is closed. The soft-edge verdict is conservative and supported by the evidence, so accepting the pair is reasonable." }, "evaluator_used": true, "retried": false, @@ -19960,7 +19957,7 @@ "soft_pairs": [ "issue:46657|issue:46692" ], - "prepared_review_unit_hash": "b1ababf53d557765f5345e0eb09700fd51373e81496a28d17db7ca6b52ed9b7b", + "prepared_review_unit_hash": "98e4d0399c058e384b25bc0ba7abde8ef44ec27987bfefbd245c091be2f9840d", "component_budget": { "node_count": 2, "item_count": 2, @@ -19982,18 +19979,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are superficially similar because both involve object detection training, but they describe different failure modes: a tensor size mismatch vs. out-of-memory after several epochs on RT-DETR v2. They do not look like the same underlying bug.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 46657 is the better representative because it describes a concrete, reusable training failure class (tensor shape mismatch) rather than a model-specific resource exhaustion symptom.", + "summary": "These two issues are not duplicates: one reports a tensor shape mismatch during object-detection training, while the other reports an out-of-memory problem after several epochs in RT-DETR v2. They may both occur in training workflows, but they describe different failure modes and likely different fixes.", + "confidence": 0.91, + "canonical_issue_reason": "Issue #46657 is the better representative because it describes a concrete, actionable code-path failure (tensor size mismatch) rather than a broad resource symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46657 is the best standalone issue to keep as the cluster representative; 46692 appears to be a separate RT-DETR v2 memory problem and should not be merged into the same bug report.", + "best_issue_reason": "#46657 is the most suitable issue to anchor the cluster because it is specific and reproducible enough to guide debugging; #46692 appears to be a separate training-memory regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46657", "right": "issue:46692", "accept": false, - "reason": "Different symptoms and likely different root causes: tensor dimension mismatch vs. OOM during later epochs. Similar training context is not enough to treat them as duplicates." + "reason": "Different underlying problems: shape mismatch vs OOM after epochs. Same training context is not enough to treat them as the same bug." } ] }, @@ -20007,7 +20004,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20053,7 +20050,7 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "d518b790b8ec27c77f52419b9e94291b2fff200b36258530da242f279b87b9d2", + "prepared_review_unit_hash": "498fdf428c0bd5df3d06fce3c155652f2f420e5c21d2add8f0ba171e6714005f", "component_budget": { "node_count": 14, "item_count": 14, @@ -20075,132 +20072,132 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated PRs that only share broad soft-similarity hints. The only true duplicate pair is the two `fixed_cross_entropy` kwarg fixes, which target the same loss-path bug.", - "confidence": 0.93, + "summary": "This cluster is mostly heterogeneous: model additions, docs, tokenizer/processor tweaks, refactors, and tests that do not describe the same underlying change. The only true duplicate-like pair is the two `fixed_cross_entropy` PRs, which both implement the same cross-entropy kwargs support for issue 43240.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the clearest canonical representative for the duplicate loss fix: it explicitly targets issue 43240 and forwards the new `cross_entropy` kwargs in `loss_utils.py`.", + "canonical_pr_reason": "PR 43251 is the best representative of the only real duplicate pair because it explicitly targets issue 43240 and presents the same `fixed_cross_entropy` kwargs expansion with clearer issue context.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the best overall match for the only real duplicate pair because it most directly and completely addresses the shared `fixed_cross_entropy` bug.", + "best_pr_reason": "PR 43251 is the strongest cluster representative for the duplicate loss change; the rest of the cluster is unrelated, and 43251 is the cleaner, issue-linked version of the same fix.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Both are unrelated model/test additions; no shared bug or code path." + "reason": "VideoPrism model/docs addition vs. an ExecuTorch dynamic-shapes export test; unrelated changes." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Mistral tokenizer methods vs MobileLLM model addition are unrelated changes." + "reason": "Mistral tokenizer helper methods vs. adding MobileLLM; different subsystems and no shared bug." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer API fix and ExecuTorch dynamic-shape test are different features." + "reason": "Tokenizer API addition vs. ExecuTorch export testing; not the same underlying change." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness work and MobileLLM model implementation do not fix the same issue." + "reason": "Processor-auto robustness fix vs. MobileLLM model addition; unrelated code paths." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function kwargs fix vs a new model addition are unrelated." + "reason": "Cross-entropy kwargs fix vs. MobileLLM model addition; no overlap in underlying bug." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Cross-entropy kwargs support and Mistral4 test fixes are different problems." + "reason": "Loss utility kwargs fix vs. Mistral4 test/model work; different bug and code path." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch export tests and Mistral4 fixes are unrelated." + "reason": "Both touch tests, but one is ExecuTorch export dynamic-shape coverage and the other is Mistral4 test/model fixes; not the same change." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export testing and MobileLLM implementation are unrelated." + "reason": "ExecuTorch dynamic-shapes export test vs. MobileLLM model addition; unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New model code and config-resolution behavior are different changes." + "reason": "New MobileLLM model support vs. a `cls.config` resolution fix; unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM addition and Falcon output-interface refactor are unrelated." + "reason": "MobileLLM model addition vs. Falcon output-collection refactor; different model families and goals." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor and model-creation skill are clearly different scopes." + "reason": "Falcon refactor vs. model-creation skill docs/workflow; not the same code change." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output collection refactor and Mistral4 test fixes are not the same bug." + "reason": "Falcon output-interface refactor vs. Mistral4 test fixes; different models and bug targets." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test fixes and `layrnorm` renaming are unrelated model-specific changes." + "reason": "Mistral4 test fixes vs. `layrnorm` renaming across CLIP-like models; unrelated fixes." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 fixes and chat response parsing updates do not share a concrete code-path bug." + "reason": "Mistral4 test fixes vs. serve chat parsing update; different subsystems and changes." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes and TDT loss kernel integration are unrelated." + "reason": "Mistral4 test fixes vs. TDT loss kernel integration; unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test fixes and config-resolution order fix are different issues." + "reason": "Mistral4 test/model fixes vs. class config resolution order fix; different code paths." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 test fixes and model-creation skill are unrelated." + "reason": "Mistral4 test fixes vs. model-creation skill docs/workflow; not duplicates." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition and skill documentation/tooling are not the same change." + "reason": "MobileLLM model addition vs. model-creation skill; unrelated." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix the same `fixed_cross_entropy` bug by adding supported kwargs to the same function in `loss_utils.py`; they are mergeable into one change." + "reason": "Both PRs change `fixed_cross_entropy` to pass additional `cross_entropy` kwargs for issue 43240; same underlying bug and mergeable into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative. The only clearly mergeable duplicate pair is PR 43251 and PR 43254: both modify `src/transformers/loss/loss_utils.py`, target issue 43240, and address the same `fixed_cross_entropy` kwargs bug. The other pair verdicts are appropriately rejected as unrelated or too weakly similar to justify a single bugfix PR." + "feedback": "Grounded and conservative overall. The summary correctly treats the cluster as heterogeneous and only elevates the 43251/43254 pair, which is supported by the packet: both touch `src/transformers/loss/loss_utils.py`, both target issue 43240, and both expand `fixed_cross_entropy` to forward `weight`/`label_smoothing` to `nn.functional.cross_entropy`. The other soft pairs are reasonably rejected as unrelated." }, "evaluator_used": true, "retried": false, @@ -20211,7 +20208,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20227,7 +20224,7 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "92ba5bc805ce01deac01342828017c3dd417ebc88926a9edd0924ca00be57dc2", + "prepared_review_unit_hash": "4265ff26508be1eb13fc57e155ae2191bf7c6187969ac8b018e57aefd7e93e9c", "component_budget": { "node_count": 2, "item_count": 2, @@ -20249,18 +20246,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs around VibeVoice overlap heavily in filenames, but they appear to address different deliverables: the main VibeVoice integration versus a separate VibeVoice Realtime addition. They should not be merged as duplicates.", - "confidence": 0.9, + "summary": "Both PRs are about VibeVoice, but they are not the same change: #40546 is the core VibeVoice implementation, while #42781 adds a separate realtime/acoustic-tokenizer path and generation changes. The filename overlap is expected for a follow-on feature, not a duplicate fix.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #40546 is the base VibeVoice implementation and the broader, more canonical integration; #42781 is a later draft for a realtime extension with additional tokenizer/generation changes.", + "canonical_pr_reason": "#40546 is the broader, original VibeVoice implementation and is the better anchor for the cluster; #42781 looks like a later extension focused on realtime support rather than the same mergeable change.", "best_issue_reason": null, - "best_pr_reason": "#40546 is the best representative PR for the cluster because it introduces the core model rather than a specialized realtime follow-on.", + "best_pr_reason": "#40546 is the most representative and complete PR for the VibeVoice model introduction, with far more review activity and broader coverage than the draft realtime add-on.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Shared VibeVoice files reflect a common subsystem, but the PRs target different changes: core VibeVoice support versus a realtime variant with extra tokenizer and generation work. They are related, not the same fix/change." + "reason": "They share the VibeVoice codebase, but #42781 is a distinct realtime feature branch (including acoustic tokenizer and generation tweaks), not the same concrete bug/change as #40546." } ] }, @@ -20274,7 +20271,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20290,7 +20287,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "b0997cb2ec00aa3f81ea7c984d0ea46747a8f83ad232985e76fcb26c6e9ddb2c", + "prepared_review_unit_hash": "fc563cab6cdbf2abe23c03b8e823c8b7168af4f1d328359e2edf8d9031933c41", "component_budget": { "node_count": 2, "item_count": 2, @@ -20312,18 +20309,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They do not look like the same underlying change, so the soft duplicate edge should be rejected.", - "confidence": 0.97, + "summary": "The two PRs are unrelated feature additions in different model families, so they should not be clustered as duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical duplicate PR emerges here because the PRs address different code paths, filenames, and model families rather than the same bug or feature.", + "canonical_pr_reason": "No canonical PR: the cluster contains two unrelated PRs, one for TimesFM covariates and one for DINOv3 image classification support.", "best_issue_reason": null, - "best_pr_reason": "PR #41224 is the cleaner representative of the cluster: it is a focused DINOv3 image-classification support change with tightly scoped model and auto-class updates, whereas #40755 is a separate TimesFM forecasting feature.", + "best_pr_reason": "Neither PR is a good cluster representative for duplicate triage because they address different code paths and different models.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different model families and different capabilities: TimesFM covariate forecasting is unrelated to DINOv3 image classification support, so they are not the same underlying change." + "reason": "Different model subsystems and different changes: TimesFM forecasting covariates vs DINOv3 image classification support. They do not look like the same underlying bug or mergeable fix." } ] }, @@ -20337,7 +20334,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20400,7 +20397,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20416,7 +20413,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "d3de80acc7cdaa88889b6f6e955301ff46d9c90f27de3d13e4116e3627e80e68", + "prepared_review_unit_hash": "51d36f3504c2d0785e6f009c3dbc87d9f6338d06c098d70bc3873e8c6653f3eb", "component_budget": { "node_count": 2, "item_count": 2, @@ -20438,18 +20435,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate GitHub Actions maintenance PRs: 42975 upgrades core workflow actions for Node 24 compatibility, while 42976 upgrades docker/build-push-action versions. They overlap in workflow files but not in the concrete change or failure mode.", - "confidence": 0.94, + "summary": "Both PRs are workflow maintenance updates, but they target different dependency bumps and different failure modes: 42975 is centered on GitHub Actions Node 24 compatibility, while 42976 upgrades Docker build actions and other workflow action versions. They overlap in CI files, but not in the same concrete change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the broader representative workflow-actions upgrade, but it is still distinct from 42976 because it targets checkout/cache/upload-artifact compatibility rather than Docker build action versions.", + "canonical_pr_reason": "No canonical PR: the PRs are related only at a broad CI-maintenance level, but they update different actions for different reasons and are not the same underlying change.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the better cluster representative only because it is the wider sweep across many workflows; however, it should not be merged with 42976 since the underlying dependencies and change goals differ.", + "best_pr_reason": "42975 is the better representative of the broader GitHub Actions compatibility effort because its title and diff clearly describe the Node 24 upgrade theme across many workflows.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Different action families and fix targets: 42975 is about Node 24 compatibility for general workflow actions, while 42976 updates docker/build-push-action. Shared workflow filenames alone are not enough to treat them as the same PR." + "reason": "Reject: both are CI/workflow dependency bumps, but 42975 upgrades actions for Node 24 compatibility while 42976 updates docker/build-push-action and other actions to newer versions. They touch overlapping files yet do not fix the same concrete code-path problem and would not naturally merge into one PR." } ] }, @@ -20463,7 +20460,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20495,7 +20492,7 @@ "pull_request:45490|pull_request:46705", "pull_request:46562|pull_request:46658" ], - "prepared_review_unit_hash": "244a5b60cd81302f5340a69f6807d680d7dae7aa3debd1c46cc13542599c5a69", + "prepared_review_unit_hash": "61fbb237f91aca9a808b29df9aef39a6c7a9adbc2f6da432d0b7aa07f0e15b53", "component_budget": { "node_count": 7, "item_count": 7, @@ -20517,84 +20514,84 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This set is mostly unrelated model-addition PRs that only look similar because they all touch docs and auto-registration files. The only plausible duplicate pair is the two Step 3.7 Flash PRs; even that is weak because they appear to be different implementation tracks with different module names and file sets.", - "confidence": 0.92, + "summary": "These PRs are mostly unrelated model-addition branches that only overlap on standard docs/auto-registration files. The only likely duplicate-like pair is the two Step 3.7 Flash PRs (46562 and 46658); everything else appears to be a distinct model or feature addition.", + "confidence": 0.78, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR fits the whole cluster: these are distinct feature additions (ViT NEPA, Molmo, Molmo2, CTSM, Step3p7, distributed config). If forced to pick an anchor, PR 46562 is the closest because it is the only one with a strong apparent duplicate candidate.", + "canonical_pr_reason": "46562 is the clearest representative of the one near-duplicate pair: it is the more polished Step 3.7 Flash model-addition PR, while 46658 looks like a WIP/companion implementation for the same feature.", "best_issue_reason": null, - "best_pr_reason": "PR 46562 is the best duplicate anchor only because it has the strongest near-match with another PR in this set; otherwise the PRs are not duplicates and should be triaged separately.", + "best_pr_reason": "46562 is the best standalone PR to anchor the cluster because it corresponds to the same Step 3.7 Flash feature as 46658 and is the more complete user-facing addition; the rest of the PRs are separate model introductions.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto-registry edits are superficial, not the same underlying change." + "reason": "Different model families (ViT NEPA vs Molmo). Shared docs/auto files are just boilerplate registration, not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo2. No shared bug or code path beyond generic model-registration plumbing." + "reason": "Different model families (ViT NEPA vs Molmo2); same scaffolding files do not indicate the same code-path change." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model families: ViT NEPA and CTSM. Similar auto-mapping/docs churn is not enough to merge as duplicates." + "reason": "Unrelated model additions (ViT NEPA vs CTSM). Overlap is only in generic docs/auto wiring." }, { "left": "pull_request:42978", "right": "pull_request:46705", "accept": false, - "reason": "ViT NEPA model addition vs distributed-config infrastructure. Completely different change surface and purpose." + "reason": "A new model PR and a distributed-config PR are different kinds of changes with no shared concrete bug or feature." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are related names but they are separate model implementations with different files and scopes, so not the same PR." + "reason": "Molmo and Molmo2 are distinct model releases/families, not the same underlying implementation or fix." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are unrelated model additions; shared docs/auto files do not indicate one underlying change." + "reason": "Molmo and CTSM are unrelated model additions; shared repo wiring does not make them duplicates." }, { "left": "pull_request:43448", "right": "pull_request:46705", "accept": false, - "reason": "Molmo model support and distributed configuration are unrelated changes." + "reason": "Model addition vs distributed configuration work are unrelated changes." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 vs CTSM are distinct model integrations with no concrete shared bug or code path." + "reason": "Molmo2 and CTSM are different models in different domains, so they are not the same change." }, { "left": "pull_request:43451", "right": "pull_request:46705", "accept": false, - "reason": "Molmo2 model support is unrelated to distributed configuration work." + "reason": "Molmo2 model support and distributed config work do not touch the same underlying code path." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "CTSM and Step3p7 are different model additions; overlap in auto-mapping/docs is too broad to treat as duplicate." + "reason": "CTSM time-series model addition and Step 3.7 Flash multimodal model addition are unrelated." }, { "left": "pull_request:45490", "right": "pull_request:46705", "accept": false, - "reason": "CTSM model addition vs distributed config are unrelated changes." + "reason": "CTSM model support and distributed config changes are different features with no shared concrete fix." }, { "left": "pull_request:46562", "right": "pull_request:46658", - "accept": false, - "reason": "Both reference Step 3.7 Flash, but the file sets and module names differ substantially (step3p7 vs step_3_7_flash, plus mini-model/scripts). This looks like adjacent work, not the same concrete PR." + "accept": true, + "reason": "Both PRs appear to implement the same Step 3.7 Flash model addition; 46658 reads like a WIP/mini-model companion for the same underlying feature and could plausibly be merged into one PR." } ] }, @@ -20608,7 +20605,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20647,136 +20644,142 @@ "pull_request:46579|pull_request:46719", "pull_request:43665|pull_request:46719" ], - "prepared_review_unit_hash": "684f76bf47f613046573c5b5434d5df7247a8c13e78cfa0a64588e082b7df0b2", + "prepared_review_unit_hash": "2231cae11425457db498de4bf865f651bdba33602cb9436e12ebb84b67394f21", "component_budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 17, - "serialized_chars": 26388, - "estimated_input_tokens": 6597, - "estimated_eval_tokens": 13450 + "serialized_chars": 26394, + "estimated_input_tokens": 6599, + "estimated_eval_tokens": 13454 }, "budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 17, - "serialized_chars": 26388, - "estimated_input_tokens": 6597, - "estimated_eval_tokens": 13450 + "serialized_chars": 26394, + "estimated_input_tokens": 6599, + "estimated_eval_tokens": 13454 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This is a loose similarity cluster, not a duplicate set: the PRs cover unrelated changes in Whisper decoding, backward-compatibility shims, placeholder test edits, dequantization, doctest cleanup, generation validation, CI memory limits, typo fixes, and sparse-attention reshaping. I would not merge any of the soft-edge pairs as duplicates.", - "confidence": 0.94, + "summary": "This cluster is mostly heterogeneous: a few weakly similar maintenance/test-fix PRs appear alongside unrelated generation, CI, FP8, sparse-attention, and typo-cleanup changes. I don\u2019t see any pair that clearly fixes the same underlying bug or could realistically be merged as one PR.", + "confidence": 0.34, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR stands out as a duplicate target because the cluster spans multiple unrelated bugfixes and maintenance changes. If forced to pick a representative standalone fix, PR #44815 is the most concrete code-path bugfix, but it is not a duplicate of the others.", + "canonical_pr_reason": "PR 46579 is the cleanest single-theme representative in the set: a broad but coherent typo/cleanup sweep across code, with a clear purpose and no sign of mixed bug classes.", "best_issue_reason": null, - "best_pr_reason": "PR #44815 is the strongest standalone bugfix in the set: it has a specific dequantization problem, touches the relevant loader/integration code, and includes a focused test surface. It is still not equivalent to the other PRs.", + "best_pr_reason": "PR 46579 is the best standalone representative because it is internally consistent and better described than the others, which are either ambiguous or clearly about different subsystems/problems.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper decoding/tokenization fix vs. dummy test-assert edits; different bugs and code paths." + "reason": "Both are test-related, but one is Whisper decoding/test fixes and the other is CLIP/ViT test edits; they do not look like the same underlying bug or change." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decode behavior vs. FP8 dequantization; unrelated subsystems." + "reason": "Whisper tokenizer/test changes and a dequantization fix in FP8/model loading are unrelated code paths." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper test/code fix vs. generation watermarking validation; not the same issue." + "reason": "These touch different areas: Whisper decoding vs generation/watermarking validation and many model tests." }, { "left": "pull_request:43498", "right": "pull_request:46579", "accept": false, - "reason": "Backward-compatibility alias for tie_weights vs. broad typo cleanup; different changes." + "reason": "Backward-compatibility for tie_weights is unrelated to a broad typo cleanup sweep." }, { "left": "pull_request:43498", "right": "pull_request:46719", "accept": false, - "reason": "tie_weights compatibility shim vs. sparse-attention/model fixes; unrelated code paths." + "reason": "Different problems: a compatibility alias for tie_weights versus sparse-attention/minimax model logic changes." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Placeholder test failures vs. CI memory/device-map integration fix; not duplicates." + "reason": "Test-only edits in CLIP/ViT do not match the CI memory/device-map integration fix in Phimoe tests and workflows." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Test-only asserts vs. dequantization bugfix; no shared underlying defect." + "reason": "A set of test edits is not the same as a dequant/fp8/model-loading bug fix." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "Test noise vs. generation config validation fix; different problems." + "reason": "These are not the same underlying change: random model test edits versus generation config/Whisper validation fixes." }, { "left": "pull_request:43665", "right": "pull_request:46579", "accept": false, - "reason": "Test-only edits vs. typo cleanup across code; not the same change." + "reason": "Model test edits are unrelated to repository-wide typo cleanup changes." }, { "left": "pull_request:43665", "right": "pull_request:46719", "accept": false, - "reason": "Dummy test changes vs. attention/model fix; unrelated." + "reason": "No shared concrete bug: broken-looking test edits vs sparse attention/minimax model code." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "FP8 dequantization fix vs. CI resource-limit workaround; different scope and code paths." + "reason": "FP8/dequantization and Phimoe CI test infrastructure are different subsystems and failure modes." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Dequantization bugfix vs. generation watermarking validation; not merge-equivalent." + "reason": "These fix different code paths: dequant/model-loading versus generation validation and model tests." }, { "left": "pull_request:44815", "right": "pull_request:46579", "accept": false, - "reason": "Concrete FP8 fix vs. typo cleanup across many files; not the same underlying issue." + "reason": "A numeric/dequantization fix is not the same as typo cleanup across many files." }, { "left": "pull_request:45114", "right": "pull_request:46579", "accept": false, - "reason": "Doctest/documentation cleanup vs. code typo cleanup; similar theme but not the same artifact." + "reason": "Doctest documentation cleanup and code typo fixes are only broadly similar, not the same PR-worthy bug/change." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation config watermarking validation vs. Phimoe CI test infrastructure; unrelated bugs." + "reason": "Generation validation/Whisper fixes do not match the Phimoe CI resource-limit workaround." }, { "left": "pull_request:46380", "right": "pull_request:46579", "accept": false, - "reason": "Specific generation validation fix vs. broad typo cleanup; not duplicates." + "reason": "These are unrelated: generation config behavior versus typo corrections in assorted code files." }, { "left": "pull_request:46579", "right": "pull_request:46719", "accept": false, - "reason": "Generic typo cleanup across code vs. sparse-attention/model fix; different underlying changes." + "reason": "Typo cleanup across code is not the same underlying change as sparse-attention and Minimax model fixes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46719", + "accept": false, + "reason": "Test-file edits and sparse-attention/model implementation changes are different bugs and should not be merged together." } ] }, @@ -20790,7 +20793,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20806,7 +20809,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "d8ca86d272c3bfeb944b7a0da603e27331afbebe0a433cf35c94b7554b76a4c8", + "prepared_review_unit_hash": "b7fed5af45d39e1e229019f3d0b9af9855f78f86bafa24dfc5d3f948ec2f79c3", "component_budget": { "node_count": 2, "item_count": 2, @@ -20828,18 +20831,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are not duplicates: one changes packaging/versioning and a Vit test for a repo-bot check, while the other rewires CI/workflow behavior and related utilities. The only overlap is an incidental shared test file, not the same underlying change.", - "confidence": 0.99, + "summary": "These two open PRs are both marked for test/bot experimentation, but they change different parts of the repo: one is a setup/dependency/version tweak, the other is a CI/workflow and testing-utility change. Their only overlap is a shared Vit test file, which is not enough to treat them as the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs target different concrete code paths and should not be clustered as duplicates.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Neither PR is a strong global duplicate representative because they solve unrelated problems; if forced, 43532 is broader, but it is not the same change as 43488.", + "best_pr_reason": "PR #43532 is the more representative item if one must be chosen because it spans the CI/testing harness surface across multiple files; #43488 is a narrower setup/dependency/test tweak and does not look like the same change.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Different edits and goals: 43488 is about setup/version/dependency bot-check behavior, while 43532 modifies CI workflows and diff-display-related utilities. Shared Vit test file is incidental." + "reason": "Reject: the overlap is only a shared Vit test file. #43488 changes setup.py/dependency metadata and a test fixture, while #43532 changes GitHub workflows, testing utilities, and related notification code. They are not the same underlying bug or fix and would not plausibly merge into one PR." } ] }, @@ -20853,7 +20856,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20869,7 +20872,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "6b4ed046949d36b046939050c8530b08a37e346a0220c75cb240301aee0b50af", + "prepared_review_unit_hash": "fd4b8ef3ceb337e9eb5d4cdd4c74b87e4855ee8156d18cae4b3380d3e5999710", "component_budget": { "node_count": 2, "item_count": 2, @@ -20891,18 +20894,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add similarly structured pipeline/documentation scaffolding, but they target different features: promptable concept segmentation vs promptable visual segmentation. They share infrastructure files, yet they are not the same underlying change.", - "confidence": 0.93, + "summary": "Both PRs add new segmentation pipelines and share scaffolding, but they target different features: promptable concept segmentation vs promptable visual segmentation. The overlap is mostly documentation and framework registration, not the same concrete change set.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs implement different segmentation pipelines with different task definitions and model/test coverage.", + "canonical_pr_reason": "No canonical PR: these are distinct pipeline additions with different underlying model/support integration and different task semantics.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good duplicate-canonical candidate because they are distinct feature additions rather than one underlying fix/change.", + "best_pr_reason": "If a representative PR is needed, 43613 is the broader implementation with more model/test integration; however, it is still not a duplicate of 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different concrete feature additions: one adds Promptable Concept Segmentation, the other Promptable Visual Segmentation. Shared docs/pipeline scaffolding is not enough to treat them as the same change." + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registration scaffolding is not enough to treat them as the same change." } ] }, @@ -20916,7 +20919,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20932,7 +20935,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "6d8e8bd9b74b118f9b7c130f0671c260e964f4cf62d462923e7b3fb7127824ed", + "prepared_review_unit_hash": "6a102a949fcab3c1aa7219fe4a021eb28daa3660f7acb6702ffdbc9c7d9972f0", "component_budget": { "node_count": 2, "item_count": 2, @@ -20954,18 +20957,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both trainer.py changes, but they address different code paths: one fixes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT. They are related in theme but not duplicates.", - "confidence": 0.96, + "summary": "These two PRs are both in `trainer.py` but they address different extension points: one changes gradient-accumulation loss scaling, the other factors out signature-based column filtering. They are not the same change and should not be clustered as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43651 is the stronger representative because it makes a concrete behavior fix in the training loss-scaling path; PR 43663 is a separate extensibility/refactor change around input column filtering.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 43651 is the best single PR to surface for this cluster since it targets a specific user-visible bug with a narrowly scoped fix. PR 43663 is unrelated enough that it should not be merged with it for triage purposes.", + "best_pr_reason": "PR 43651 is the stronger representative of the cluster only because it makes a concrete runtime behavior change in `training_step`; however, it is not a duplicate of 43663, which adds a separate customization hook for column filtering.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both touch trainer.py and custom-trainer behavior, but they fix different underlying problems: gradient accumulation loss scaling vs. signature-column filtering. They are not the same concrete change and would not plausibly be merged into one PR." + "reason": "Different code paths and behaviors: 43651 changes loss scaling during training, while 43663 refactors model signature inspection for dataset column filtering. They are not the same underlying fix and would not be merged as one PR." } ] }, @@ -20979,7 +20982,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20995,7 +20998,7 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "e8a92d58421c596bd7de40f2c0dd65d8cbcbdbe4345dcf61c58ae3212be834c5", + "prepared_review_unit_hash": "2f2fc1bf7bf5366a35e51061ce23bc3653e26c64dfba490238e5957b5d3304d1", "component_budget": { "node_count": 2, "item_count": 2, @@ -21017,18 +21020,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs touch `src/transformers/cli/serve.py` and address TypeAdapter-related import-time typing issues, but they are not the same patch: one is a broad `from __future__ import annotations` cleanup, the other is a narrow `pydantic`-availability fix.", - "confidence": 0.8, + "summary": "Two open PRs touch the same CLI file and both fix annotation/import-time NameError risks, but they target different runtime failures. They should stay separate; no duplicate merge is warranted.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43656 is the broader CLI typing fix and directly targets the reported TypeAdapter NameError in the transformers CLI.", + "canonical_pr_reason": "PR #43842 is the most direct match to the TypeAdapter/pydantic NameError: it moves the TypeAdapter import under TYPE_CHECKING and quotes the annotation in `_validate_request`.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the more focused fix for the concrete missing-pydantic NameError path, so it is the stronger standalone patch.", + "best_pr_reason": "PR #43842 is the best representative because it is narrowly scoped to the concrete TypeAdapter NameError path, whereas #43656 is a broader annotation/postponed-evaluation cleanup affecting other symbols.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Same file and related symptom, but different concrete fixes: 43656 changes overall annotation handling with a future-import, while 43842 specifically guards `TypeAdapter` for the no-pydantic case. They look like adjacent fixes rather than the same underlying change." + "reason": "Both are annotation-related changes in `src/transformers/cli/serve.py`, but they fix different problems: #43656 addresses runtime NameErrors from unquoted model/processor type hints, while #43842 fixes `TypeAdapter` resolution when pydantic is absent. They are related, not the same concrete bug or mergeable duplicate." } ] }, @@ -21042,7 +21045,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21108,7 +21111,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -21220,343 +21223,385 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "0157009372dcd263b20827ff691e4606d95967f469734d808243b3d0ede36b52", + "prepared_review_unit_hash": "ed60f99e4255811baefba09b4014c66eca51e8803c436ae19412ebcad2532001", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84906, - "estimated_input_tokens": 21227, - "estimated_eval_tokens": 42710 + "serialized_chars": 84899, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75102, - "estimated_input_tokens": 18776, - "estimated_eval_tokens": 37808 + "serialized_chars": 75095, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly independent, model-specific output-tracing refactors. It contains at least two strong duplicate-like subgroups (GPT-J/CodeGen and GPT-Neo) and another plausible same-family subgroup around Flaubert/XLM; most other items still look separate, but the evidence is not strong enough to treat the rest as definitively unrelated.", - "confidence": 0.84, - "canonical_issue_reason": "No issue node was supplied in the items; the shared issue #43979 is only an external reference used by several PRs.", - "canonical_pr_reason": "PR 44066 is the strongest representative of the GPT-J/CodeGen subgroup: it has an explicit link to #43979 and overlaps the same output-tracing refactor path as 44085/44722.", - "best_issue_reason": "No issue node was supplied in the items; the shared issue #43979 is only an external reference used by several PRs.", - "best_pr_reason": "PR 44066 is the best overall representative of one of the clearest duplicate-like subgroups in the cluster, while the rest of the cluster appears to split into other model-specific refactors rather than one single duplicate set.", + "summary": "This cluster is mostly a batch of model-specific output-tracing refactors tied to the same umbrella effort (#43979), but most artifacts are separate model implementations and should not be merged as duplicates. Only a few pairs cover the same concrete model code path.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best anchor: it explicitly targets #43979 and captures the core GPT-J/CodeGen output-tracing migration in a representative, non-trivial patch.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the strongest representative of the cluster because it directly implements the standardized output-tracing refactor for the core GPT-J/CodeGen path and links the work to the tracking issue.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model code paths (CVT/FNet vs GPT-J); shared tracking issue alone is not enough." + "reason": "Different model families and code paths (CVT/FNet vs GPT-J), so this is not the same underlying change." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different models and implementations; both are output-tracing refactors but not the same change." + "reason": "CVT/FNet and DeBERTa v2 are unrelated implementations; same umbrella theme, but not the same bug/change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same code path; 44066 also includes CodeGen, so they could plausibly be merged into one PR." + "reason": "Both are GPT-J output-tracing refactors on the same code path; 44085 looks like a narrower variant of the same change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet and EfficientNet are separate model refactors, not the same underlying bug/change." + "reason": "RegNet and EfficientNet are separate model implementations; these are distinct refactors, not duplicates." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "Different models (EfficientNet vs GPT-J/CodeGen); only the theme is shared." + "reason": "EfficientNet output tracing and GPT-J/CodeGen output tracing are different model code paths." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "Unrelated model code paths; GPT-J/CodeGen refactor vs EfficientNet refactor." + "reason": "GPT-J/CodeGen versus EfficientNet; same tracing pattern, but not the same underlying code-path problem." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are unrelated model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different models; this is only a thematic similarity." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are separate model implementations, so these are not duplicates." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both target GPT-Neo output tracing on the same model path; they appear to be the same refactor expressed in slightly different patches." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J and GPT-Neo are different model code paths, even though both are output-tracing refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations, so not the same underlying change." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated implementations; same theme, not same bug/change." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are separate model paths, so these are distinct changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo versus GPT-J/CodeGen: same family of refactor, but different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are unrelated model implementations." }, { - "left": "pull_request:44013", - "right": "pull_request:44044", + "left": "pull_request:44018", + "right": "pull_request:44116", "accept": false, - "reason": "Different models (MobileNetV2 vs DeBERTa-v2), so not a duplicate." + "reason": "GPT-Neo and Flaubert are different code paths." }, { - "left": "pull_request:44066", - "right": "pull_request:44086", + "left": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "Different model families and code paths; same issue target is too broad to merge these." + "reason": "MPT and GPT-J/CodeGen are unrelated models." }, { - "left": "pull_request:44066", - "right": "pull_request:44071", + "left": "pull_request:44018", + "right": "pull_request:44085", "accept": false, - "reason": "GPT-J/CodeGen refactor vs MPT are separate refactors, not one code-path fix." + "reason": "GPT-Neo and GPT-J are different model code paths." }, { "left": "pull_request:44018", - "right": "pull_request:44068", - "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors on the same model and file set; they look like the same underlying change." + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated implementations." }, { - "left": "pull_request:44066", - "right": "pull_request:44068", + "left": "pull_request:44068", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations." + "reason": "GPT-Neo and MGP-STR are different model code paths." }, { - "left": "pull_request:44018", - "right": "pull_request:44066", + "left": "pull_request:44059", + "right": "pull_request:44086", "accept": false, - "reason": "Different models (GPT-Neo vs GPT-J/CodeGen)." + "reason": "GPT2 and MGP-STR are unrelated model implementations." }, { - "left": "pull_request:44019", - "right": "pull_request:44722", + "left": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "ResNet refactor vs GPT-J/CodeGen refactor; not the same change." + "reason": "GPT-J and MGP-STR are different code paths." }, { - "left": "pull_request:44086", - "right": "pull_request:44722", + "left": "pull_request:44098", + "right": "pull_request:44101", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are unrelated code paths." + "reason": "ViLT and XLM/Flaubert are separate model families; not the same change." }, { - "left": "pull_request:44019", - "right": "pull_request:44085", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "ResNet vs GPT-J/CodeGen refactor; same broad theme, but not the same bug/change." + "reason": "GPT2 and GPT-J/CodeGen are different model code paths." }, { "left": "pull_request:44018", - "right": "pull_request:44019", + "right": "pull_request:44071", "accept": false, - "reason": "Different models and output paths; not mergeable as one PR." + "reason": "GPT-Neo and MPT are unrelated implementations." }, { - "left": "pull_request:44019", - "right": "pull_request:44071", + "left": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "ResNet and MPT are separate model refactors." + "reason": "MPNet and LongT5 are different model code paths." }, { - "left": "pull_request:44068", - "right": "pull_request:44722", + "left": "pull_request:43997", + "right": "pull_request:44017", "accept": false, - "reason": "GPT-Neo vs GPT-J/CodeGen; similar refactor style, different code paths." + "reason": "RegNet and SegFormer are distinct vision model implementations; same broad area, but not the same bug/change." }, { - "left": "pull_request:44068", - "right": "pull_request:44116", + "left": "pull_request:44071", + "right": "pull_request:44161", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." + "reason": "MPT and LongT5 are unrelated model code paths." }, { - "left": "pull_request:44018", - "right": "pull_request:44116", + "left": "pull_request:43997", + "right": "pull_request:44056", "accept": false, - "reason": "GPT-Neo vs Flaubert; different underlying changes." + "reason": "RegNet and MPNet are separate model implementations." }, { "left": "pull_request:44071", - "right": "pull_request:44076", + "right": "pull_request:44086", "accept": false, - "reason": "MPT and ImageGPT are separate model changes." + "reason": "MPT and MGP-STR are unrelated models." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and Flaubert/XLM are different model code paths." + "reason": "TextNet and XLM/Flaubert are different code paths." }, { "left": "pull_request:43997", - "right": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet vs Segformer; only the refactor pattern matches, not the code path." + "reason": "MPNet and LongT5 are different model code paths." }, { "left": "pull_request:43997", - "right": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet and MPNet are different model refactors." + "reason": "RegNet and LongT5 are unrelated model implementations." }, { - "left": "pull_request:44056", + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are different code paths." + }, + { + "left": "pull_request:44018", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are different models; not a duplicate pair." + "reason": "GPT-Neo and LongT5 are unrelated models." }, { "left": "pull_request:43996", - "right": "pull_request:44007", + "right": "pull_request:44161", "accept": false, - "reason": "CVT/FNet and RegNet are separate model refactors." + "reason": "CVT/FNet and LongT5 are unrelated model implementations." }, { "left": "pull_request:43996", - "right": "pull_request:44129", + "right": "pull_request:44017", "accept": false, - "reason": "CVT/FNet and SpeechT5 are different code paths." + "reason": "CVT/FNet and SegFormer are different code paths." }, { "left": "pull_request:44010", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBERT and MPT are different model implementations." + "reason": "SqueezeBERT and MPT are unrelated model implementations." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBERT and VisualBERT share a BERT-family theme but not the same code-path fix." - }, - { - "left": "pull_request:44044", - "right": "pull_request:44074", - "accept": false, - "reason": "DeBERTa-v2 and TextNet are unrelated model changes." - }, - { - "left": "pull_request:44044", - "right": "pull_request:44101", - "accept": false, - "reason": "DeBERTa-v2 and Flaubert/XLM are distinct refactors." + "reason": "SqueezeBERT and VisualBERT are different model code paths." }, { - "left": "pull_request:44017", - "right": "pull_request:44161", + "left": "pull_request:44010", + "right": "pull_request:44018", "accept": false, - "reason": "Segformer and LongT5 are different models and fixes." + "reason": "SqueezeBERT and GPT-Neo are unrelated implementations." }, { - "left": "pull_request:43996", - "right": "pull_request:44007", + "left": "pull_request:44074", + "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet vs RegNet are separate model refactors." + "reason": "TextNet and SpeechT5 are different model code paths." }, { - "left": "pull_request:43996", - "right": "pull_request:44129", + "left": "pull_request:44073", + "right": "pull_request:44101", "accept": false, - "reason": "CVT/FNet and SpeechT5 are different code paths." + "reason": "VisualBERT and XLM/Flaubert are unrelated model implementations." }, { - "left": "pull_request:44010", - "right": "pull_request:44018", + "left": "pull_request:44056", + "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBERT and GPT-Neo are unrelated model changes." + "reason": "MPNet and MPT are separate model code paths." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBERT and DeBERTa-v2 are different model refactors." + "reason": "SqueezeBERT and DeBERTa v2 are unrelated models." }, { "left": "pull_request:44044", - "right": "pull_request:44154", + "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa-v2 and VITS are different models; same style of refactor is not enough." + "reason": "DeBERTa v2 and TextNet are different model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J/CodeGen output-tracing refactors on the same concrete code path, so they plausibly belong in one PR." + "reason": "Same GPT-J/CodeGen standardized output-tracing refactor on the same core code path; these look like duplicate versions of the same change." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both target GPT-J output tracing on the same file and change set style; they look mergeable into a single PR." + "reason": "Both are GPT-J output-tracing refactors on the same model path; 44722 is the broader version of the same change." }, { "left": "pull_request:44085", "right": "pull_request:44129", - "accept": true, - "reason": "Both are narrow output-tracing refactors on the same underlying decorator-style pattern and appear close enough to merge if the shared implementation were unified." + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model code paths." }, { "left": "pull_request:43996", "right": "pull_request:44007", - "accept": true, - "reason": "Both are model-specific output-tracing refactors with the same underlying pattern and could plausibly be combined for the same shared change family." + "accept": false, + "reason": "CVT/FNet and RegNet are distinct model implementations." }, { "left": "pull_request:43996", "right": "pull_request:44129", - "accept": true, - "reason": "Both are model-specific output-tracing refactors in the same standardized tracing effort and could plausibly be merged together." + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated code paths." }, { "left": "pull_request:44044", "right": "pull_request:44129", - "accept": true, - "reason": "Both are output-tracing refactors in the same standardized interface effort and look close enough to be one change family." + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are different model implementations." }, { "left": "pull_request:44044", "right": "pull_request:44154", - "accept": true, - "reason": "Both are output-tracing refactors that remove explicit output-flag plumbing from the same family of model code." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44068", - "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors on the same model and file, so they look like the same underlying change." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44129", - "accept": true, - "reason": "Both are narrow output-tracing refactors on the same underlying decorator-style pattern and appear close enough to merge if the shared implementation were unified." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44129", - "accept": false, - "reason": "GPT-J/CodeGen and SpeechT5 are different model code paths, so this is not the same underlying change." - }, - { - "left": "pull_request:44017", - "right": "pull_request:44161", - "accept": false, - "reason": "Segformer and LongT5 are different models and fixes." - }, - { - "left": "pull_request:44098", - "right": "pull_request:44101", - "accept": false, - "reason": "ViLT and Flaubert/XLM are separate model refactors." - }, - { - "left": "pull_request:44056", - "right": "pull_request:44101", "accept": false, - "reason": "MPNet and Flaubert/XLM are not the same underlying change." + "reason": "DeBERTa v2 and ViTs are unrelated model code paths." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The packet supports the conservative parts of the review, but the summary overstates the evidence by claiming that only the GPT-J/CodeGen and GPT-Neo variants are duplicate-like. The cluster also contains another plausible same-family subgroup around Flaubert/XLM (e.g. 44101 and 44116), so that exclusivity claim is too strong. The pairwise accept verdicts that were marked true look reasonable, but the summary should be softened to avoid implying the rest are definitively separate beyond the evidence provided." + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly same umbrella work with separate model-specific code paths, and the only accepted pairs are the ones that clearly share the same model path and appear mergeable as duplicate/refinement PRs. No pair verdict appears overstated." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -21564,7 +21609,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 36, + "component_count": 35, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -21591,14 +21636,14 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "8efa79b64d0c583ca01a5db8ccb490dc00ec2aee7239fd65d402134d4224df6a", + "prepared_review_unit_hash": "28e62bb5b94574c104a3bfe1cd809f804c8bf8dcd74f8c59a6e421d2d43a4a9f", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84906, - "estimated_input_tokens": 21227, - "estimated_eval_tokens": 42710 + "serialized_chars": 84899, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 }, "budget": { "node_count": 8, @@ -21613,48 +21658,48 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all thematically similar standardized output-tracing refactors, but they touch different model code paths and are not duplicates of the same change. The soft edges should be rejected because they are model-specific implementations rather than one mergeable PR.", - "confidence": 0.95, + "summary": "All items are pull requests in the same broad Hugging Face refactor wave for standardized output tracing, but they touch different model code paths and should not be deduplicated as the same fix. No issue is present in the cluster. The best representative PR is SpeechT5 because it has the most concrete scoped change and an explicit issue link.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the clearest anchor: it has an explicit issue target, a concrete SpeechT5-specific refactor, and enough scope to represent the cluster\u2019s general pattern without pretending the other model-specific refactors are duplicates.", + "canonical_pr_reason": "PR 44129 is the strongest canonical representative: it has an explicit issue target, a clear refactor title, and a focused change to one model\u2019s output tracing implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the best standalone representative because it is well-scoped, explicitly tied to an issue, and clearly documents the output-tracing migration within one model.", + "best_pr_reason": "PR 44129 is the best overall representative of the cluster because it is concrete, well-scoped, and clearly tied to a tracked issue; the other PRs are similar but less anchored.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are output-tracing migrations, but FocalNet and ImageGPT are unrelated model implementations with different forward paths and hooks; they are not the same underlying change." + "reason": "Both are standardized-output refactors, but they apply to different model files and different execution paths (FocalNet vs ImageGPT), so they are not the same underlying change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet have different architectures and different tracing/output mechanics; these are parallel refactors, not one duplicate bug or one mergeable PR." + "reason": "These touch VisualBert and TextNet respectively; despite similar decorator migration, they are separate model-specific refactors and not one mergeable fix." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert are distinct model code paths; the similarity is only the shared decorator/refactor pattern, not the same concrete change." + "reason": "MPNet and VisualBert have distinct tracing implementations and model internals; this is a shared refactor theme, not the same bug or change." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "MPNet and TextNet are different models with different output-capture wiring; they are not duplicates of the same fix." + "reason": "MPNet and TextNet are unrelated code paths; the similarity is only the output-capture migration pattern." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated architectures; despite similar output-tracing edits, they are separate model-specific refactors." + "reason": "RegNet and SqueezeBert are separate model implementations, so these are parallel refactors rather than duplicates of one concrete fix." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 change different model internals and output signatures; they share only the refactoring theme, not the same underlying code-path problem." + "reason": "SqueezeBert and SpeechT5 each change their own output tracing logic; they are not the same underlying code-path problem and would not plausibly be merged into one PR." } ] }, @@ -21668,7 +21713,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21734,274 +21779,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "be7046f55a795044a1edf09d22c7cee333a015a32cca709fb36198db2118635a", + "prepared_review_unit_hash": "0d24dfe595ab6f3ac4550a79accf0e97a4fc4d2615ded0e990cb9f818ede3f0e", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40177, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40177, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All items are related by a common refactor theme around output tracing, but they touch different model implementations and code paths, so they are not duplicates of each other.", - "confidence": 0.9, + "summary": "This is a cluster of similar-looking output-tracing refactors across many different model implementations, not a set of duplicate fixes for one shared bug or code path.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue artifact is present; as a PR anchor, #44003 is the strongest representative because it has the broadest concrete code change and adds explicit output-capturing plumbing across related Mamba-family modules.", + "canonical_pr_reason": "PR 44003 is the strongest representative of the cluster: it is a broader two-file Mamba/FalconMamba refactor and includes the shared output-capture machinery, so it best captures the overall theme.", "best_issue_reason": null, - "best_pr_reason": "#44003 is the best representative PR for the cluster: it is a substantive output-tracing refactor, spans multiple related files, and introduces the supporting capture/output-recording machinery rather than only a narrow signature cleanup.", + "best_pr_reason": "PR 44003 is the best single representative because it is the most substantial and central example of the output-tracing refactor pattern in this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths (timm_backbone vs upernet); same refactor theme only." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different models (timm_backbone vs vision_text_dual_encoder); not the same underlying change." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different implementations and architectures; both only refactor output tracing." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "timm_backbone and Mamba-family models are unrelated code paths; not one concrete fix." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model families and forward paths; not a duplicate change." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 vs UnivNet are distinct models; shared pattern is too broad." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model implementations and outputs; same refactor style only." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "MobileNetV1 and Depth Anything are separate model paths with no common bug." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different models; output-tracing cleanup is not the same concrete change." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 vs speech encoder-decoder are unrelated code paths." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-text dual encoder and UnivNet are unrelated model implementations." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different models/subsystems; both only adjust output-tracing behavior." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different architectures and code paths; no shared concrete bug." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families; same refactor theme is too broad." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are separate implementations; not one underlying fix." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Mamba-family output plumbing vs UnivNet tracing cleanup are distinct changes." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different models and forward signatures; not the same bug or patch." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are unrelated model paths; same refactor theme only." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different architectures and files; not a single mergeable fix." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Mamba/Falcon-Mamba output tracing vs SuperPoint are unrelated concrete changes." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba-family output recording and CodeGen cache/output refactor are different code paths." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "timm_backbone and CodeGen are unrelated models; shared wording is not enough." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet vs CodeGen are different architectures and forward logic." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen do not share the same underlying change." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Vision-text dual encoder and CodeGen are unrelated code paths." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and VisionEncoderDecoder are separate model families; not the same concrete fix." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different model implementations and output signatures; not duplicates." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything vs SuperPoint are unrelated models." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different architectures (depth estimation vs RWKV); same refactor theme only." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything and DPR are distinct model paths." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44029", "accept": false, - "reason": "Vision and speech encoder-decoder wrappers are separate implementations, not one bug." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44027", "accept": false, - "reason": "Different model families and code paths; not a duplicate PR." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44028", "accept": false, - "reason": "VisionEncoderDecoder and RWKV are unrelated concrete changes." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different models and forward behavior; not the same fix." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech encoder-decoder and SuperPoint are unrelated implementations." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different model families; same refactor pattern is insufficient." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Speech encoder-decoder vs DPR are separate code paths." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV have no shared concrete bug or patch." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different model implementations; not mergeable as one PR." + "reason": "Different model families and code paths; only the refactor pattern is shared." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are unrelated model paths; only the refactor theme matches." + "reason": "Different model families and code paths; only the refactor pattern is shared." } ] }, @@ -22015,7 +22060,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22031,12 +22076,12 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "0d0363d53a2391bd540362812656373d83317812622d56061c41a01349c96e93", + "prepared_review_unit_hash": "5939ab1f80e8a160ad19dcefa2bf471a31d686ec2c91dd75f3410b0192091d81", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6552, + "serialized_chars": 6551, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -22044,7 +22089,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6552, + "serialized_chars": 6551, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -22053,18 +22098,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate model-addition PRs for different models (PenguinVL vs Voxtral TTS), so they should not be clustered as duplicates.", - "confidence": 0.98, + "summary": "These are two different model-addition PRs that only overlap on shared Transformers plumbing and docs/auto-registry files. They target different models and code paths, so they should not be merged as duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No clear duplicate pair here; if forced to pick a representative, PR #45401 is the more focused and better-scoped change because it has an explicit issue target and is centered on one concrete TTS feature.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR #45401 is the better standalone representative for the cluster because it is narrowly tied to Voxtral TTS support and includes an explicit tracking issue, whereas PR #44662 is a broader PenguinVL model implementation.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different underlying changes: one adds PenguinVL, the other adds Voxtral TTS. Shared auto/docs scaffolding is expected for new model integrations, not evidence of the same fix or feature." + "reason": "Different model implementations (PenguinVL vs Voxtral TTS) with different filenames, model classes, and functionality. Shared auto/config/docs files are expected for any new model integration and do not indicate the same underlying change." } ] }, @@ -22078,7 +22123,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22094,7 +22139,7 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "e3bab5f65f7a67b2c1e484f5e154dca33ee4d81d0f239b2a4bfcbed8677845b1", + "prepared_review_unit_hash": "3ab961925b038cad1e75aa3374c84841413c7e15ab0900818122a6f3fe41eac1", "component_budget": { "node_count": 2, "item_count": 2, @@ -22116,18 +22161,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related kernel-fusion feature PRs, but they target different layers of the design: one introduces a generic Module Fusion API, the other implements n-to-1 kernel fusion via KernelConfig and integration hooks. They are not the same concrete change and should not be merged as duplicates.", - "confidence": 0.85, + "summary": "These PRs are related thematically around module/kernel fusion, but they are not the same change: #44979 introduces a new `module_fusion.py` API scaffold, while #45363 wires `KernelConfig` into model loading and adds integration/example code for n-to-1 kernel fusion. They should not be deduplicated.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45363 is the more concrete end-user feature PR: it wires KernelConfig into modeling/integration code and adds a runnable example, whereas #44979 is a foundational API layer with no direct overlap in filenames or specific code path.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45363 is the best standalone candidate because it has a clear behavior change, touches the actual runtime paths, and is more directly actionable for review/merge.", + "best_pr_reason": "#45363 is the better cluster representative because it is the more concrete end-user feature: it spans loader/integration code, `KernelConfig`, and a usage example, whereas #44979 is mainly the lower-level module fusion API groundwork.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Different changes: #44979 defines a generic module fusion API in a new module, while #45363 adds n-to-1 kernel fusion plumbing through KernelConfig and integration/model loading paths. Related conceptually, but not the same underlying bug or mergeable PR." + "reason": "Related fusion work, but different concrete changes and code paths: one adds a module-fusion API module, the other adds kernel-config-based n-to-1 fusion integration. They are not mergeable as one PR without conflating separate features." } ] }, @@ -22141,7 +22186,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22157,7 +22202,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "615c0eaa8faa46864d32b5b34c81cc3d9f70f37e54f1f701b540b8cb326ecd4d", + "prepared_review_unit_hash": "81dcc477885e4f5405e1199838c536737f2a35baf3b13306b4e15186151ed7d9", "component_budget": { "node_count": 2, "item_count": 2, @@ -22179,18 +22224,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same Mistral regex AttributeError in `_patch_mistral_regex` on the same code path, with the same core one-line fix; 45317 is the more complete version because it also adds tests.", - "confidence": 0.97, + "summary": "Both pull requests fix the same Mistral regex AttributeError in the same code path, with 45317 appearing to be the more complete version because it adds tests. They are strong duplicates and can be treated as one change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same fix as 45086 plus test coverage, making it the more complete and merge-ready change.", + "canonical_pr_reason": "PR 45317 is the better canonical PR because it implements the same fix as 45086 and also adds test coverage, making it the more complete patch.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 best represents the cluster because it fixes the bug on the same file and includes validation via tests, whereas 45086 is only the minimal code change.", + "best_pr_reason": "PR 45317 is the best representative for the cluster since it contains the same bug fix plus tests, while 45086 is the narrower earlier version.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in `_patch_mistral_regex`, same target issue, and the diffs overlap on the exact code-path change; 45317 simply extends the patch with tests." + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same file and code path, with 45317 extending 45086 by adding tests." } ] }, @@ -22204,7 +22249,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22224,12 +22269,12 @@ "pull_request:45569|pull_request:46565", "pull_request:46332|pull_request:46565" ], - "prepared_review_unit_hash": "9c21b812fd448ba6b982fe8201c2289dd09e8763682d56a82e21a3c1c2ed79ba", + "prepared_review_unit_hash": "7192bd8d301fc5cd72a77703e7ae5546e6dd48d26587f30f68544af8318bbca8", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 13739, + "serialized_chars": 13738, "estimated_input_tokens": 3435, "estimated_eval_tokens": 7126 }, @@ -22237,7 +22282,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 13739, + "serialized_chars": 13738, "estimated_input_tokens": 3435, "estimated_eval_tokens": 7126 }, @@ -22246,30 +22291,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related by the Nemotron naming family, but they are not duplicates: one is a NemotronH architecture/refactor PR, one splits NemotronH into dense/sparse variants, and two are separate ASR streaming additions for different model generations.", - "confidence": 0.92, + "summary": "These PRs are all Nemotron-related but split across different model families and feature tracks: NemotronH refactors versus Nemotron ASR streaming. The overlaps are broad thematic similarity, not duplicate code-path changes.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the cluster mixes distinct changes for NemotronH core refactoring/splitting and separate Nemotron ASR streaming support, so none clearly subsumes the others.", + "canonical_pr_reason": "PR 45569 is the most central NemotronH architecture change and most likely subsumes the narrower 45115 refactor, making it the best representative PR in this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the broadest representative of the NemotronH model work, but it is still not a duplicate of the ASR streaming PRs.", + "best_pr_reason": "PR 45569 has the clearest, broadest scope within the NemotronH line and is the strongest anchor for this cluster; the ASR PRs are separate model additions rather than the same change.", "soft_edge_verdicts": [ { - "left": "pull_request:45115", - "right": "pull_request:45569", + "left": "Refactor/nemotron h inherit granitemoehybrid", + "right": "Proper nemotron H and 3 and 2", "accept": false, - "reason": "Related NemotronH work, but different concrete changes: 45115 is a conversion-mapping and inheritance refactor, while 45569 splits NemotronH into dense/sparse variants and adds docs/tests. Not the same underlying fix." + "reason": "Related model family, but 45115 is a narrow inheritance refactor while 45569 is a broader split/cleanup across NemotronH dense/sparse models; not the same concrete change." }, { - "left": "pull_request:45569", - "right": "pull_request:46565", + "left": "Proper nemotron H and 3 and 2", + "right": "Add Nemotron 3.5 ASR Streaming", "accept": false, - "reason": "Both touch auto/model plumbing, but they target different models and features: NemotronH core refactor vs Nemotron 3.5 ASR streaming. Shared files are broad infrastructure, not the same bug/change." + "reason": "Both touch Nemotron naming and auto/docs plumbing, but one is NemotronH architecture work and the other is ASR streaming; different model families and code paths." }, { - "left": "pull_request:46332", - "right": "pull_request:46565", + "left": "Add Nemotron ASR Streaming", + "right": "Add Nemotron 3.5 ASR Streaming", "accept": false, - "reason": "Both are ASR-streaming related, but for different model families/generations (Nemotron ASR vs Nemotron 3.5 ASR). Same subsystem, different concrete code paths and releases." + "reason": "Same ASR area, but these are separate model-version additions, not one shared bug or a single change that could plausibly be merged as one PR." } ] }, @@ -22283,7 +22328,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22299,7 +22344,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "d6a3f5be27f57c9041d6caa5497c1fe1cf11eabfa65fde870c4aa08883451cb6", + "prepared_review_unit_hash": "3b1168a35d2cdffd5711712378df7332a21cbc4da2e396294ca7c254924c7ace", "component_budget": { "node_count": 2, "item_count": 2, @@ -22321,18 +22366,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related pieces of the heterogeneous-model initiative, but they are not duplicates: #45333 adds per-layer heterogeneous config plumbing, while #45332 adds broader heterogeneous modeling/runtime support that consumes that config machinery.", + "summary": "These are related PRs in the same heterogeneous-config initiative, but they are not duplicates: one adds the per-layer config plumbing, the other consumes it to add broader modeling/runtime support.", "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "#45332 is the broader umbrella change and the more complete representative of the heterogeneous-model support work; #45333 is a narrower config-only slice.", + "canonical_pr_reason": "PR #45333 is the narrower foundational change: it introduces per-layer configuration support in `configuration_utils` and the heterogeneity helpers, which the larger modeling PR appears to build on.", "best_issue_reason": null, - "best_pr_reason": "#45332 best represents the cluster because it covers the wider end-to-end implementation, not just the configuration layer.", + "best_pr_reason": "PR #45333 is the better representative of the cluster because it is the core abstraction change and is more self-contained than the downstream modeling integration in #45332.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Related area, but not the same underlying change: #45333 is focused on per-layer config support, while #45332 adds broader runtime/modeling changes and cache handling. They would not plausibly collapse into one duplicate PR." + "reason": "Same feature area, but not the same concrete change. #45333 adds per-layer config infrastructure; #45332 adds model/cache/masking behavior that uses that infrastructure. They could be merged separately and are not duplicate fixes." } ] }, @@ -22346,7 +22391,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22362,12 +22407,12 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "0de48ea6d8fa215bac09f56216e4d1f70bf1e501b9a24940042093405a8d6516", + "prepared_review_unit_hash": "602bcd3606ceb09dbb5cb2e2b280822fff4853137cbe6131249cae39a9c89d50", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5513, + "serialized_chars": 5514, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -22375,7 +22420,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5513, + "serialized_chars": 5514, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -22384,18 +22429,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related typing-oriented PRs, but they are not the same change: one is a broad type-checking pass over many files, while the other adds specific typing aliases/protocols to satisfy particular model/integration sites.", - "confidence": 0.88, + "summary": "These are both typing-related PRs and share a few files, but they address different problems: #45415 is a broad repository-wide type-checking cleanup, while #45425 adds concrete typing protocols/aliases to support specific integrations and model utility code. They should not be merged as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No clear canonical duplicate PR: 45415 is a sweeping type-checking cleanup, while 45425 is a narrower set of typing compatibility fixes and helper protocols.", + "canonical_pr_reason": "PR #45425 is the better canonical PR for this cluster because it has a focused, concrete typing change set centered on `_typing.py`, `modeling_utils.py`, and PEFT-related integration code, rather than a broad sweep of type-checking adjustments.", "best_issue_reason": null, - "best_pr_reason": "PR 45425 is the better representative of a concrete code-change because it targets specific typing gaps in core helpers, whereas 45415 is a broad repository-wide type-checking effort.", + "best_pr_reason": "PR #45425 is the best representative because it targets a specific compatibility/typing gap with clear implementation scope, whereas #45415 is a much wider type-checking pass across many files and is not the same underlying change.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-related and overlap in a few files, but they address different scopes: 45415 is a broad static-type cleanup across many modules, while 45425 adds specific type aliases/protocols for particular APIs. They would not naturally merge into one single duplicate fix." + "reason": "Shared filenames and general typing theme are not enough: #45415 is a broad `src/transformers/*py` type-checking enablement pass, while #45425 introduces specific typing aliases/protocols for modeling_utils/PEFT integration. They fix different code-paths and are not plausibly one merged PR." } ] }, @@ -22409,7 +22454,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22425,7 +22470,7 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "a8404d4418e04730ddaec4bbeb10bd05e80dceaa685e05f0670d69ef8e14aaba", + "prepared_review_unit_hash": "437e784edbdaa3f47d75e1c4503c42fbeb34f701c30f420fd7fef323ab63e2d8", "component_budget": { "node_count": 2, "item_count": 2, @@ -22447,18 +22492,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one fixes Qwen3.5/Qwen3_next/Olmo hybrid cached multi-token forward behavior, while the other fixes Mamba2 cache handling for seq_len > 1. They touch different model families and different code paths.", - "confidence": 0.98, + "summary": "These are two unrelated open PRs that both fix cached-forward bugs, but in different model families and with different code paths: Qwen3.5/OLMo hybrid linear-attention cached multi-token forward vs. Mamba2 use_cache with seq_len>1. They should not be clustered as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR candidate: the PRs address unrelated bugs in different model implementations and are not mergeable into one change.", + "canonical_pr_reason": "PR #46084 is the cleaner canonical PR: it has an explicit issue target, a narrowly scoped fix for a concrete seq_len>1 cache bug, and a focused test update.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good global representative of the other; the fixes are separate and should be triaged independently.", + "best_pr_reason": "PR #46084 is the best representative of the cluster because it is the most self-contained and explicitly tied to a tracked bug, whereas #45513 spans multiple model files and a broader model-family-specific fix.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different model families and distinct cache/forward bugs; no shared concrete code-path issue." + "reason": "Different underlying bugs and different model code paths. #45513 targets Qwen3.5/OLMo hybrid linear-attention cached multi-token forward; #46084 targets Mamba2 cache handling when seq_len > 1. Shared theme is too broad to treat as the same change." } ] }, @@ -22472,7 +22517,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22488,7 +22533,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "02409383fcb01fe1ae50d5d594190a07c2d32541b7cdc00c0db32550ad4077ca", + "prepared_review_unit_hash": "68bc090dd8ee2951d321d3065f29a70e5a1db7331c01abb6d201c54e6f5b77a4", "component_budget": { "node_count": 2, "item_count": 2, @@ -22510,18 +22555,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs touch expert-parallel loading paths, but they fix different bugs: 45548 is about EP/DeepSpeed ZeRO-3 loading and config serialization during accelerate launch, while 45662 fixes FSDP2 expert tensors being overwritten by rank-0 broadcast/post-load handling.", - "confidence": 0.95, + "summary": "The two pull requests both touch EP-related loading code, but they fix different failure modes: one is about preserving distributed config / enabling EP-aware loading, the other is about preventing expert tensors from being overwritten during FSDP2/rank-0 broadcast. They are related, but not duplicates.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 45548 and 45662 address different concrete EP-loading failures and should not be deduplicated.", + "canonical_pr_reason": "PR #45662 is the better canonical candidate because it targets a concrete data-corruption bug in a specific loading path (experts being silently overwritten) and has the more self-contained, user-visible fix.", "best_issue_reason": null, - "best_pr_reason": "45662 is the more concrete standalone bug fix, but it targets a different code path than 45548 (FSDP2 overwrite vs. ZeRO-3/accelerate loading).", + "best_pr_reason": "PR #45662 is the strongest representative of the cluster topic: it fixes a specific EP/FSDP2 loading bug with a clear code-path and direct correctness impact, whereas #45548 is more about compatibility plumbing for accelerate/ZeRO-3.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Reject: only broad EP similarity and one shared file. 45548 fixes config/ZeRO-3 loading via accelerate launch; 45662 fixes a separate FSDP2 expert-overwrite path. Not the same underlying bug." + "reason": "Related EP loading changes, but they address different bugs and code paths: #45548 preserves distributed config / adds EP awareness, while #45662 prevents expert parameters from being overwritten during sharded loading. They would not plausibly merge as one PR." } ] }, @@ -22535,7 +22580,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22551,7 +22596,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "ad3176edfbbc204a9b63bbc587c5688eafda27dcfaab5c87cfdbc31a16cf1b60", + "prepared_review_unit_hash": "333577ab690e2f9b0f5dc24a7ebf48a4aa53109fe7a04d36172550526cc50503", "component_budget": { "node_count": 2, "item_count": 2, @@ -22573,18 +22618,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same xdist-related captured_info collision in testing_utils; PR 45645 appears to subsume the narrower path fix from 45639 and adds the CI-side log aggregation needed to preserve all debug logs.", - "confidence": 0.9, - "canonical_issue_reason": "Issue 45561 is the shared tracked bug for xdist-safe captured_info/debug log handling.", - "canonical_pr_reason": "PR 45645 is the stronger canonical PR because it includes the core worker-specific file fix plus the workflow changes needed to collect multiple captured_info files in CI.", - "best_issue_reason": "Issue 45561 is the only explicit issue target and matches the underlying bug described by both PRs.", - "best_pr_reason": "PR 45645 is the best representative because it covers the same fix as 45639 and extends it to the full end-to-end log preservation flow.", + "summary": "Both PRs address the same xdist-related captured_info collision in `testing_utils.py`; PR #45645 is the more complete version because it also updates the CI workflow to read multiple per-worker log files.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#45645 is the better canonical PR because it contains the core xdist-safe logging fix plus the downstream workflow change needed to consume the new per-worker artifacts.", + "best_issue_reason": null, + "best_pr_reason": "#45645 is the best representative of the cluster since it fixes the full concrete problem end-to-end, while #45639 only covers the writer side.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Accept: both PRs address the same concrete xdist captured_info collision in transformers/testing_utils; 45645 appears to include the same fix as 45639 plus the related CI log collection updates, so they could plausibly be merged into one PR." + "reason": "Same underlying bug: `captured_info.txt` collisions under pytest-xdist. #45645 appears to be a superset of #45639's fix and could plausibly be merged as one PR." } ] }, @@ -22598,70 +22643,7 @@ { "label": "pull_request", "component_index": 24, - "component_count": 36, - "review_unit_index": 1, - "review_unit_count": 1, - "status": "reviewed", - "reason": null, - "source": "llm", - "cache_hit": false, - "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-45670-2", - "nodes": [ - "pull_request:45670", - "pull_request:46320" - ], - "soft_pairs": [ - "pull_request:45670|pull_request:46320" - ], - "prepared_review_unit_hash": "80d613b52dba613c8e1fd8139954fe31d3d1369b8f91bc8c003100fa37c299ec", - "component_budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4376, - "estimated_input_tokens": 1094, - "estimated_eval_tokens": 2444 - }, - "budget": { - "node_count": 2, - "item_count": 2, - "soft_pair_count": 1, - "serialized_chars": 4376, - "estimated_input_tokens": 1094, - "estimated_eval_tokens": 2444 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "The two PRs are unrelated: one is a small AutoModel registry addition for glmasr/dia, while the other adds attention-implementation dispatch support for RecurrentGemma with corresponding model/test/doc changes.", - "confidence": 0.97, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46320 is the stronger standalone change: it fixes a concrete code path in RecurrentGemma and includes implementation plus tests/docs, whereas #45670 is a narrow model-mapping tweak.", - "best_issue_reason": null, - "best_pr_reason": "PR #46320 is the best representative item because it is the more substantive, self-contained change and has clearer user-facing impact.", - "soft_edge_verdicts": [ - { - "left": "pull_request:45670", - "right": "pull_request:46320", - "accept": false, - "reason": "Different targets and code paths: #45670 only updates AutoModel registry entries for glmasr/dia, while #46320 adds attn_implementation dispatch for RecurrentGemma. They are not the same underlying bug or change." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, - "error_kind": null, - "error_message": null - }, - { - "label": "pull_request", - "component_index": 25, - "component_count": 36, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22677,7 +22659,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "cff188121138774dd4fde37037f3be49bc1115f2e89eb1e4de6cd93e339383f3", + "prepared_review_unit_hash": "93f25b81be90a6bb698fcdb5aa39b292cdc2940497bd1de9995c2b93559c47aa", "component_budget": { "node_count": 2, "item_count": 2, @@ -22699,18 +22681,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in the same Transformers codebase, but they implement different linting-rule refactors: one adds @can_return_tuple to model forward methods (TRF003), the other changes _init_weights patterns for many models (TRF018). They are not the same underlying change.", - "confidence": 0.95, + "summary": "Two open PRs in the same repository touch some of the same model files, but they implement different linting rules and fix different code patterns. The overlap looks incidental rather than duplicate work.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45720 is the cleaner representative of this pair: it is a focused, concrete modeling-forward decorator change with a clear code-path scope.", + "canonical_pr_reason": "No single canonical PR: #45720 is about TRF003 (return-tuple/forward signature cleanup across model forwards), while #46259 is about TRF018 (model weight initialization refactor). They target different underlying changes.", "best_issue_reason": null, - "best_pr_reason": "PR #45720 is the best representative because its intent is narrowly defined and the diff is easier to characterize than the broader cross-model initialization refactor in #46259.", + "best_pr_reason": "Neither PR is a good duplicate target for the other. If forced to pick the more representative cluster member, #45720 is slightly broader and more central to the shared file overlap, but the pair is not a true duplicate set.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Reject: these are different linter-rule implementations affecting different code paths. #45720 adds return-tuple handling to forward methods, while #46259 updates weight initialization logic; they are not mergeable as one PR for the same bug/change." + "reason": "Different linter rules and different code paths: TRF003 changes forward return-tuple handling, TRF018 changes initialization logic. Shared model filenames do not indicate the same underlying fix." } ] }, @@ -22723,8 +22705,8 @@ }, { "label": "pull_request", - "component_index": 26, - "component_count": 36, + "component_index": 25, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22789,8 +22771,8 @@ }, { "label": "pull_request", - "component_index": 27, - "component_count": 36, + "component_index": 26, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22806,7 +22788,7 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "10280828d0062c2b43ee46136fd1a89268a08100fc337a58d7053d46e65c4d92", + "prepared_review_unit_hash": "c6d96aba025fd2fd8f43cfd5e55625391506506e68ae1aabb317289e7449bce6", "component_budget": { "node_count": 2, "item_count": 2, @@ -22828,18 +22810,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates: one is a broad experimental/tracing and model-infra change, the other is a test-migration PR with kernel/test additions. They do not target the same concrete bug or code path.", - "confidence": 0.95, + "summary": "The two PRs are not duplicates. PR 45932 is a mixed CI/telemetry and model-config change, while PR 46100 is a test migration adding kernel tests and related test infra. They share only broad testing/CI context, not the same concrete bug or change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR as a duplicate representative; the two PRs address different areas and should remain separate.", + "canonical_pr_reason": "No canonical PR as a duplicate pair; the changes target different code paths and different goals.", "best_issue_reason": null, - "best_pr_reason": "PR #46100 is the better standalone candidate because its scope is more focused and internally coherent, centered on test migration and related test files.", + "best_pr_reason": "PR 46100 is the better cluster representative because it is a more coherent, self-contained test migration/change set. PR 45932 is mostly unrelated CI/workflow and model configuration churn.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Different change intents and file sets: #45932 mixes workflow secrets, model implementation tweaks, and mapping/config changes; #46100 is a test-migration PR with kernel test additions. They do not look like the same underlying fix and would not plausibly be merged into one PR." + "reason": "Different underlying changes: 45932 updates CI/workflow secrets and several model configs, while 46100 adds kernel tests and test-migration files. They do not appear mergeable into one PR for the same concrete fix/change." } ] }, @@ -22852,8 +22834,8 @@ }, { "label": "pull_request", - "component_index": 28, - "component_count": 36, + "component_index": 27, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22869,40 +22851,40 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "113ae2aca6fd89a162c1ab6d6ccb06eb67107d683dedbd9dde20a0f5d10b2368", + "prepared_review_unit_hash": "5898c557ca3f5acf198a58842cd631e4cf16a9d595966c7679dbf537968c9d7e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5523, - "estimated_input_tokens": 1381, - "estimated_eval_tokens": 3018 + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5523, - "estimated_input_tokens": 1381, - "estimated_eval_tokens": 3018 + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs overlap in distributed/parallelism code, but they are not the same change: #46126 is a focused EP + Trainer integration for router/expert-parallel behavior, while #46269 is a broad distributed-branch base touching many unrelated areas including CI, FSDP, integrations, and multiple models.", - "confidence": 0.89, + "summary": "These are not duplicates: PR #46126 is a focused EP/Trainer integration change around tensor-parallel router handling, while PR #46269 is a broad distributed branch base touching many unrelated distributed/FSDP areas. They share some filenames, but not the same concrete fix or mergeable code path.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "#46126 is the more specific, concrete feature change centered on EP/router behavior and Trainer integration; #46269 is a broad umbrella branch with many unrelated distributed updates rather than the same single fix.", + "canonical_pr_reason": "PR #46126 is the more specific, self-contained change focused on EP router/tensor-parallel integration and trainer wiring; it has a clear single-purpose scope.", "best_issue_reason": null, - "best_pr_reason": "#46126 is the best representative of the cluster because it has a narrow, identifiable code-path change and a clear implementation target, whereas #46269 is too broad and mixed to serve as a clean duplicate anchor.", + "best_pr_reason": "PR #46126 is the best representative because it targets one concrete feature/fix, whereas PR #46269 is a large umbrella branch-base PR with mixed distributed changes.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Shared filenames only indicate the same subsystem. The PRs do not appear to fix the same concrete bug or implement the same change: #46126 adds EP/router masking and Trainer integration, while #46269 is a large distributed branch base spanning CI, FSDP, integrations, and multiple unrelated edits." + "reason": "Shared files alone are not enough here. PR #46126 is an EP/router-specific integration, while PR #46269 is a broad distributed branch base covering multiple unrelated subsystems (FSDP, CI, integrations, model loading). They do not appear to fix the same underlying bug or change, and would not plausibly merge as one PR." } ] }, @@ -22915,8 +22897,8 @@ }, { "label": "pull_request", - "component_index": 29, - "component_count": 36, + "component_index": 28, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -22968,208 +22950,208 @@ "pull_request:46504|pull_request:46563", "pull_request:46504|pull_request:46716" ], - "prepared_review_unit_hash": "19018fde7d8ffda2eaca6d99a911ef37392e64e4409b0ac1f85aebaceb3dc707", + "prepared_review_unit_hash": "005b295efe4efd27f43c7a72ac004245220cb69c99a1b2cdf6067be4a00b79c3", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 29, - "serialized_chars": 25929, - "estimated_input_tokens": 6483, - "estimated_eval_tokens": 13222 + "serialized_chars": 25928, + "estimated_input_tokens": 6482, + "estimated_eval_tokens": 13220 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 29, - "serialized_chars": 25929, - "estimated_input_tokens": 6483, - "estimated_eval_tokens": 13222 + "serialized_chars": 25928, + "estimated_input_tokens": 6482, + "estimated_eval_tokens": 13220 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate workflow maintenance PRs; despite some filename-based similarity, they address different CI/security tweaks and are not duplicates. The closest cluster is `pr-ci-security-gate.yml`, but the edits target different behavior changes.", - "confidence": 0.95, + "summary": "This cluster is mostly unrelated workflow-maintenance PRs that were grouped by title similarity. The only true overlap is that several PRs touch .github/workflows/pr-ci-security-gate.yml, but they make different security-related changes and should not be merged as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46554 is the most substantive and central workflow-security change in the set, with concrete logic changes to the PR CI security gate rather than a small metadata tweak.", + "canonical_pr_reason": "No single canonical PR: the items are distinct workflow edits, and even the repeated pr-ci-security-gate changes are separate follow-up changes rather than one duplicate fix.", "best_issue_reason": null, - "best_pr_reason": "PR 46554 is the best representative PR because it makes the clearest behavioral change to a shared security-sensitive workflow and sits at the center of the soft-similarity cluster.", + "best_pr_reason": "PR 46554 is the most substantive representative of the pr-ci-security-gate sequence, but it still is not a duplicate of the later edits.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and unrelated behavior: self-comment CI message piping vs PR security-gate hardening." + "reason": "Different workflows: self-comment CI vs PR CI security gate, with unrelated behavior changes." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Different workflow files and different changes; no shared concrete bug or fix." + "reason": "Different workflows and unrelated changes; only title style is similar." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different workflows and unrelated edits; one updates self-comment CI, the other adds comments to a security-gate workflow." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46138", "right": "pull_request:46716", "accept": false, - "reason": "Different workflows and different security-context changes; not the same underlying fix." + "reason": "Different workflows and unrelated changes." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow files and distinct shell-script handling changes; not duplicate fixes." + "reason": "Different workflows: stdin handling in self-comment CI vs commit-message parsing in build-ci-docker-images." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "Unrelated workflows: SSH runner setup vs benchmark trigger conditions." + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs benchmark dispatch trigger." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "Different workflows and distinct permission/setup adjustments; not the same bug." + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs anti-slop permissions." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "SSH runner changes are unrelated to assign-reviewer checkout behavior." + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs reviewer-assignment checkout ref." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and unrelated security changes; no shared concrete code-path problem." + "reason": "Different workflows and unrelated concerns." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "Different files and different workflow concerns; not duplicative." + "reason": "Different workflows and unrelated concerns." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Both are workflow chores, but they change different jobs and triggers in different files." + "reason": "Different workflows with different purposes: anti-slop permissions vs benchmark dispatch support." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Different workflow files and different fixes: permissions scoping vs checkout ref selection." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Different workflows and different shell/env handling; not the same underlying change." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Same broad CI area, but different workflows and different behavior changes." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Distinct workflow logic in different files; no same bug or change." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow files and unrelated fixes." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different security/checkout concerns; not mergeable as one duplicate fix." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow behaviors: schedule/manual trigger vs shell quoting fix." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Benchmark trigger change is unrelated to PR security-gate logic." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different shell-safety vs security-gate behavior." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "Different files and unrelated changes." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same workflow file, but different parts and different fixes: one adds checkout/diff-based security checks, the other changes permissions and approval gating." + "reason": "Same file, but different concrete security-gate edits: one adds repo checkout and file checks, the other changes permissions/approval flow." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "Same workflow file, but one adds enforcement logic while the other adds explanatory comments; not the same underlying change." + "reason": "Same file, but different changes: security checks vs reusable-workflow safety notes." }, { "left": "pull_request:46554", "right": "pull_request:46716", "accept": false, - "reason": "Same workflow family, but different changes: explicit checks vs changing inherited secrets to empty." + "reason": "Same file, but different changes: security checks vs secrets inheritance tightening." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Same file but different fixes: permissions/approval flow vs comment-only documentation." + "reason": "Same file, but different concrete changes in the security-gate workflow." }, { "left": "pull_request:46558", "right": "pull_request:46716", "accept": false, - "reason": "Both touch the gate workflow, but one adjusts job permissions/flow while the other removes inherited secrets; different changes." + "reason": "Same file, but different concrete changes in the security-gate workflow." }, { "left": "pull_request:46563", "right": "pull_request:46716", "accept": false, - "reason": "Comments explaining safety are not the same change as removing inherited secrets." + "reason": "Same file, but different concrete changes in the security-gate workflow." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Different workflows and unrelated edits." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46504", "right": "pull_request:46716", "accept": false, - "reason": "Different workflows and unrelated changes." + "reason": "Different workflows and unrelated behavior changes." } ] }, @@ -23182,8 +23164,8 @@ }, { "label": "pull_request", - "component_index": 30, - "component_count": 36, + "component_index": 29, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23199,12 +23181,12 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "fb57bc315b43217166c21aeede4a2553491cc33f505de9d846392e0e10f5b4e6", + "prepared_review_unit_hash": "117b492514d80860027b0d8b0ace4c6e0ac2f66988160d944f217880e36fa847", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4275, + "serialized_chars": 4276, "estimated_input_tokens": 1069, "estimated_eval_tokens": 2394 }, @@ -23212,7 +23194,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4275, + "serialized_chars": 4276, "estimated_input_tokens": 1069, "estimated_eval_tokens": 2394 }, @@ -23221,18 +23203,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address source-introspection code that previously relied on opening files, but they fix different call sites in different modules and are not the same underlying change.", + "summary": "These PRs share a theme of avoiding brittle source-file reads, but they fix different code paths in different modules: one in `import_utils.py` and one in `modeling_utils.py` with tests. They are not the same underlying change.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46207 is the clearer representative of the cluster theme: it fixes a specific code path in `modeling_utils.py` and includes tests. PR 46152 is a separate lower-level helper change in `import_utils.py`.", + "canonical_pr_reason": "PR #46207 is the better representative of the cluster because it is a concrete, self-contained fix with test coverage for a specific source-reading heuristic.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the strongest standalone candidate because it targets a concrete behavior change with test coverage; PR 46152 is related in spirit but addresses a different source-reading helper.", + "best_pr_reason": "PR #46207 is the strongest standalone candidate: it targets a precise bug in `_can_set_attn_implementation` and includes validation tests.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Related theme (replacing file-open source lookup), but different modules, different functions, and no shared concrete bug path." + "reason": "Same broad idea of avoiding `open()` for source inspection, but they touch different functions and modules with different failure modes; not mergeable as one PR." } ] }, @@ -23245,8 +23227,8 @@ }, { "label": "pull_request", - "component_index": 31, - "component_count": 36, + "component_index": 30, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23262,7 +23244,7 @@ "soft_pairs": [ "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "3fc5effeb8787f7c9c2c4541afb86ba414bcf2d98c4247405094dea2ba967ffc", + "prepared_review_unit_hash": "2cfc50d71a1d48c5db64dd8806aa030e0713b8b8344b44695b6f9999c163f511", "component_budget": { "node_count": 2, "item_count": 2, @@ -23284,18 +23266,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests add SDPA attention dispatch support, but they target different model families and distinct codepaths: T5 vs wav2vec2/audio models. They are related in theme, not duplicates.", + "summary": "Both PRs add SDPA attention dispatch support, but they target different model families and code paths (T5 vs wav2vec2/audio models). They are related in theme, not duplicates.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the two PRs solve separate model-specific implementation tasks, so neither is a duplicate of the other.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "If a representative PR is needed, #46196 is broader and central to the larger SDPA rollout across audio models, but it is still not a substitute for #46154.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "Same general feature label (SDPA dispatch), but different model families and concrete code paths; they would not plausibly be merged into one PR." + "reason": "Same broad goal, but different architectures and implementation paths. These would not plausibly be merged into one PR as a duplicate fix." } ] }, @@ -23308,8 +23290,8 @@ }, { "label": "pull_request", - "component_index": 32, - "component_count": 36, + "component_index": 31, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23374,8 +23356,8 @@ }, { "label": "pull_request", - "component_index": 33, - "component_count": 36, + "component_index": 32, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23391,7 +23373,7 @@ "soft_pairs": [ "pull_request:46310|pull_request:46546" ], - "prepared_review_unit_hash": "8f2d1ea8ed8fd22ac21be70692ec23d9ee27fa6988059aeb9081255fc2021e7a", + "prepared_review_unit_hash": "482a13eb32ea69b10273afaa02091c07a9fec2845d3d4e4c0b29a4deef73f98f", "component_budget": { "node_count": 2, "item_count": 2, @@ -23413,18 +23395,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same right-window boundary bug in the Moonshine streaming sliding-window mask; PR 46546 is the more complete patch because it also adds the Gemma4 counterpart and tests.", - "confidence": 0.93, + "summary": "Both PRs fix the same sliding-window right-boundary off-by-one bug (`<` to `<=`) in Moonshine streaming; PR 46546 is the broader, more complete version and also patches Gemma4 plus tests.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46546 is the more complete implementation of the same boundary fix: it includes the Moonshine streaming mask correction, the modular mirror, a related Gemma4 fix, and test coverage.", + "canonical_pr_reason": "PR 46546 is the best canonical PR because it contains the shared Moonshine streaming fix, extends the same bug fix to Gemma4, and includes test coverage, making it the more complete representative of this change.", "best_issue_reason": null, - "best_pr_reason": "PR 46546 is the best representative because it covers the same underlying bug as 46310 but with broader, more complete coverage and tests.", + "best_pr_reason": "PR 46546 is the best representative for the cluster since it addresses the same concrete boundary bug while covering more affected code paths and adding tests.", "soft_edge_verdicts": [ { "left": "pull_request:46310", "right": "pull_request:46546", "accept": true, - "reason": "Same underlying sliding-window boundary bug in Moonshine streaming; the Moonshine-specific diff matches exactly, and the additional Gemma4 change in 46546 does not make it a different problem." + "reason": "Same underlying off-by-one boundary bug in the sliding-window mask; the Moonshine change is identical, and 46546 simply broadens the fix to another model plus tests, so they could plausibly be merged into one PR." } ] }, @@ -23437,8 +23419,8 @@ }, { "label": "pull_request", - "component_index": 34, - "component_count": 36, + "component_index": 33, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23454,12 +23436,12 @@ "soft_pairs": [ "pull_request:46376|pull_request:46417" ], - "prepared_review_unit_hash": "df6655b7645090c2c88efef0311c00dc6e6bc02d018d2b90733888e1af0a21ce", + "prepared_review_unit_hash": "325f8cb6ab627d525a6b73e291606663699f044fef36ac74f94c9fb7c510404e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5937, + "serialized_chars": 5939, "estimated_input_tokens": 1485, "estimated_eval_tokens": 3226 }, @@ -23467,7 +23449,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5937, + "serialized_chars": 5939, "estimated_input_tokens": 1485, "estimated_eval_tokens": 3226 }, @@ -23476,18 +23458,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one adds multimodal support to continuous batching, while the other adds a new HunYuan VL model package with docs/tests. They should not be merged as duplicates.", + "summary": "The two pull requests are unrelated: one adds multimodal support to continuous batching internals, while the other introduces the HunYuan VL model package, docs, and tests. They do not appear to be duplicates or mergeable into a single PR.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46417 is the better representative PR for this cluster because it is a self-contained, concrete model addition with extensive supporting files, while 46376 is a draft infrastructure change for continuous batching and targets a different code path.", + "canonical_pr_reason": "PR 46417 is the more self-contained and clearly scoped change set, with a complete model addition across code, docs, and tests. PR 46376 is a separate draft infrastructure refactor for continuous batching.", "best_issue_reason": null, - "best_pr_reason": "PR 46417 is the most suitable standalone artifact here: it is complete enough to stand on its own and has the clearest, well-scoped change set. PR 46376 is draft and focused on a separate subsystem.", + "best_pr_reason": "PR 46417 is the strongest standalone representative of the cluster because it is a complete feature addition. PR 46376 is a different subsystem change and not a duplicate of it.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different underlying changes: continuous batching multimodal support vs. HunYuan VL model implementation. They touch different features and code paths, so they are not duplicates." + "reason": "Different underlying changes and code paths: continuous batching multimodal cache support versus adding a new HunYuan VL model. No shared filenames, targets, or concrete bug/fix overlap." } ] }, @@ -23500,8 +23482,8 @@ }, { "label": "pull_request", - "component_index": 35, - "component_count": 36, + "component_index": 34, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23517,7 +23499,7 @@ "soft_pairs": [ "pull_request:46574|pull_request:46580" ], - "prepared_review_unit_hash": "1bad85f6ba1d7d57e61608dad870e7f081b94c02e50e8f7c2d252ecdd59e40a1", + "prepared_review_unit_hash": "f4d6feade69219ea43c36ad1c6db0333769718b98c3bebc8f2a5ddfe064885cf", "component_budget": { "node_count": 2, "item_count": 2, @@ -23539,18 +23521,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated tokenizer fixes: one corrects CodeLlama decode behavior that strips real leading spaces, while the other corrects DeepSeek checkpoint routing to the right tokenizer class in auto-loading.", + "summary": "Two unrelated tokenizer fixes: one for CodeLlama decode stripping, one for DeepSeek checkpoint/tokenizer-class resolution. They share a broad theme but not the same underlying bug or code path.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR for duplication: PR 46574 is a decoder/strip fix in CodeLlama, while PR 46580 changes auto-tokenizer model-class selection for DeepSeek checkpoints.", + "canonical_pr_reason": "PR #46574 is the narrower, cleaner bug fix centered on a concrete decode regression in CodeLlama and includes a targeted tokenizer change plus tests.", "best_issue_reason": null, - "best_pr_reason": "PR 46580 is the broader tokenizer-loading fix and the cleaner representative of the cluster\u2019s auto-loading side, but it is not the same bug as PR 46574.", + "best_pr_reason": "PR #46574 is the stronger representative because it fixes a single, well-scoped runtime bug with a direct code-path change; PR #46580 is a separate auto-tokenizer mapping issue.", "soft_edge_verdicts": [ { "left": "pull_request:46574", "right": "pull_request:46580", "accept": false, - "reason": "Different underlying bugs and code paths: CodeLlama decode stripping vs DeepSeek tokenizer-class misrouting. They are not mergeable into one PR." + "reason": "Different problems and code paths: CodeLlama decoder initialization/strip behavior vs DeepSeek checkpoint-prefix tokenizer-class override. They would not plausibly be merged into one PR as a single fix." } ] }, @@ -23563,8 +23545,8 @@ }, { "label": "pull_request", - "component_index": 36, - "component_count": 36, + "component_index": 35, + "component_count": 35, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -23580,7 +23562,7 @@ "soft_pairs": [ "pull_request:46603|pull_request:46604" ], - "prepared_review_unit_hash": "e80b428d929648fb27d335707c35ce36e7c4fc0c8b49052effd9660d0b65a440", + "prepared_review_unit_hash": "35bad15ba8f2447d982fa3aeb685899268b0425ec47a56caa7e93e2e96186940", "component_budget": { "node_count": 2, "item_count": 2, @@ -23602,18 +23584,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target Mistral tokenization internals and touch many of the same files, but they address different changes: one is a refactor moving `MistralConverter` into an integration package, while the other adds tekken tokenizer detection/conversion/save utilities. They are related, but not the same underlying fix/change, so they should not be merged as duplicates.", - "confidence": 0.93, + "summary": "These are both Mistral-related PRs, but they address different changes: one is a package refactor/move of MistralConverter, the other adds tekken tokenizer detection/conversion/save support. They are related but not duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46604 is the more substantive feature change, adding tekken tokenizer detection, conversion, and save utilities rather than only relocating existing converter code.", + "canonical_pr_reason": "PR #46604 is the better canonical PR because it contains the more substantive user-facing tokenizer functionality; #46603 is mostly a code-organization move of MistralConverter.", "best_issue_reason": null, - "best_pr_reason": "PR #46604 is the strongest representative of the cluster because it introduces the broader Mistral tokenizer functionality and is the more self-contained user-facing change.", + "best_pr_reason": "PR #46604 is the best representative of the cluster since it introduces the main behavioral change, while #46603 is a narrower refactor in the same area.", "soft_edge_verdicts": [ { "left": "pull_request:46603", "right": "pull_request:46604", "accept": false, - "reason": "Shared Mistral tokenizer files, but the scopes differ: #46603 is a code move/refactor, while #46604 implements tekken tokenizer support and save utilities. They do not appear to fix the same concrete code-path problem." + "reason": "Different concrete changes: #46603 moves MistralConverter into a new package location, while #46604 adds tekken tokenizer detection/conversion/save utilities. They touch similar files but not the same underlying fix and would not naturally merge into one PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 27630afddf493eb23c5d64caffc4ecff0827597a..0a354539a1a12bf855849e3dca2533754c7fd625 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260617t180027z", + "analysis_id": "hybrid-model-20260618t000034z", "archived_artifacts": { - "hybrid": "snapshots/20260617T180027Z/analysis-runs/hybrid-model-20260617t180027z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260617T180027Z/analysis-runs/hybrid-model-20260617t180027z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-17T18:13:21Z", + "published_at": "2026-06-18T00:09:37Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260617T180027Z", + "snapshot_id": "20260618T000034Z", "variant": "hybrid" } diff --git a/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..8d563ac2f7b11dbc60a5b82a4314e0fc0c944a05 --- /dev/null +++ b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json @@ -0,0 +1,2220 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260618T000034Z", + "generated_at": "2026-06-18T00:09:35Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 41116, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.809, + "size_similarity": 0.893, + "file_overlap": 0.75, + "area_overlap": 0.871, + "patch_similarity": 0.669, + "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": [ + [ + 737, + 744 + ] + ], + "right_ranges": [ + [ + 730, + 737 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 301, + 307 + ], + [ + 742, + 748 + ], + [ + 1360, + 1366 + ] + ], + "right_ranges": [ + [ + 720, + 726 + ], + [ + 1331, + 1337 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 149 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 550 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 394 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45987-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45987.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45987, + "canonical_pr_number": 46173, + "issue_numbers": [ + 45987 + ], + "pr_numbers": [ + 46173, + 46474 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 46173, + "right_pr_number": 46474, + "code_similarity": 0.65, + "size_similarity": 1.0, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 1.0, + "shared_filenames": [ + "src/transformers/cache_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cache_utils.py", + "left_ranges": [ + [ + 69, + 75 + ], + [ + 375, + 381 + ] + ], + "right_ranges": [ + [ + 79, + 85 + ], + [ + 461, + 467 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.494, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.434, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3595, + 3602 + ], + [ + 3819, + 3853 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46305-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46305.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46305, + "canonical_pr_number": 46546, + "issue_numbers": [ + 46305 + ], + "pr_numbers": [ + 46310, + 46546 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46310, + "right_pr_number": 46546, + "code_similarity": 0.566, + "size_similarity": 0.444, + "file_overlap": 0.5, + "area_overlap": 0.867, + "patch_similarity": 0.16, + "shared_filenames": [ + "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "left_ranges": [ + [ + 356, + 362 + ] + ], + "right_ranges": [ + [ + 355, + 361 + ] + ] + }, + { + "filename": "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py", + "left_ranges": [ + [ + 269, + 275 + ] + ], + "right_ranges": [ + [ + 269, + 275 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46562-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46562, + "issue_numbers": [], + "pr_numbers": [ + 46562, + 46658 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46562, + "right_pr_number": 46658, + "code_similarity": 0.168, + "size_similarity": 0.401, + "file_overlap": 0.032, + "area_overlap": 0.0, + "patch_similarity": 0.523, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 441, + 447 + ], + [ + 591, + 598 + ], + [ + 823, + 829 + ], + [ + 883, + 889 + ], + [ + 1082, + 1088 + ], + [ + 1205, + 1211 + ] + ], + "right_ranges": [ + [ + 362, + 368 + ], + [ + 795, + 801 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 41116, + "duplicate_pr_numbers": [ + 45613 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-45987-3", + "canonical_pr_number": 46173, + "duplicate_pr_numbers": [ + 46474 + ], + "target_issue_number": 45987, + "reason": "PRs in cluster-45987-3 are treated as duplicates because they converge on issue #45987 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 46546, + "duplicate_pr_numbers": [ + 46310 + ], + "target_issue_number": 46305, + "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46562-2", + "canonical_pr_number": 46562, + "duplicate_pr_numbers": [ + 46658 + ], + "target_issue_number": null, + "reason": "PRs in cluster-46562-2 are treated as duplicates because they share soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-41211-3", + "issue_number": 41211, + "reason": "Issue #41211 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 6 discussion comments plus 6 inbound references.", + "score": 148.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..a381346b4d56ac13c0eb962994198b8a8e36b533 --- /dev/null +++ b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,23610 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260618T000034Z", + "generated_at": "2026-06-18T00:09:35Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43425", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44841", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46468", + "issue:46515", + "issue:46661" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:46468", + "issue:45237|issue:45507", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:46165|issue:46468", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:44841|issue:46661", + "issue:24643|issue:30277", + "issue:34928|issue:44805", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:44292|issue:45237", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:45800|issue:46515", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43828|issue:44292", + "issue:43541|issue:43716", + "issue:43716|issue:46468", + "issue:43716|issue:45685", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:44918|issue:45161", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805" + ], + "prepared_review_unit_hash": "7c86c1552f899ba60896cee129b60c46ec4b166293331562afb6bb8f15ec6dd4", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 62, + "serialized_chars": 45731, + "estimated_input_tokens": 11433, + "estimated_eval_tokens": 23122 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous runtime/docs/model-compatibility issues. I found one clear duplicate pair: the two DeepSpeed ZeRO-3 + BERT index-out-of-bounds reports (43596/43638). The rest are only loosely similar and should stay separate.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43596 is the cleanest representative of the only true duplicate pair in the set: both it and 43638 report the same DeepSpeed ZeRO-3 / BERT index-0-out-of-bounds failure during model init.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 is the best issue anchor because it captures the duplicate subcluster most clearly and is the cleaner title of the matching pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: Qwen3.5 config attribute access vs chat/completions streaming inference error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different models/backends: GPT-OSS on AMD GPUs vs MoE MPS histogram unsupported for Int." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different problem classes: image preprocessor dtype mismatch vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch is not the same bug as GPT-OSS-20B on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "One is GPT-OSS on AMD GPUs; the other is a broken Nemotron example. Not the same underlying issue." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific failures; no shared concrete code path." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Deepspeed collective mismatch vs ModernBERT ONNX export error are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Both are model export/runtime issues, but different models and different failure modes." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3 auto device mapping are unrelated." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Both concern docs/examples, but they target different features and different doc problems." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanup issues, but they reference different removed commands/tasks and are not the same change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ModernBERT ONNX export and GenerationMixin import error are different regressions." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "Different trainer/resume scenarios and different reported failures." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Generic mask/tensor shape mismatch is not the same as Gemma4 device-map forward mismatch." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are index/shape errors, but the model paths and root causes differ." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed ZeRO-3 BERT init error is unrelated to resume-training with trust_remote_code." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Similar error family, but not the same code path or model setup." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Mask shape mismatch and video token/feature mismatch are different bugs." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Generic tensor mask mismatch is not the same as GPT-OSS MoE tensor-parallel failure." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Resume-training with remote code and a broken Nemotron example are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration in chat template is unrelated to resume-training issues." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed ZeRO-3 BERT init failure is unrelated to resume-training with remote code." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset and ModernBERT ONNX export are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Different distributed failures; only superficial similarity in deepspeed wording." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying bug: DeepSpeed ZeRO-3 + BERT initialization hits the same index-0-out-of-bounds failure." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy call path is unrelated to GPT-OSS tensor-parallel issues." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different model/tokenization issue vs LoRA merge/save/reload mismatch." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA reload inconsistency are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral and remote-code config_class error are different problems." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both mention DeepSpeed, but one is a weight-shape RuntimeError and the other is a collective mismatch." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation checkpointing/FSDP recompute mismatch is not the same as a generic mask/index shape error." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same component family, but different reported bugs: CPU offload incompatibility vs uninitialized memory after materialization." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "BERT ZeRO-3 init failure and GPT-OSS MoE TP failure are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP checkpointing recompute mismatch and CustomTrainer.compute_loss signature break are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different models and code paths; no duplicate evidence." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific device/dtype issues." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different GPT/Qwen model/runtime failures and different hardware paths." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Different multimodal bugs: reverse embedding lookup vs video feature/token mismatch." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config attribute error and Mistral regex patch crash are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Gemma4 device mismatch and reverse embedding lookup are different bugs." + }, + { + "left": "issue:45800", + "right": "issue:46515", + "accept": false, + "reason": "Both mention torch compatibility, but they are different incompatibility reports." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch 2.10 incompatibility and Qwen-3-8B-NVFP4 runtime error are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation checkpointing and Qwen3 auto device mapping are different failures." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch is not the same as MPS histogram not implemented for Int." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different models and different failure modes." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral torch dynamo grouped_mm tracing error is unrelated to image preprocessor dtype mismatch." + }, + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Different model families and unrelated failures." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different model/backend issues; no same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the summary stays conservative, and the only accepted soft pair (issue:43596/issue:43638) is well supported by the titles. The other verdicts mostly avoid overclaiming and do not look like obvious duplicates from the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 2, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44841", + "issue:44863", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45237", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304", + "issue:46402", + "issue:46661" + ], + "soft_pairs": [ + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720", + "issue:43901|issue:44512", + "issue:44589|issue:46304", + "issue:45237|issue:45901", + "issue:43039|issue:44560", + "issue:43975|issue:46015", + "issue:45823|issue:46294", + "issue:44841|issue:45823", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:45812|issue:45920", + "issue:43866|issue:44863", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:43596|issue:45800", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:36010|issue:39290", + "issue:43541|issue:43828", + "issue:43404|issue:43866", + "issue:44841|issue:45084", + "issue:41093|issue:41762", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823", + "issue:43296|issue:43366", + "issue:44387|issue:45005" + ], + "prepared_review_unit_hash": "26b5b8b76dd8f0aa7f4295f1f0d464629f541f93743428f673d7b3e955fd9d0a", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39792, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issue reports; only two soft pairs look like true duplicates: the AutoTokenizer v5 token-ID regression and a DeepSpeed ZeRO-3 index-out-of-bounds loading failure.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45920 is the broad, current umbrella for the AutoTokenizer v5 token-ID regression across multiple model families, so it best represents the duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the most useful triage target: open, broad, and already framed as the shared tokenizer regression rather than a narrow model-specific instance.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Different model families and failure modes: corrupted-checkpoint/load failure vs missing text-encoder weight loading." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Both are compatibility-related, but one is a torch version incompatibility and the other is support for USE_DISTRIBUTED=0." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "PaddleOCR-VL/vLLM model load failure vs distributed-build support regression; unrelated code paths." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 init/loading failure class: index 0 out of bounds from empty/size-0 tensors while loading models." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different bugs: Gemma3 sliding_window_pattern config issue vs Qwen3 MoE auto device-map CUDA assert." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs-related, but they cover different APIs and different stale references." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup error vs distributed support regression; not the same underlying bug." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "AMD GPU support failure vs table-question-answering task crash; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger cross-entropy path issue vs Qwen3-vl video StopIteration; different subsystems." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Tokenizer detokenization regression vs outdated RAG examples; unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 device-mapping mismatch vs Qwen3.5 video/chat-template mismatch; different code paths." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral vs Gemma4 PLE device mismatch; not the same issue." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint-loading bug vs docs/examples update; unrelated." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code model breakage vs distributed-build support regression; different failures." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same AutoTokenizer v5 token-ID regression; 45920 is the broader umbrella report spanning more affected model families." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint/load issue vs NemotronH implementation cannot load checkpoints; different models and symptoms." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params API incompatibility vs distributed support regression; unrelated." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error vs RAG docs issue; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but Gemma3TextConfig/vLLM and Qwen3-MoE are different model/config bugs." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are ZeRO-3 index errors, but one is a mask/tensor shape mismatch and the other is an empty-index out-of-bounds during BertModel init." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper model load failure vs table-question-answering crash; unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash test loading failure vs Ovis2 checkpoint corruption; different model issues." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 incompatibility vs pydantic PretrainedConfig field regression; different incompatibilities." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Tokenizer detokenization bug vs table-question-answering crash; unrelated." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper load failure vs pydantic model field regression; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Both are version-sensitive issues, but ZeRO-3 empty-index init failure and torch/transformers compatibility are not the same bug." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 failures: chat crash vs streaming inference API failure." + }, + { + "left": "issue:45685", + "right": "issue:46304", + "accept": false, + "reason": "MPS histogram int error in MoE vs distributed support regression; unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Both are MoE-related, but grouped_mm tracing and MPS histogram failures are different code paths." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin vs Gemma3TextConfig sliding_window_pattern; unrelated." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure vs autocast dtype mismatch; different MoE bugs." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Tied lm_head weights in Mistral3 vs Ovis2 checkpoint corruption; unrelated." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor load failure vs compile non-template nodes; unrelated." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are ZeRO-3 index errors, but the reported mechanics differ: mask-shape mismatch vs empty-index out-of-bounds." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "flash-attn fallback gating vs remote-code model_class/config_class error; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLm pad_token_idx regression vs pydantic PretrainedConfig field regression; different config bugs." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers compatibility issue vs Gemma4 device mismatch; different failures." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL/vLLM loading failure vs GGUF gpt-oss support request; unrelated." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 CUDA memory regression vs tied-weights translation-model issue; different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The triage is grounded in the packet: the canonical issue choice is well-supported by the open, broader AutoTokenizer v5 regression report, and the soft-edge verdicts stay conservative overall. The two accepted pairs both have strong title-level alignment (shared exact regression/error class), while the rejected pairs are mostly kept separate for sensible model- or subsystem-specific differences." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 3, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:39186", + "issue:40990", + "issue:41093", + "issue:42915", + "issue:43054", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43873", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44589", + "issue:44805", + "issue:44863", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45161", + "issue:45183", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294", + "issue:46463", + "issue:46489", + "issue:46491", + "issue:46640", + "issue:46661" + ], + "soft_pairs": [ + "issue:44805|issue:44918", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:43531|issue:45070", + "issue:43541|issue:44589", + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:41093|issue:43596", + "issue:44479|issue:46294", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:46640|issue:46661", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:40990|issue:44960", + "issue:45632|issue:45698", + "issue:43645|issue:45399", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44863|issue:45005", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:45183|issue:46463", + "issue:46489|issue:46491", + "issue:43975|issue:44292", + "issue:44292|issue:45507" + ], + "prepared_review_unit_hash": "5b3e321af2af67b518dfbbf07c0092ec498de914d803c8df5e35d584e9832298", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36254, + "estimated_input_tokens": 9064, + "estimated_eval_tokens": 18384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "A few pairs look like the same underlying regression, but most are only superficially similar and should remain separate. No PRs are present.", + "confidence": 0.67, + "canonical_issue_reason": "issue:44479 is the clearest representative of an accepted duplicate-like set: it describes the broad v5.3.0 video-input regression across several Qwen VL / Qwen3.5 variants and names the concrete broken path.", + "canonical_pr_reason": null, + "best_issue_reason": "Best overall issue to anchor triage because it states a concrete regression with a wide affected surface and is the most representative of the accepted video-input bug family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both mention mask/shape failures, but one is a generic mask-index mismatch and the other is a TRL SFT/input-embedding failure on Qwen3.5; too little evidence of the same root cause." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Different models and different failures: grouped_mm tracing error in Mixtral vs tensor-parallel-only failure in GPT-OSS MoE." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a streaming chat/completions failure on Qwen3.5-0.8B. Different code paths." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support for gpt-oss is unrelated to a Qwen3.5 streaming API failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "`init_empty_weights`/_is_hf_initialized construction error is a different bug from the pydantic `PretrainedConfig` field regression." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and a pydantic model-field regression are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Grouped_mm tracing on Mixtral and missing Float8 storage deserialization are different runtime failures." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Mistral-3 image-preprocessor dtype mismatch is unrelated to Qwen3.5 video token/feature mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Remote-code breakage and torch/transformers version incompatibility are too broad to be the same concrete bug." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors, but one is a mask-length mismatch and the other is a zero3/BertModel empty-tensor index error; different failure modes." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "Same Qwen video-input regression family: both describe mismatches between video features and video tokens/chat templating on Qwen3.5 variants." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": true, + "reason": "Both are SigLIP2-specific reports of bad/nonsensical outputs, suggesting the same underlying model implementation regression." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite-config unwrapping in AutoModelForCausalLM and uninitialized rotary embedding memory after to_empty() are different bugs." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 integration and remote-code model_class/config_class lookup are separate failures." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different models and symptoms: Mistral-3 image dtype mismatch vs Qwen3.5 input-embedding/TRL packing failure." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight-shape initialization error is unrelated to Qwen3.5 input-embedding unpacking." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are not the same bug." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": true, + "reason": "Both point to remote-code/custom-module resolution loading the wrong artifact because the cache/path identity is not unique enough." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization in v5 and Flash-Attn2 fallback gating are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Untied lm_head weights in Mistral3 are not the same as translation-model tied-weight issues in v5." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FineGrainedFP8 failure and Flash-Attn2 fallback gating are different subsystems and failure modes." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and tied-weights translation-model regressions do not share a concrete code path." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Offloading/quantization behavior and tied-weight translation-model regressions are too far apart." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer ReLU discussion is unrelated to a documentation typo about the removed run command." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading and translation-model tied weights are different defects." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "Cache-strategy typo and removed-doc-command mismatch are just documentation issues, not the same one." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3.5 input-embedding unpacking and GraniteMoEHybrid invalid-method calls are unrelated." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": true, + "reason": "Both are input-validation error-message bugs in the shared decoder/input-checking path; T5 is a concrete instance of the broader XOR validation problem." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Different tokenizer regressions on different model families; same general area, but not the same underlying bug." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "DeepSeek-Coder v1 detokenization regression and an NVFP4 loading/runtime error are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "NVFP4 runtime failure and GraniteMoEHybrid invalid method are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary matches the packet, the accepted soft edges are plausible from the titles alone, and most non-matches are conservatively rejected. No PRs are present." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 4, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:24643", + "issue:28530", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43782", + "issue:43824", + "issue:43828", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44661", + "issue:44910", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45399", + "issue:45520", + "issue:45663", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:45920", + "issue:46489" + ], + "soft_pairs": [ + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:36331|issue:41762", + "issue:43872|issue:44589", + "issue:43828|issue:44928", + "issue:45920|issue:46489", + "issue:43866|issue:44403", + "issue:43054|issue:45701", + "issue:38175|issue:43054", + "issue:43950|issue:45902", + "issue:41553|issue:43883", + "issue:44910|issue:45859", + "issue:43054|issue:43493", + "issue:42947|issue:44387", + "issue:45663|issue:45902", + "issue:41553|issue:43881", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:44403|issue:44488", + "issue:24643|issue:28530", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:44991|issue:45399", + "issue:43824|issue:43872", + "issue:45685|issue:45823", + "issue:43782|issue:44960", + "issue:43881|issue:43883", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:45081|issue:45399", + "issue:42491|issue:44368", + "issue:43828|issue:45823" + ], + "prepared_review_unit_hash": "e555f838f81ea1a151821e8da191de576f1c86f0fb68392264a188b9d61cab82", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38112, + "estimated_input_tokens": 9528, + "estimated_eval_tokens": 19312 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: many unrelated model-specific bugs, tokenizer/loading regressions, and training/runtime issues. The only clearly duplicate subcluster is the Transformers v5 non-persistent-buffer regression; the tokenizer mis-tokenization pair also looks like the same underlying v5 tokenizer bug family.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43950 is the clearest root report for the Transformers 5.x non-persistent-buffer regression, and it best anchors the strongest duplicate subcluster (with 43644/44534/45902).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the best representative because it describes the core bug directly, rather than a downstream symptom or a broader follow-on report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same Transformers v5 non-persistent-buffer regression; 44534 is the same corruption/junk-buffer bug described in 43950." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is a docs/example mismatch and quantization failure while the other is degraded text embeddings; not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same v5 non-persistent-buffer corruption bug; 43644 and 43950 describe the same failure mode with different wording." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Different failures: a custom `compute_loss` signature mismatch vs a DeepSpeed ZeRO-3 loading index error." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different quantization/storage incompatibilities (`_is_hf_initialized` vs missing `Float8_e4m3fnStorage`), not the same code path." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different model/runtime problems: autocast dtype mismatch on Phi-tiny-MoE vs RLHF instability from Qwen3.5 3D `position_ids` handling." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": true, + "reason": "Same tokenizer regression family in v5+; 46489 is a specific DeepSeek-Coder instance of the broader wrong-token-IDs bug." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "A corrupted checkpoint and a loading-noise issue are unrelated." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "One is a SigLIP2 embedding-quality complaint, the other is a tokenizer-version behavior change; too broad to be the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both involve SigLIP2 output quality, but they describe different symptoms and no shared concrete failure path." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": true, + "reason": "Same non-persistent-buffer materialization regression; 45902's uninitialized `inv_freq` is a concrete manifestation of 43950's bug." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated issues: bad AutoTokenizer error for Voxtral vs missing `all_tied_weights_keys` on MolmoForCausalLM." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3.5 issues: FlashAttention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Related to SigLIP2, but one reports poor embeddings and the other reports HF/JAX parity mismatch; not clearly the same concrete bug." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Different training issues: ineffective gradient checkpointing with PEFT LoRA vs higher reserved memory under int4 quantization." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Different root causes: Gemma4 FSDP2 `shared_kv_states` KeyError vs Qwen3_5Moe `inv_freq` uninitialized memory." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading failures with no shared concrete code path." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "A generic loading-noise report and a specific tokenizer/model loading failure are not duplicates." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "DeepSeek-Coder detokenization and a Python 3.13 flash_attn import error are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Unnecessary load noise vs a specific model loading failure; not the same bug." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "Training with DeepSpeed causing a weight-dimension error is unrelated to early stopping metric resolution." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both touch `remote_code`, but one is a missing-fields config bug and the other is a broad breakage claim; not the same concrete failure." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different memory-related training problems with different mechanisms." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading error and Qwen3 MoE/LoRA compatibility issue are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure and flash-attn fallback gating are separate subsystems and separate bugs." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Importing a missing model class and a bitsandbytes Int8Params constructor incompatibility are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram implementation gap vs Gemma4 device-map mismatch are different errors." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different model-specific loading issues with no shared code path." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different model-loading failures: GLM4V load failure vs missing `all_tied_weights_keys`." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Broad remote-code breakage and a Gemma4 device mismatch are not the same issue." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer-mapping helper failure and remote_code breakage are different problems." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral tokenizer regex patch crash and flash-attn2 fallback gating are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen3/MoE/LoRA, but one is a version-compatibility break and the other is a configuration warning." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different runtime/device issues: dtype mismatch vs device-map placement mismatch." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the cluster is clearly heterogeneous, and the duplicate calls are mostly conservative. The non-persistent-buffer v5 regression subcluster is well supported by the titles, and the tokenizer regression pair is also reasonably framed as a related v5-family issue rather than overclaimed as a universal duplicate." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 5, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43596", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43872", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44560", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45356", + "issue:45488", + "issue:45588", + "issue:45663", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45902", + "issue:45923", + "issue:46315", + "issue:46402", + "issue:46489", + "issue:46515", + "issue:46661" + ], + "soft_pairs": [ + "issue:45020|issue:45356", + "issue:45715|issue:46315", + "issue:43827|issue:44509", + "issue:45488|issue:46489", + "issue:43404|issue:43950", + "issue:45759|issue:46661", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:45020|issue:46661", + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:41553|issue:44991", + "issue:42947|issue:45923", + "issue:36331|issue:41093", + "issue:43940|issue:45313", + "issue:43854|issue:44863", + "issue:44479|issue:44560", + "issue:41553|issue:44488", + "issue:41093|issue:44805", + "issue:44928|issue:45588", + "issue:24643|issue:30064", + "issue:44863|issue:46402", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:43596|issue:46515", + "issue:43366|issue:43531", + "issue:45325|issue:45902", + "issue:45663|issue:45823" + ], + "prepared_review_unit_hash": "49c326d8de779fa026a1f13201a85b863a11a0777f31c6b98b9001c1b09ecdb2", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36628, + "estimated_input_tokens": 9157, + "estimated_eval_tokens": 18570 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous, but a few pairs look like true duplicates: stale v5 docs references, DeepSeek tokenizer regressions, the `_is_hf_initialized` init-empty-weights/typeerror path, and the identical mask-shape IndexError. Most other pairs are only loosely related by subsystem or model family and should stay separate.", + "confidence": 0.67, + "canonical_issue_reason": "Issue 45020 is the broadest umbrella report here: it describes a general recent-version regression affecting remote-code models, which makes it the best anchor for the cluster even though most other items are narrower and not duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the most representative issue for the overall cluster because it captures the common theme of v5 regressions breaking previously working model-loading paths, while the rest are mostly model-specific edge cases.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Related regression theme, but not the same bug: one is a broad remote_code breakage, the other is a specific Kimi-K2.5 tokenizer/codec regression." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they hit different methods (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and different code paths." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs cleanup problem: both report stale references to pipeline/tasks removed in v5." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": true, + "reason": "Same tokenizer regression family: DeepSeek tokenization is broken because v5/LlamaTokenizer no longer respects the tokenizer.json pre-tokenizer behavior." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Different bugs: tied `lm_head` weights in Mistral3 vs corruption of non-persistent buffers during `from_pretrained()`." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "Both involve model-loading APIs, but one is config unwrapping for composite multimodal configs and the other is a remote-code `config_class` attribute failure." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying `_is_hf_initialized` incompatibility during empty-weight/model materialization." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Both mention model-init/dtype/FP8-related failures, but the concrete bugs and code paths are different." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Both are broad model-loading regressions, but the specific failure modes differ: remote_code breakage vs missing `config_class` on `from_pretrained()`." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Too generic to merge confidently; the titles suggest a Whisper run failure and a specific fbank feature-extraction issue, but not the same concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Both involve DeepSpeed and Qwen3-family models, but one is weight conversion/loading and the other is gradient averaging with `model_accepts_loss_kwargs=True`." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/loading regressions, but they affect different models and different failure points." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both touch gradient checkpointing, but one is a PEFT/LoRA effectiveness bug and the other is a missing capability flag on a trust-remote-code model." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different trainer/runtime failures: an unexpected `num_items_in_batch` kwarg versus a tensor mask-shape mismatch." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Similar DeepSpeed ZeRO-3 symptom, but the model variants and reported missing-weight patterns differ enough that this is not safely the same bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Both are checkpoint/model-loading failures, but they concern different architectures with different load paths." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input regressions, but they affect different model families and different processing stages." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer/model-loading issues, but the failures and targets are different." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same underlying IndexError pattern: a mask/tensor length mismatch at index 0 with the same shape-mismatch symptom." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Different problems: RLHF NaNs from Qwen3.5 position-id/SDPA behavior versus a flash-attention AttributeError on `s_aux=None`." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Completely unrelated: DeepSpeed training weight-shape error versus image-processor void segmentation-map handling." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Different model families and different loading failures; not the same checkpoint bug." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Same broad area of model support, but one is a PagedAttentionCache runtime crash and the other is a checkpoint-loading failure." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Both are model/runtime compatibility issues, but one is GGUF architecture support and the other is a dtype mismatch under autocast." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "No concrete overlap: ZeRO-3/Bert init failure versus torch-version incompatibility." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Different Qwen3-related bugs: GGUF architecture support versus sliding_window behavior." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Both involve positional/rotary embeddings, but the symptoms and code paths are distinct." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same Gemma4 family, but one is an FSDP2 shared-kv-state KeyError and the other is a device-mapping mismatch in forward." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet, and the representative issue choice is reasonable. The soft-edge decisions are mostly conservative: the accepted pairs are well-supported by near-identical titles/root causes (docs pipeline removals, `_is_hf_initialized`, mask-shape IndexError) or a clearly shared tokenizer regression family, while the rejected pairs are appropriately kept separate when the concrete failure modes differ. No obvious overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 6, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43720", + "issue:43824", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45276", + "issue:45584", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:46459", + "issue:46489", + "issue:46512", + "issue:46515" + ], + "soft_pairs": [ + "issue:43638|issue:46515", + "issue:43366|issue:44589", + "issue:43716|issue:44560", + "issue:44155|issue:45584", + "issue:43950|issue:45834", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:44560|issue:45685", + "issue:43645|issue:43646", + "issue:44560|issue:44918", + "issue:43550|issue:44910", + "issue:44683|issue:46515", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:45812|issue:46489", + "issue:45823|issue:46515", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:38175|issue:43994", + "issue:43632|issue:45834", + "issue:44877|issue:45030", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:43317|issue:44756", + "issue:44661|issue:44805", + "issue:42491|issue:43720", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:41720|issue:42915", + "issue:42915|issue:44756", + "issue:42491|issue:45276", + "issue:44451|issue:44488", + "issue:44960|issue:44977", + "issue:46459|issue:46512" + ], + "prepared_review_unit_hash": "96260b857ebdddf979cce0768b06c9525effbc7295edda1bdd54886b22fb8880", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37772, + "estimated_input_tokens": 9443, + "estimated_eval_tokens": 19142 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated Transformers regressions. The only clear duplicate pair is the non-persistent buffer corruption reports (43644/44534); a SigLIP2 output issue and a custom-model-init regression also look duplicate-like, but most other pairs are only broadly similar and should stay separate.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43644 is the strongest canonical anchor: it\u2019s the earliest, clearest statement of the non-persistent-buffer regression, and issue:44534 is an exact wording follow-up of the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best overall issue to keep as the representative duplicate target because it has the cleanest, most specific bug description and an exact near-duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "Different failures: Deepspeed ZeRO-3/Bert shape issue vs a torch 2.4.1 incompatibility report." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated topics: GGUF gpt-oss support vs Float8 storage lookup failure." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different model families and code paths: Mistral-3 image dtype mismatch vs Qwen3-VL video StopIteration." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo track embedding leak and Whisper empty transcription are different bugs." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer loading regression vs Kosmos2.5 long-OCR index error; not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing a missing class and meta-device loading failures are different code paths." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3-MoE FP8 failure and meta-device loading breakage are separate issues." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-VL video StopIteration and MoE MPS int-histogram error are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Both describe the same v5 regression breaking custom model initialization; the notebook mention is just a narrower manifestation." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen3.5/TRL failures and different symptoms; not the same underlying bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to Qwen3.5 flash-attention illegal access." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and torch 2.4.1 incompatibility are different version-specific problems." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py run failure is unrelated to GGUF gpt-oss support." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "ImportError for GenerationMixin and Qwen3-MoE FP8 failure are unrelated." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Both concern tokenization, but they target different models and likely different tokenizer bugs." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 device-mismatch during forward is not the same as a generic torch 2.4.1 incompatibility." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "add-new-model-like validation failure and Gemma4 PLE device mismatch are different issues." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Qwen video StopIteration and a mask shape IndexError are not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 model and same wrong-output symptom; the second report is a rephrasing/expansion of the first." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized flag breakage and Kosmos2.5 OCR index errors are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Strict config blocking granite_speech and glm4v config validation are separate config problems." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "ZeRO-3/Bert empty-shape index error and add-new-model-like tokenizer-mapping failure are different code paths." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Both are index/config style failures, but the Bert ZeRO-3 error is not the same bug as tokenizer-mapping validation." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "gpu+cpu offload dequantized-model loading and Strix Halo mmap OOM are different loading issues." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3-MoE LoRA incompatibility and BitNet packed-weight loading are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact same bug: Transformers v5 fills non-persistent buffers with junk; 44534 is just a later re-report." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but 39692 is a docs example/model-processor mismatch and quantization issue, not the same output-corruption bug." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and Qwen3-MoE FP8 failure are different failures." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3-MoE FP8 and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Qwen3-MoE LoRA hf4.x/hf5.x incompatibility and Gemma4 resize_token_embeddings are different regressions." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading failures, but the reports name different BERT repos and don\u2019t clearly point to one concrete shared bug." + }, + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype, but one is silent dtype loss on checkpoint load and the other is stale config.dtype after from_config; different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The judgments are conservative and grounded in the packet. The three accepted soft pairs are plausible near-duplicates, and the rejected pairs mostly stay on the cautious side given the differing titles and symptom details. The canonical choice of issue:43644 is well supported by the exact follow-up issue:44534." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 7, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44402", + "issue:44488", + "issue:44756", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45507", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45859", + "issue:46050", + "issue:46468" + ], + "soft_pairs": [ + "issue:45313|issue:45357", + "issue:45507|issue:46468", + "issue:44488|issue:44991", + "issue:43856|issue:43873", + "issue:43644|issue:43645", + "issue:43856|issue:44387", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:42491|issue:45292", + "issue:43708|issue:44075", + "issue:42491|issue:43931", + "issue:44960|issue:46050", + "issue:45488|issue:45812", + "issue:44164|issue:44291", + "issue:45362|issue:45406", + "issue:45276|issue:45335", + "issue:44910|issue:45588", + "issue:44402|issue:44488", + "issue:43404|issue:43454", + "issue:43425|issue:43957", + "issue:43873|issue:44387", + "issue:39692|issue:44960", + "issue:41720|issue:44155", + "issue:42947|issue:45704", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:36010|issue:39404", + "issue:44387|issue:44756", + "issue:45663|issue:45859" + ], + "prepared_review_unit_hash": "9dd6c2b5f3fc619eaf3dab73aa30fbade302cbce61e10e5c992d15535ed5ce80", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36205, + "estimated_input_tokens": 9052, + "estimated_eval_tokens": 18360 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is still mostly heterogeneous. The two resize_token_embeddings and lm_head-weight-tying pairs are only model-family similarities in the packet, not strong enough to treat as confirmed duplicates or one mergeable fix, so they should remain separate.", + "confidence": 0.52, + "canonical_issue_reason": "Issue 45276 is the clearest anchor because it is a concrete, narrowly scoped resize_token_embeddings propagation bug; 45335 is related but not strong enough evidence of the same exact defect.", + "canonical_pr_reason": null, + "best_issue_reason": "45276 remains the best representative issue because it is the most specific and technically well-scoped item in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 multimodal issues, but one is ZeRO-3 loading for language_model and the other is incorrect visual-encoder save keys; different code paths." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "No concrete shared bug or fix path is visible; they concern different models and different failures." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but the model-specific symptoms point to different underlying paths." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "These are both memory/quantization adjacent, but one is training memory inefficiency and the other is offloading with quantization." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "One is about non-persistent buffers, the other about notebook custom-model initialization; not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory, but the reported failures are different enough that a single fix is not supported by the packet." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and different loading failures; no shared concrete defect is evident." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "These are unrelated loading/conversion failures across different models and subsystems." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a docs example with multiple errors and the other is bad AutoModel/pipeline outputs; not clearly the same bug." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "A Qwen3 MoE compatibility issue is not the same as a resize_token_embeddings propagation bug." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint-resume math and SGD argument handling are separate issues." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "These are different Qwen3-related failures with different symptoms and likely different fixes." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "A model title and a quantized save/load regression do not indicate the same underlying bug." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but the affected models and likely root causes differ." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Saving/loading extra_state and init_empty_weights argument handling are unrelated serialization issues." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different models and different crashes; no strong evidence of one shared fix." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "They share a resize_token_embeddings bug class, but the packet only shows Gemma4 vs T5Gemma model-specific variants, not enough to confirm one bug or one mergeable fix." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash_attention, but one is a Qwen3.5 memory access crash and the other is an AttributeError on s_aux=None; different paths." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab-size mismatch and tokenizer loading failure are different problems." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "They both involve lm_head tying, but the packet only shows Mistral3 vs AyaVision variants; that is not enough to confidently call them the same bug or one fix." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility and meta-device loading errors are distinct regressions." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but the specific failures differ too much to merge as one bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "No meaningful overlap beyond both being model-related issues." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "A Qwen3 device-mapping crash and an AudioFlamingo3 batched inference leak are unrelated." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing with PEFT LoRA and an Apex RMSNorm memory leak are separate problems." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 attention-state rebuilding and Qwen multimodal text_config unwrapping are different initialization bugs." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected _is_hf_initialized handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "GenerationMixin import failure and Whisper pipeline return_language regression are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 reserved-memory OOM and disabling mmap on Strix Halo are different memory issues with different causes." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 FSDP2 KeyError and Qwen3.5 MoE CPU-offload incompatibility are separate bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly grounded, but the two accepted soft edges overstate the evidence. In both accepted pairs, the titles only show a shared bug class across different models, not enough to confidently say they are the same bug or a single mergeable fix. In particular, 45276\u219445335 is Gemma4 vs T5Gemma, and 43404\u219443454 is Mistral3 vs AyaVision; those are plausible similarities, but not conservative enough to mark true from the packet alone. The rest of the verdicts look appropriately cautious." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 8, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:39404", + "issue:42796", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43906", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44292", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44530", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:45084", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45685", + "issue:45698", + "issue:45701", + "issue:45812", + "issue:45823", + "issue:46315", + "issue:46326", + "issue:46468", + "issue:46491", + "issue:46665" + ], + "soft_pairs": [ + "issue:44292|issue:44912", + "issue:43957|issue:44756", + "issue:43335|issue:45663", + "issue:44155|issue:45823", + "issue:42947|issue:43381", + "issue:44863|issue:46468", + "issue:45276|issue:45292", + "issue:45685|issue:46468", + "issue:44479|issue:44877", + "issue:44164|issue:44756", + "issue:43381|issue:44928", + "issue:43854|issue:45216", + "issue:45684|issue:45698", + "issue:45588|issue:45663", + "issue:42796|issue:46326", + "issue:45292|issue:45335", + "issue:43828|issue:43957", + "issue:43646|issue:43708", + "issue:44402|issue:44451", + "issue:43645|issue:44164", + "issue:44589|issue:44756", + "issue:43425|issue:43828", + "issue:45812|issue:46315", + "issue:46315|issue:46491", + "issue:39404|issue:45701", + "issue:43611|issue:43645", + "issue:39404|issue:46665", + "issue:44530|issue:45084", + "issue:44683|issue:45823", + "issue:43425|issue:45823", + "issue:43425|issue:43716", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:43906|issue:44530" + ], + "prepared_review_unit_hash": "d0a88a5bd7125803c3028500bfede436526631cf852999066dad252097fa008c", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37882, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated bug reports. The only clear duplicate family is the resize_token_embeddings/output embedding regression affecting multiple model variants; the rest should stay separate.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45292 is the cleanest, most generic report of the shared resize_token_embeddings/output_embeddings bug, and it aligns with the model-specific variants in the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45292 is the best cluster representative because it names the core code-path problem in model-agnostic terms and sits at the center of the accepted duplicate family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both are quantization/loading regressions, but they involve different models, quantization formats, and failure modes; not the same concrete bug." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading breakage and Strix Halo mmap OOM are different root causes and fixes." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Different model/configuration bugs: sparse-layer construction vs FSDP2 KeyError in Gemma4 attention." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 track leakage and Gemma4 PLE device mismatch are unrelated inference bugs." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA behavior and the other is eval-mode usage; different issues." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "NemotronH checkpoint loading failure and official Nemotron example breakage are related by model family only, not the same bug." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same underlying resize_token_embeddings regression: Gemma4-specific failure to update embeddings matches the generic output_embeddings bug." + }, + { + "left": "issue:45685", + "right": "issue:46468", + "accept": false, + "reason": "MPS histogram unsupported for MoE and a broken Nemotron example are different failures." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression and strict-config granite_speech loading are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling and mmap OOM avoidance are not the same code-path." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode checkpointing and RLHF NaN explosions from 3D position_ids are different bugs." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different model-loading regressions: unit test loading of GLM-4.7-Flash vs Qwen3.5 save_pretrained correctness." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom-model save/load, but one is file permissions and the other is wrong module resolution after save_pretrained." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sink-less crash and Gemma4 FSDP2 KeyError are distinct attention-path issues." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer BPE-marker handling and continue_final_message behavior are different tokenizer regressions." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings bug, with 45335 as a model-specific report that decoder.embed_tokens is not updated." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Both are dtype-related, but one is autocast mismatch and the other is meta-device loading failure." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization in notebooks and Trainer resume max_steps calculation are unrelated." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab-size mismatch and ScandiBERT loading failure are different tokenizer/model issues." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and extra_state save/load handling are separate bugs." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure and mmap OOM avoidance are unrelated runtime problems." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch 2.10 incompatibility and autocast dtype mismatch are not the same concrete bug." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong Granite token IDs and slow convert_tokens_to_ids are separate tokenizer regressions." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "Performance regression in convert_tokens_to_ids and CodeLlama space-roundtrip regression are different issues." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper return_language pipeline regression and version-dependent tokenization changes are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both concern custom models, but one is base_model_prefix loading and the other is notebook initialization semantics." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Whisper pipeline return_language and from_pretrained-in-classes behavior are different bugs." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache group-type crash and compile non-template-nodes error are unrelated." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are different failure modes." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility and Mistral-3 image-preprocessor/model dtype mismatch are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression and mask-shape IndexError are distinct code-path failures." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage error and compile non-template-nodes error are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Isolated reproduction of issue 38071 and PagedAttentionCache crash are not the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet and stays conservative: most pairs are reasonably rejected as different bugs, and the accepted resize_token_embeddings/output_embeddings family is well supported by the titles. The chosen representative issue:45292 is a sensible generic center for that duplicate family." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 9, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:42796", + "issue:42915", + "issue:43299", + "issue:43404", + "issue:43425", + "issue:43452", + "issue:43541", + "issue:43596", + "issue:43611", + "issue:43643", + "issue:43644", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44451", + "issue:44462", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45859", + "issue:46050", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489" + ], + "soft_pairs": [ + "issue:43901|issue:44509", + "issue:43425|issue:44918", + "issue:43994|issue:44079", + "issue:44671|issue:44991", + "issue:44265|issue:44898", + "issue:36032|issue:43452", + "issue:39404|issue:44912", + "issue:44164|issue:45759", + "issue:46050|issue:46326", + "issue:43299|issue:43931", + "issue:43906|issue:46326", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:43596|issue:44683", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:44488|issue:44987", + "issue:43611|issue:43644", + "issue:44462|issue:46489", + "issue:42796|issue:43404", + "issue:36032|issue:42915", + "issue:43716|issue:45084", + "issue:46285|issue:46402", + "issue:32090|issue:39404", + "issue:44737|issue:45859", + "issue:44661|issue:44683", + "issue:43643|issue:44530", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:42915|issue:44530" + ], + "prepared_review_unit_hash": "d61d21249559251a0e8f7783990b87f8cc78d864ecd5fdb0b6d87321dd8bc57c", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36807, + "estimated_input_tokens": 9202, + "estimated_eval_tokens": 18660 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: most items are unrelated model-, tokenizer-, export-, or runtime-specific regressions, so there is no single strong duplicate center. If one issue must represent the cluster, issue 43611 is the broadest loading-regression report.", + "confidence": 0.42, + "canonical_issue_reason": "Issue 43611 is the broadest and most central representative here because it describes a general Transformers 5.0 loading regression affecting models with `base_model_prefix`, which is closer to the common theme than the narrower model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43611 is the best global representative for this cluster since it captures the widest class of failures among the items; the rest are mostly distinct edge cases or model-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but one is about `return_all_scores` and the other about removed pipeline tasks; different documentation issues, not the same bug." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "One is a Torch 2.10 compatibility issue, the other is a TRL SFT trainer/model unpacking failure; no shared code-path bug." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "`siglip2` bad outputs and `ModelOutput` key assignment are different failure modes in different areas." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT prediction regression and tokenizer loading for `EMBIDDIA/est-roberta` are unrelated tokenizer/model issues." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "`torch.export` failing on `torch_compilable_check` is a compiler/export problem; Perceiver non-default resolution is a separate model-shape issue." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer method conflict and `gguf_file` loading breakage are different tokenizer/loading bugs." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline language return regression and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "`extra_state` save/load handling and `from_config` unwrapping of `text_config` are different serialization/configuration issues." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized `sinq` save/load breakage and mistral_common `continue_final_message` behavior are distinct bugs." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a dev0 regression in Moe model loading and the other is a weight-shape mismatch for a specific checkpoint; not the same concrete defect." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "An isolated reproduction of an older issue and a mistral_common prompt-format bug are not the same underlying change." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer failures: loading `est-roberta` vs Mistral regex patching with `fix_mistral_regex=True`." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "One is training memory inefficiency in Qwen3 MoE, the other is mmap-related OOM on Strix Halo; same broad symptom, different causes." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed zero3/BertModel indexing error and compiled flex_attention torch version failure are unrelated runtime/compiler issues." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Save/load `extra_state` and video-input regression in Qwen VL models are separate problems." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading `sleng-bert` and Mistral tokenizer regex patching are different tokenizer paths." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "`sleng-bert` loading failure and `physical-intelligence/fast` loading failure are separate model-loading regressions." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are Transformers 5 loading regressions, but one is about `base_model_prefix` handling and the other about non-persistent buffers; different bugs." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Ignoring repository `tokenizer.json` and DeepSeek-Coder tokenizer output regression are related to tokenizers but not the same defect." + }, + { + "left": "issue:42796", + "right": "issue:43404", + "accept": false, + "reason": "Ministral tokenizer BPE-marker handling and Mistral3 weight tying are different model/tokenizer issues." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer load conflict and Qwen3Moe FP8 config failure are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image-preprocessor dtype mismatch and template compilation error are unrelated." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structure inconsistency and Sam3Model text-encoder loading are both model-loading regressions but on different architectures and code paths." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and Whisper return_language regression do not share a concrete underlying bug." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "XLNet CPU device-placement bug and Qwen3_5Moe CPU-offload incompatibility are different device-management problems." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer-mapping codegen failure and flex_attention torch version failure are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Missing fields with `trust_remote_code=True` and PagedAttention cache crashes on `linear_attention` are different issues." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Different model-loading failures for different checkpoints; no shared code-path evidence." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Grouped_mm tracing failure in Mixtral and template-node compilation failure are different compiler/runtime problems." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Both mention Qwen3-family models, but one is FP8 fine-grained config loading and the other is paged-attention cache group-type handling; not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 10, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43317", + "issue:43377", + "issue:43404", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43582", + "issue:43632", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43746", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44062", + "issue:44206", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45987", + "issue:46178", + "issue:46326" + ], + "soft_pairs": [ + "issue:40990|issue:43421", + "issue:43582|issue:43632", + "issue:43644|issue:43646", + "issue:34567|issue:40444", + "issue:42175|issue:46178", + "issue:44206|issue:44220", + "issue:44811|issue:45584", + "issue:43883|issue:44488", + "issue:40444|issue:46326", + "issue:36032|issue:43645", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:43262|issue:43377", + "issue:43746|issue:46326", + "issue:42915|issue:45084", + "issue:44403|issue:45081", + "issue:43883|issue:44991", + "issue:44464|issue:45987", + "issue:43646|issue:44534", + "issue:43404|issue:44466", + "issue:43645|issue:44534", + "issue:44849|issue:45325", + "issue:43377|issue:43819", + "issue:44451|issue:45081", + "issue:42915|issue:44291", + "issue:43994|issue:44610", + "issue:44987|issue:45081", + "issue:28530|issue:34567", + "issue:36032|issue:43742", + "issue:42915|issue:44589", + "issue:44062|issue:44987" + ], + "prepared_review_unit_hash": "5e64b6a05cf15e73449095944cf32a49920c920a2fc747d3ce14c29b8a83aff7", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37698, + "estimated_input_tokens": 9425, + "estimated_eval_tokens": 19106 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is noisy and mostly contains unrelated bug reports that only share broad Transformer/v5/loading terminology. I did not find any soft pair that looks like the same underlying issue, so all soft-edge candidates are rejected. The broadest representative issue is the generic custom-model initialization regression.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43646 is the broadest, most representative Transformers v5 regression in the set: a generic custom-model initialization breakage, which best matches the cluster\u2019s overall similarity theme.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best single issue to anchor this cluster because it is the most general symptom among the v5 regression reports and is more representative than the highly specific model- or backend-specific failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Different problems: perplexity/eval behavior vs tokenizer backend post-processor updates." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Unrelated failures: Apple Silicon warmup TypeError vs `_is_hf_initialized` flag regression." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 regressions, but they describe different bugs: non-persistent buffer serialization vs custom model initialization." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Trainer token accounting and multimodal iterable-dataset loading are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Both mention backends, but one is packaging extras and the other is an unsupported backend mapping entry." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Same audio area, but different concrete code paths and symptoms." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both involve Whisper, but one is `batch_decode` skipping special tokens and the other is an empty-transcription/alignment failure." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading errors with different missing attributes and failure modes." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Different models and different features: multimodal iterable dataset loading vs `continue_final_message` handling." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer class-name conflict and Jupyter custom-model initialization are separate bugs." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve GGUF loading, but one is a generic API break and the other is a specific Qwen2.5-GGUF regression." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Different loading contexts: device-map dequantized offload vs Qwen2.5-GGUF failure." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error and GraniteMoEHybrid invalid-method call are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading regressions, but they affect different models and likely different tokenizer/model code paths." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask batching are distinct issues." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT local-checkpoint loading and mistral_common `continue_final_message` are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FP8 config loading failure and template compilation failure do not share the same bug." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Unnecessary loading noise and a Mistral regex patch crash are different issues." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "No substantive overlap beyond both being model-loading failures." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency and `StaticCache.get_seq_length()` type mismatch are different code paths." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5-related custom-model regressions, but one is initialization and the other is buffer serialization." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Related to tied weights, but one is untied weights in a model class and the other is device-dependent serialization behavior." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen model bugs with different mechanisms." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI batching/padding and DAC latent/STE behavior are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Different model loading failures with different root causes." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Both touch v5/init behavior, but the failures are different: FP8 config vs unexpected `_is_hf_initialized` argument." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Different models and different symptoms: nonsensical predictions vs image-size mismatch." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different loading bugs for different models/tokenizers." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early stopping configuration and token-count tracking are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer load conflict and MobileLLM key error are different failures." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Both involve newer v5/FP8-related regressions, but the specific errors and affected subsystems differ." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken keyword conflict and model-loading failure for a specific model are not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 11, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41553", + "issue:41720", + "issue:42796", + "issue:43089", + "issue:43296", + "issue:43322", + "issue:43381", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43746", + "issue:43792", + "issue:44079", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44877", + "issue:44991", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45538", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:45987", + "issue:46315", + "issue:46491", + "issue:46710" + ], + "soft_pairs": [ + "issue:43089|issue:45987", + "issue:44265|issue:44610", + "issue:43322|issue:43540", + "issue:44479|issue:45325", + "issue:45216|issue:45357", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:44991|issue:46315", + "issue:43425|issue:44403", + "issue:43296|issue:43531", + "issue:44220|issue:44683", + "issue:43611|issue:44534", + "issue:40444|issue:45325", + "issue:45812|issue:46491", + "issue:44403|issue:44991", + "issue:43632|issue:44756", + "issue:43381|issue:45923", + "issue:43792|issue:44589", + "issue:42796|issue:43746", + "issue:33666|issue:40444", + "issue:44291|issue:44589", + "issue:44079|issue:45663", + "issue:40990|issue:44811", + "issue:42796|issue:46710", + "issue:43482|issue:43525", + "issue:35707|issue:40444", + "issue:43482|issue:45538", + "issue:41553|issue:43582", + "issue:43632|issue:44534" + ], + "prepared_review_unit_hash": "eff243808bda245775a5482ee86bfa192b114af6b302f128646db0275f40be7b", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35649, + "estimated_input_tokens": 8913, + "estimated_eval_tokens": 18082 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated; only the Qwen3.5 `save_pretrained` regression pair looks like a true duplicate. The rest span distinct models, APIs, and failure modes, so they should stay separate.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 45216 is the best canonical representative: it describes the broader Qwen3.5 `save_pretrained` checkpoint regression, and 45357 looks like a narrower follow-up on the same serialization path (incorrect visual encoder keys).", + "canonical_pr_reason": null, + "best_issue_reason": "45216 is the strongest issue candidate overall because it is the broader, earlier regression report and subsumes the narrower 45357 symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Both mention generation/caching behavior, but one is a broad per-token performance complaint and the other is a `get_seq_length()` contract issue; not the same bug." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "`torch.export` with `torch_compilable_check` is unrelated to a processor/model input-size mismatch." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Different models and different failures: Llava Next loading segfault vs Qwen3OmniMoe video-input `ValueError`." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen-VL/video-related, but one is a broad video regression across several models while the other is a still-image `position_ids` scaling bug in Qwen2.5-VL." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same model and same `save_pretrained` serialization path; 45357 reads like a narrower symptom of the broader Qwen3.5 checkpoint corruption reported in 45216." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer method-name conflict is unrelated to the Qwen3 auto-device CUDA assert." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Sliding-window handling for Qwen3-MoE is unrelated to strict config loading for GraniteSpeech." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Both involve tokenizers in v5, but one is a load failure for a specific model and the other is a performance regression in `convert_tokens_to_ids`." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch version incompatibility is unrelated to loading noise during transformer initialization." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM is unrelated to Qwen3-MoE sliding-window behavior." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction failure is unrelated to compiled flex attention on newer torch versions." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "`base_model_prefix` loading failure and non-persistent buffer corruption are distinct v5 regressions with different code paths." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Multi-image IterableDataset finetuning failure is not the same as the still-image temporal position-id scaling bug." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Different tokenizers and different symptom classes: wrong Granite token IDs vs CodeLlama leading-space round-trip regression." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "General loading noise is not the same as a concrete tokenizer load failure." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "`_is_hf_initialized` regression is unrelated to mmap/OOM behavior on Strix Halo." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode gradient checkpointing behavior is not the same as a missing `supports_gradient_checkpointing` flag." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure is unrelated to the Float8 storage-object error." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "Tokenizer BPE-marker handling is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen-VL training issues, but one is multi-GPU training and the other is multi-image prompt handling; not the same bug." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument during init-empty-weights is unrelated to missing Float8 storage registration." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "`ModelOutput` key assignment bug and Gemma-4 FSDP2 `shared_kv_states` `KeyError` are different problems." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on a benchmark model is unrelated to Whisper `batch_decode` ignoring `skip_special_tokens`." + }, + { + "left": "issue:42796", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer BPE-marker handling is unrelated to a downstream accuracy regression in DeepSeek-R1-Distill-Llama-8B." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading under v5 and Llama4 `pad_token_id` missing are distinct model/config loading issues." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to finetuning with multiple images per prompt." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure and CLIPTokenizer `model_max_length` misuse are different tokenizer/model-loading bugs." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "AutoTokenizer error messaging for Voxtral is unrelated to Apple Silicon `caching_allocator_warmup` `TypeError`." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "These are separate v5 regressions affecting different internals: HF init flags vs non-persistent buffers." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall and conservative in the soft-edge calls. The only accepted pair, issue:45216 vs issue:45357, is reasonably supported by the titles as the same Qwen3.5 save_pretrained regression path, and the rest of the rejections stay on the safe side without overclaiming duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 12, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:40444", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43421", + "issue:43454", + "issue:43482", + "issue:43493", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44462", + "issue:44466", + "issue:44534", + "issue:44568", + "issue:44829", + "issue:44877", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45812", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:43531|issue:44291", + "issue:42175|issue:45542", + "issue:43421|issue:44568", + "issue:44829|issue:45479", + "issue:44079|issue:44534", + "issue:41720|issue:45081", + "issue:43818|issue:46402", + "issue:44462|issue:45812", + "issue:43818|issue:45357", + "issue:43611|issue:43646", + "issue:43377|issue:45491", + "issue:43644|issue:44079", + "issue:42175|issue:43296", + "issue:43749|issue:44079", + "issue:44265|issue:45072", + "issue:43792|issue:43957", + "issue:43454|issue:44466", + "issue:45325|issue:45381", + "issue:43493|issue:43994", + "issue:40444|issue:45381", + "issue:45325|issue:45834", + "issue:42915|issue:43482", + "issue:44291|issue:44877", + "issue:43632|issue:45072", + "issue:43873|issue:45701", + "issue:43701|issue:45325", + "issue:42796|issue:45381", + "issue:43296|issue:45800" + ], + "prepared_review_unit_hash": "81a98c3c684600504f736b889d2648128c660c7f77e7009956b62e44545ca7c5", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36010, + "estimated_input_tokens": 9003, + "estimated_eval_tokens": 18262 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; the soft links are largely topical rather than true duplicates. The safest anchor is the high-activity quantization/offloading report #43873, but the only clearly conservative stance on the proposed soft edges is to reject them absent stronger body-level overlap.", + "confidence": 0.74, + "canonical_issue_reason": "#43873 has the strongest discussion/inbound-reference signal and is a broad, representative issue in an otherwise mixed cluster, so it is the best canonical issue anchor for triage.", + "canonical_pr_reason": null, + "best_issue_reason": "#43873 is the highest-signal issue here; it is broad enough to serve as the cluster representative even though the set is not a tight duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different bugs: device_map/offload loading failure vs Qwen3 MoE training memory usage." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Both are v5 regressions, but they affect different code paths and symptoms; not the same concrete bug." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both mention TF/backends, but one is packaging/install coverage and the other is a runtime backend error; not a duplicate." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Related tokenizer/runtime symptoms, but the packet does not establish the same underlying bug; keep this conservative." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Different failure modes: flash_attention_3 training degeneration vs single-label zero-loss behavior." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer junk filling are unrelated regressions." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Different models and code paths: auto device-map cuda assert vs Mistral regex patch crash." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Video-LLaVA temporal-attention/weight issue is unrelated to Sam3Model text-encoder loading." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer-loading and tokenizer-id regression are only loosely related; titles/metadata do not establish a shared concrete bug." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different model-specific regressions: Video-LLaVA tower/temporal attention vs Qwen3.5 visual-encoder save_pretrained keys." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 init/load issues, but they are distinct bugs affecting different initialization behavior." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Missing padding-mask support in MIMI is not the same as Gemma3 mixed-length NaNs." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer junk filling is unrelated to ModelOutput key assignment." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Backend packaging/install issues are unrelated to PaddleOCR-VL model loading in vLLM." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export/torch_compilable_check failure and bf16 dtype mismatch are different problems." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper loading failure and meta-device loading regression are not the same bug." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention lm_head/tied weights, but one is a model-specific tie omission and the other is device-dependent serialization." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "These are related Qwen2.5-VL position-id regressions, but the packet does not establish that they are the same underlying defect." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL position-id scaling and Kosmos2.5 OCR long-input index error are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Qwen3MoE FP8 failure and Qwen2.5-GGUF v5 loading failure are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different regressions." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized handling and bf16 dtype mismatch are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantized offloading behavior and tokenizer changes across versions are different issues." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL position-id scaling." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE-marker handling and Qwen2.5-VL video position ids are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and torch/transformers version incompatibility are not the same bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The triage summary is mostly grounded, but several soft-edge acceptances are too strong for the evidence in the packet. In particular, 43421\u219444568, 44462\u219445812, and 45325\u219445381 are not established duplicates from titles/metadata alone; they look related at best, and the packet\u2019s soft-pair signal is explicitly non-deterministic/negative. Please keep those verdicts conservative (false or maybe-related, not accepted) unless there is stronger body-level evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 13, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:36331", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45584", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315" + ], + "soft_pairs": [ + "issue:44479|issue:45823", + "issue:43606|issue:44265", + "issue:43262|issue:43746", + "issue:40990|issue:45584", + "issue:35707|issue:41720", + "issue:34634|issue:35707", + "issue:43881|issue:44488", + "issue:43454|issue:43550", + "issue:43606|issue:43819", + "issue:39692|issue:40444", + "issue:44186|issue:44265", + "issue:41720|issue:45823", + "issue:46305|issue:46315", + "issue:43377|issue:44610", + "issue:43257|issue:45923", + "issue:44568|issue:45920", + "issue:43454|issue:45072", + "issue:42796|issue:46209", + "issue:43317|issue:43873", + "issue:43454|issue:43720", + "issue:43377|issue:44898", + "issue:40990|issue:42947", + "issue:42617|issue:42915", + "issue:43749|issue:44534", + "issue:43572|issue:45759", + "issue:43322|issue:46050", + "issue:41720|issue:43322", + "issue:43526|issue:43883", + "issue:36331|issue:42617", + "issue:43881|issue:44991" + ], + "prepared_review_unit_hash": "fa3c2e8c5ac34900f39046796677e2e813f7d5264d13399eff03257ac50692a5", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36268, + "estimated_input_tokens": 9067, + "estimated_eval_tokens": 18390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is a grab-bag of unrelated issues across tokenization, loading, quantization, multimodal processors, training, and offloading. None of the soft pairs look like true duplicates or the same concrete fix; they mostly share only broad subsystem keywords. The broadest and most discussion-heavy issue is 43873, but it is still not a duplicate of the others.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43873 is the broadest active report in the set and has the most follow-on discussion/references around offloading/quantization, so it is the best umbrella candidate, though it does not actually duplicate the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the most representative/open issue by scope and activity, but only as a broad umbrella; it is not a true duplicate target for the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Different models and failures: a video-input regression in Qwen VL family vs a Gemma4 PLE device-mismatch during forward." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Bark CPU-offload device mismatch and torch.export failure with torch_compilable_check are unrelated code paths." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are distinct bugs." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Perplexity on gpt-oss-20b and Whisper empty-transcription generation failure do not share the same underlying defect." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Inputs_embeds/past_key_values generation issue is unrelated to Qwen3 auto device-map cudaErrorAssert." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "BarkProcessor voice_preset failure is a different feature path from progressive generation with past_key_values." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they concern different models and likely different tokenizer/model initialization failures." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Weight-tying in AyaVision and torch.compile+SDPA failure in Bamba-9B-v2 are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU-offload device mismatch in Bark-small is not the same as DAC from_latents vs forward mismatch." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "A documentation/example mismatch in SigLIP2 is unrelated to Qwen2.5-VL fine-tuning with iterable multi-image batches." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and torch.export failures are different subsystems and failures." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both mention device_map=auto, but the concrete failures and models differ; not the same bug." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "A sliding-window mask off-by-one is unrelated to tokenizer speed regression." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask behavior and OmDet-Turbo image-size mismatch are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading and a missing gradient-checkpointing flag are distinct issues." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but one is missing BOS/EOS for a specific tokenizer while the other is wrong token IDs across several models; not the same concrete fix." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "lm_head weight tying and bfloat16 dtype mismatches are unrelated." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer BPE-marker handling and FP8 inference on B200 are unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both involve loading/offloading, but one is dequantized model loading with device_map=auto and the other is a broader quantization/offload issue; not clearly the same bug." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision weight tying and BitNet packed-weight unpacking during accelerate loading are different code paths." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI batched-padding-mask bug and Perceiver non-default-resolution failure are unrelated model components." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a language-model eval and ineffective gradient checkpointing with LoRA are different problems." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py execution failure and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU-efficient loading and non-persistent buffer initialization junk are not the same concrete defect." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx in a config and failure to unwrap text_config for composite multimodal configs are distinct." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next segfault on load and sinq quantized save/load breakage are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Different models and symptoms: Qwen3 auto device-map cudaErrorAssert vs Llava Next loading segfault." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and Molmo missing all_tied_weights_keys are unrelated." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss signature mismatch and 3d_parallel.py runtime failure are different issues." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "GLM-4V loading failure and EMBEDDIA/est-roberta tokenizer loading regression are separate model/tokenizer problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 14, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43317", + "issue:43404", + "issue:43450", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:43873", + "issue:44008", + "issue:44186", + "issue:44387", + "issue:44423", + "issue:44466", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46291", + "issue:46402", + "issue:46665" + ], + "soft_pairs": [ + "issue:44387|issue:45704", + "issue:44466|issue:45357", + "issue:41720|issue:44977", + "issue:36331|issue:42915", + "issue:44877|issue:44991", + "issue:43450|issue:45701", + "issue:40990|issue:45701", + "issue:43065|issue:46402", + "issue:44186|issue:45072", + "issue:43873|issue:44756", + "issue:41720|issue:42617", + "issue:26413|issue:28530", + "issue:43550|issue:43720", + "issue:43577|issue:43708", + "issue:43404|issue:45357", + "issue:43295|issue:45701", + "issue:43550|issue:43606", + "issue:43646|issue:46665", + "issue:44186|issue:44610", + "issue:44610|issue:44898", + "issue:44008|issue:45823", + "issue:42915|issue:43317", + "issue:36010|issue:42617", + "issue:44423|issue:45406", + "issue:45632|issue:45684", + "issue:41720|issue:45161", + "issue:28530|issue:33666", + "issue:43531|issue:43572", + "issue:44671|issue:44743", + "issue:42915|issue:46291", + "issue:36331|issue:41720" + ], + "prepared_review_unit_hash": "60a5a47097de6ebc1094993e94a11b6688bd34b67cc03a6b694511c4051fca14", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37112, + "estimated_input_tokens": 9278, + "estimated_eval_tokens": 18812 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most links are superficial topical overlaps (same model family, same subsystem, or same symptom class) rather than true duplicates. I would not merge any of the soft-edge pairs.", + "confidence": 0.93, + "canonical_issue_reason": "No clear canonical issue exists for this cluster; the items span unrelated bugs across training, loading, quantization, processors, serving, and model-specific regressions, so the similarity edges are too weak to justify a single duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the broadest and most central-looking report in the set (generic version/tokenization regression with multiple references), but it still only fits a small subset of the items and is not a true umbrella for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both are memory-related, but one is an int4 quantization/OOM regression and the other is a T5/apex RMSNorm leak; different models and different code paths." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve save/serialization, but one is tied-weight serialization and the other is incorrect visual encoder keys for Qwen3.5; not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Same Qwen family, but multi-GPU training failure and flash-attention generation failure are distinct issues with different failure modes." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "A trainer API compatibility error and a FineGrainedFP8 model failure are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading regressions, but they affect different objects (config vs tokenizer) and different failure causes." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Batched video processor shape handling is unrelated to a broad tokenization-behavior change." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "High perplexity may be influenced by tokenization, but this is not the same concrete regression as a general tokenization change." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Both touch Sam3, but one is a dummy Conv2d in the pixel decoder and the other is missing text encoder weights; different components." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crashes and dtype mismatches are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both mention memory/offload, but one is quantization offloading behavior and the other is mmap-related OOM on a specific platform." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Different features and failures; shared context is only that both are training/runtime problems." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "Checkpoint resumption bug and early-stopping metric lookup bug are separate trainer behaviors." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "One is a torch.compile/SDPA model bug; the other is an accelerate loading issue with packed BitNet weights." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Model dtype loading and trainer max_steps recomputation are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Tied lm_head weights in Mistral3 and Qwen3.5 visual encoder key saving are different model-specific serialization problems." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor/tokenizer API regression is not the same as a general version-caused tokenization change." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are device/compile related, but one is SDPA under compile and the other is CPU offload device mismatch." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Both involve model initialization/loading, but they describe different failure surfaces and likely different fixes." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer padding/truncation crash and image-size mismatch are distinct preprocessing bugs." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both relate to image resolution, but one is a processor/model size mismatch and the other is a Perceiver interpolation failure." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Different Gemma generations and different bugs: variable collision in forward vs PLE device mismatch with auto device mapping." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Both involve quantized/dequantized loading, but the concrete failure modes and models differ." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "Importing GenerationMixin and running 3d_parallel.py are unrelated." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both crash in serving paths, but with different processors, symptoms, and root causes." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "One is a cache-path collision, the other is permission propagation during save_pretrained; not the same bug." + }, + { + "left": "issue:45701", + "right": "issue:45704", + "accept": false, + "reason": "A tokenization regression and a T5 RMSNorm memory leak are unrelated." + }, + { + "left": "issue:45161", + "right": "issue:41720", + "accept": false, + "reason": "Tensor-parallel-only failure for GPT-OSS MoE is not the same as Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and a missing StableLmConfig pad_token_idx are different issues." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM regression and recurrent-state reset in modular_qwen3_5 are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "A model/quantization failure and a kernels package import requirement change are not duplicates." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer loss-signature incompatibility is unrelated to multimodal multi-GPU training." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 15, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28530", + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:44077", + "issue:44186", + "issue:44361", + "issue:44387", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44589", + "issue:44625", + "issue:44913", + "issue:44991", + "issue:45059", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:46402", + "issue:46665" + ], + "soft_pairs": [ + "issue:44186|issue:44361", + "issue:43526|issue:43881", + "issue:43316|issue:44913", + "issue:42915|issue:43322", + "issue:28530|issue:30064", + "issue:33357|issue:35707", + "issue:30064|issue:34634", + "issue:43873|issue:45704", + "issue:43550|issue:43653", + "issue:29159|issue:36032", + "issue:43540|issue:43782", + "issue:44387|issue:45701", + "issue:43606|issue:43720", + "issue:44492|issue:44509", + "issue:44534|issue:46665", + "issue:43531|issue:45759", + "issue:41720|issue:43746", + "issue:42490|issue:43329", + "issue:43388|issue:45059", + "issue:43606|issue:43653", + "issue:45406|issue:45464", + "issue:34567|issue:35707", + "issue:34634|issue:41720", + "issue:45357|issue:46402", + "issue:43531|issue:44589", + "issue:44451|issue:44991", + "issue:45759|issue:45859", + "issue:44077|issue:44625" + ], + "prepared_review_unit_hash": "10bf255cf011f863e35bbcfb6c7b2c36f2e108b724009b69a9e128cd447122d0", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36010, + "estimated_input_tokens": 9003, + "estimated_eval_tokens": 18262 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a loose mix of superficially related Transformers issues (tokenizers, configs, loading/offload, multimodal models), but none of the soft pairs look like the same underlying bug or change. I rejected all soft edges.", + "confidence": 0.66, + "canonical_issue_reason": "`issue:44534` is the broadest representative of the cluster: a Transformers v5 runtime regression affecting core model state handling. Even so, the cluster is too heterogeneous to treat as one true duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:44534` is the most generally framed and potentially high-impact issue among an otherwise unrelated set, so it is the best single issue to stand in as the cluster anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer failures, but they hit different tokenizer classes and different paths: NER/batched padding crash vs task-time AttributeError." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "`reduce_labels` in BeitImageProcessorFast is unrelated to a GLM-4V model loading failure; different component and failure mode." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config-related, but one is a Gemma3TextConfig API mismatch and the other is GPTNeoX rotary_pct not persisting on reload\u2014different fields, different bug." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated model-loading/runtime problems." + }, + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early stopping metric selection and void segmentation map processing are completely separate features." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error with a CLIP model vs generation with inputs_embeds/past_key_values; no shared code path." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image processor void segmentation maps and BarkProcessor voice_preset are unrelated processor bugs." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and an apex RMSNorm memory leak are different runtime issues with different root causes." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile + SDPA on Bamba-9B-v2 is unrelated to BigBirdTokenizer special-token registration/empty decode behavior." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is a different bug from T5Tokenizer's add_special_tokens method conflict during load." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "One is a video-input processing error in Qwen3OmniMoe; the other is a Qwen3VL from_pretrained weight_only loading error." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "CUDA reserved-memory regression under int4 quantization is not the same as a tokenization change across versions." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offload, but one is a CPU-offload device mismatch for Bark and the other is missing unpacking of packed BitNet weights during accelerate loading." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategy typo and stale docs mentioning removed pipeline tasks are both doc-level issues, but not the same defect." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer initialization junk in v5 and `from_pretrained` behavior inside Python classes are separate problems with different mechanics." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and unwrapping `text_config` in composite Qwen configs are different config/code-path bugs." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Auto device mapping CUDA assert on A800 is unrelated to GraniteSpeech PEFT local-checkpoint loading failures." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "A trailing-slash import warning is not the same as an undefined-function bug in the video branch of multimodal token counting." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Dropping tuple-label elements in `gather_for_metrics` is unrelated to SAM3 text/bounding-box behavior." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Bark CPU-offload device mismatch and BigBirdTokenizer special-token registration are different bugs in different subsystems." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing `_tokenizer` and Qwen3.5 streaming chat/completions failure are distinct API/runtime issues." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "`num_input_tokens_seen` not updating is a TrainerState bookkeeping bug, not a generation bug involving embeds and cache state." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset failure and Qwen3 auto device mapping cudaErrorAssert are unrelated." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Incorrect Qwen3.5 visual-encoder keys on save and Sam3Model text-encoder weights not loading are both multimodal, but they affect different models and directions (save vs load)." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "A policy change around optional `post_init` and Qwen3.5 `num_labels` propagation are separate config-validation issues." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and missing Float8 storage objects are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Loading `vesteinn/ScandiBERT` and tokenizer loading for EMBEDDIA/est-roberta are both model-loading reports, but they do not share the same failure mode or code path." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both are Qwen3.5/3.6-related config bugs, but one is about `from_config` unwrapping and the other is CPU-offload compatibility in rotary embedding forward; different fixes." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "`post_init` allowance and `num_labels` propagation are different configuration semantics issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 16, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43644", + "issue:43645", + "issue:43653", + "issue:43746", + "issue:43782", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44186", + "issue:44368", + "issue:44451", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45704", + "issue:45759", + "issue:45907", + "issue:45923", + "issue:46032", + "issue:46050", + "issue:46463", + "issue:46655", + "issue:46665", + "issue:46693" + ], + "soft_pairs": [ + "issue:44625|issue:45479", + "issue:43644|issue:46665", + "issue:44062|issue:44451", + "issue:44898|issue:45072", + "issue:44987|issue:45042", + "issue:44368|issue:45357", + "issue:43381|issue:43844", + "issue:39692|issue:44610", + "issue:41720|issue:45704", + "issue:43421|issue:43927", + "issue:46655|issue:46693", + "issue:45704|issue:45923", + "issue:43381|issue:45704", + "issue:43531|issue:43645", + "issue:43746|issue:46050", + "issue:44918|issue:44936", + "issue:42886|issue:44843", + "issue:45759|issue:45923", + "issue:44514|issue:44545", + "issue:46032|issue:46463", + "issue:39692|issue:43873", + "issue:43746|issue:43873", + "issue:43994|issue:44671", + "issue:43421|issue:43653", + "issue:30064|issue:36010", + "issue:43421|issue:44186", + "issue:43782|issue:45381", + "issue:45081|issue:45356", + "issue:45356|issue:45907" + ], + "prepared_review_unit_hash": "104aa2775cad7eae59b90ace02052b55298a590ff35f5846e5ab2d32a1544215", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36804, + "estimated_input_tokens": 9201, + "estimated_eval_tokens": 18658 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated and should not be clustered together. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False (44514/44545).", + "confidence": 0.94, + "canonical_issue_reason": "issue:44514 is the best canonical issue because it is the older, clearer report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash; 44545 is a near-verbatim duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44514 is the strongest representative issue in the set since it matches the one true duplicate pair and has the cleaner original description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both are classification-label bugs, but they concern different model families and different failure modes." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Unrelated bugs: non-persistent buffer initialization vs from_pretrained behavior inside Python classes." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer AddedToken argument handling and loading a specific model are different issues." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different subsystems and symptoms: Perceiver image resolution handling vs bfloat16 dtype mismatches." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Model loading regression and PIL/torchvision dependency behavior are not the same bug." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Tie_word_embeddings warning is unrelated to incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction is distinct from ZeRO-3 gradient inflation." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Docs/example mismatch and quantization failures are different from OmDet-Turbo processor image-size mismatch." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Auto device mapping CUDA assert and T5 Apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both involve tokenizers, but one is runtime post-processor updates and the other is save/load of custom token IDs." + }, + { + "left": "issue:46655", + "right": "issue:46693", + "accept": false, + "reason": "Video padding calculation and long-context FA2 performance regression are different problems." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "Memory-leak behavior in T5 is unrelated to a missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "These are different training/runtime bugs with no shared concrete code path." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to Jupyter custom model initialization." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "Local checkpoint loading with PEFT adapters is unrelated to sinq quantization save/load breakage." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "TRL SFT embedding unpacking and trainer.evaluate after trainer.train are different trainer issues." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention HF_HUB_OFFLINE, but one is general tokenizer cache loading and the other is a specific Mistral regex patch calling model_info(); not clearly the same bug." + }, + { + "left": "issue:45759", + "right": "issue:45923", + "accept": false, + "reason": "Composite Qwen config handling and a missing capability flag are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash; 44545 is effectively a duplicate of 44514." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "Mamba2Mixer cache correctness and T5 decoder input validation messaging are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 documentation errors and offloading with quantization are different issues." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT local checkpoint loading and quantization/offloading behavior are not the same underlying bug." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different model families and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked LM prediction regression." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both touch special tokens, but one is runtime post-processor syncing and the other is BigBirdTokenizer decode output registration." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image processor void segmentation maps and a GenerationMixin import error are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Tokenizer post-processor updates and LayoutLMv2Tokenizer NER/padding crashes are different code paths." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL video vision_position_ids are separate multimodal bugs." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer regressions, but one is an AttributeError in fix_mistral_regex and the other is a Kimi-K2.5 codec/warning regression." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression and list_repo_files revision forwarding are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary matches the packet\u2019s lone deterministic duplicate pair (44514/44545), and the other soft-edge rejections stay conservative by distinguishing different titles, model families, or failure modes without overclaiming. The canonical choice of 44514 is also supported by the earlier timestamp and clearer wording." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 17, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43335", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43611", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43927", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44442", + "issue:44448", + "issue:44485", + "issue:44488", + "issue:44509", + "issue:44661", + "issue:44671", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45381", + "issue:45584", + "issue:45834", + "issue:46294", + "issue:46665", + "issue:46693" + ], + "soft_pairs": [ + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:43295|issue:45834", + "issue:43596|issue:44220", + "issue:43550|issue:45072", + "issue:43550|issue:43927", + "issue:44062|issue:44488", + "issue:43540|issue:46294", + "issue:45381|issue:46693", + "issue:43618|issue:43697", + "issue:43531|issue:43643", + "issue:43611|issue:46665", + "issue:44220|issue:44661", + "issue:43653|issue:43819", + "issue:44186|issue:44898", + "issue:30064|issue:36331", + "issue:43643|issue:44877", + "issue:43335|issue:43756", + "issue:41720|issue:44991", + "issue:43550|issue:43819", + "issue:44843|issue:45081", + "issue:44821|issue:45084", + "issue:43540|issue:44008", + "issue:39692|issue:45834", + "issue:44186|issue:44442", + "issue:44485|issue:44509", + "issue:44448|issue:44671", + "issue:43644|issue:43749", + "issue:45584|issue:45834", + "issue:34634|issue:36010" + ], + "prepared_review_unit_hash": "d4254d125a320477cf7f806e4ce2dce4633906b055930bea508130186bc2df80", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36749, + "estimated_input_tokens": 9188, + "estimated_eval_tokens": 18632 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is highly heterogeneous, with most soft pairs describing distinct model-, tokenizer-, or loading-specific bugs rather than true duplicates. The closest common thread is Transformers v5 regression/loading behavior, so 43611 is the best broad anchor, but no pair is strong enough to merge.", + "confidence": 0.42, + "canonical_issue_reason": "issue:43611 is the broadest, most representative v5 loading-regression report in the set; it best matches the small subset of config/load-related issues, though the cluster overall is mixed.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the best triage anchor because it is user-facing and broadly about model loading regressions, making it more representative than the narrower model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different bugs: packed-weight loading in BitNet vs DAC latent decoding/STE mismatch." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug vs accelerate loading of packed weights; unrelated code paths." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal, but one is a processor/tokenizer regression and the other is a long OCR input index error in a different model." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed ZeRO3/BertModel init failure is unrelated to audio fbank extraction." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different failure modes: torch.compile/SDPA in Bamba vs bfloat16 dtype mismatch in different models." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba compile/SDPA bug is unrelated to DiaConfig token ID persistence and generation index errors." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken keyword duplication is unrelated to tokenizer loading for a specific model." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both concern video input, but they describe different Qwen model families and different mismatches." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "Qwen2.5-VL vision_position_ids bug is unrelated to a performance regression after 5.2.0." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both are v5 regressions, but one is missing CLIP attentions and the other is RT-DETR output drift; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to AutoConfig missing fields with trust_remote_code." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Both mention from_pretrained/loading, but one is a base_model_prefix regression and the other is inconsistent class-wrapper behavior." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction bug vs tokenizer mapping failure when adding a new model-like entry." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DAC latent decoding are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer crash on NER padding/truncation is unrelated to Perceiver image classification with non-default resolution." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image processor void-segmentation failure is unrelated to CustomTrainer.compute_loss signature changes." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Missing fields from trust_remote_code configs is different from strict config blocking granite_speech loading." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction bug and Smollm3 RoPE-layer dropping are different model/config issues." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DAC forward/from_latents mismatch." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper function, but different bugs: offline model_info call vs backend_tokenizer attribute error." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading failure is unrelated to template compilation TypeError." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different multimodal code paths: Qwen3OmniMoe video processing vs Gemma3n audio_mel_mask variable collision." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc example errors are unrelated to Kosmos2.5 long OCR index errors." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to AutoTokenizer failing on FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipeline tasks." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Different model regressions: Pegasus output drift vs CamemBERT masked LM prediction errors." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer junk in v5 is not the same as FSDP CPU RAM efficient loading being broken." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription failure after token alignment is unrelated to Kosmos2.5 OCR length handling." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice preset issue is unrelated to the GenerationMixin import error." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 18, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:42175", + "issue:43322", + "issue:43388", + "issue:43421", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43645", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43825", + "issue:43827", + "issue:43950", + "issue:43994", + "issue:44008", + "issue:44291", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45685", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46032", + "issue:46192", + "issue:46291", + "issue:46294", + "issue:46463", + "issue:46512", + "issue:46665" + ], + "soft_pairs": [ + "issue:43322|issue:45685", + "issue:43606|issue:45072", + "issue:45183|issue:45479", + "issue:43716|issue:44008", + "issue:43388|issue:46032", + "issue:44479|issue:45834", + "issue:43950|issue:45987", + "issue:43825|issue:43994", + "issue:44991|issue:46291", + "issue:43582|issue:46512", + "issue:43577|issue:43818", + "issue:46294|issue:46463", + "issue:43782|issue:44977", + "issue:44561|issue:45020", + "issue:43421|issue:44811", + "issue:44805|issue:44936", + "issue:43540|issue:45381", + "issue:44610|issue:45072", + "issue:43825|issue:43827", + "issue:43720|issue:45072", + "issue:28218|issue:29159", + "issue:43749|issue:44291", + "issue:43645|issue:46665", + "issue:45987|issue:46192", + "issue:44671|issue:45701", + "issue:43577|issue:43582", + "issue:42175|issue:45081" + ], + "prepared_review_unit_hash": "81e26f8591ecab93bddb46e33f003f90182a5341eee6776de92f91a138e20ebd", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36320, + "estimated_input_tokens": 9080, + "estimated_eval_tokens": 18416 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a thematically related but heterogeneous mix of Transformers issues: tokenizer regressions, model loading/dtype problems, multimodal/video bugs, and a few v5 breakages. I do not see any soft pair that clearly describes the same concrete bug or change, so I would keep them separate.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set for version-related remote_code breakages, and is the closest thing here to a cluster anchor; however, the overall set is still too mixed for a strong duplicate canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best representative of the main recurring theme (recent-version regressions affecting model loading/execution), even though many other items are only loosely related and not duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Different failures in different subsystems: Llava Next loading segfault vs MPS histogram on Int in moe." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both are runtime/backend mismatches, but they involve different models and distinct code paths (CPU offload vs bfloat16 dtype handling)." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "One is about input-validation error messaging; the other is a training loss bug for single-label classification." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Distinct multimodal bugs: image preprocessor/model dtype mismatch vs an AttributeError from a variable-name collision." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Different core problems: metric gathering drops labels in the last batch vs Mamba2 cache generation producing wrong results." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Both are vision/multimodal issues, but the concrete failures are unrelated (Qwen video-input regression vs Kosmos2.5 OCR index error)." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "From-pretrained non-persistent buffer corruption is a loading regression; StaticCache.get_seq_length is a return-type/contract bug." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an incorrect pipeline error message; the other is a model producing nonsensical outputs via AutoModel/pipeline." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading failure for one model vs importing transformers.pipelines failing due to kernels version requirements." + }, + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Both mention dtype-ish behavior, but one is an Apple Silicon warmup TypeError and the other is stale config dtype after from_config()." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype persistence issue vs Video-LLaVA architecture/weight-sharing problem; not the same bug." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen3.5-9B video token/feature mismatch is unrelated to T5 decoder-input validation errors." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Both are Qwen-related, but one is a weight_only loading error and the other is a flash-attention generation bug; different code paths." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "44561 is one specific remote_code breakage caused by removing is_torch_fx_available; 45020 is a broader report of recent-version remote_code failures, not the same concrete bug." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Special-token post-processor updates and Whisper batch_decode skip_special_tokens are separate tokenizer/decoder behaviors." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "A tensor shape IndexError is not the same issue as trainer.evaluate() failing after trainer.train()." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs, but the reported mismatches differ (processing ValueError vs wrong vision_position_ids)." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch for OmDet-Turbo vs dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Runtime error-message regression vs stale docs after pipeline removals; related topic, but not the same underlying bug/change." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight loading bug in BitNet is unrelated to generic bfloat16 dtype mismatch issues." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "The first is specifically about an added-token spacing bug; the second is a broader slow-vs-fast tokenizer inconsistency report." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading breakage and init_empty_weights rejecting _is_hf_initialized are different loading-path regressions." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Jupyter custom-model initialization in v5 vs from_pretrained inside Python classes are different usage patterns and failures." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache length contract bug and return_tensors being ignored when text_kwargs is passed are unrelated API issues." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression and generic tokenization changes in new versions are too broad to be the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype mismatch and AppleSilicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Packaging/backend dependency issue vs tokenizer regex patch crash; no shared code-path bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 19, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:36331", + "issue:42175", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43582", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43782", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44991", + "issue:45003", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45538", + "issue:45600", + "issue:46209", + "issue:46285", + "issue:46291", + "issue:46326", + "issue:46697", + "issue:46728" + ], + "soft_pairs": [ + "issue:34634|issue:36331", + "issue:45003|issue:46285", + "issue:46697|issue:46728", + "issue:43653|issue:44912", + "issue:43262|issue:43819", + "issue:43582|issue:44857", + "issue:45081|issue:45406", + "issue:43742|issue:45020", + "issue:44589|issue:45042", + "issue:43782|issue:45325", + "issue:45372|issue:46326", + "issue:44060|issue:44368", + "issue:45084|issue:45538", + "issue:43550|issue:44912", + "issue:43329|issue:45600", + "issue:43526|issue:43697", + "issue:43452|issue:44991", + "issue:43526|issue:44554", + "issue:43257|issue:43329", + "issue:36010|issue:36331", + "issue:43906|issue:46209", + "issue:36010|issue:43645", + "issue:43819|issue:43927", + "issue:43531|issue:45042", + "issue:43638|issue:44936", + "issue:42175|issue:46291", + "issue:43720|issue:44610", + "issue:42490|issue:43257" + ], + "prepared_review_unit_hash": "f1b8bdf905d98ec4de92b88ae407f0c6e2c4763e614a8991b51000f36d0aa2ba", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft pairs mostly share only broad subsystem or keyword similarity, not the same concrete bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated bugs in tokenizers, multimodal processors, training, quantization, and backend/package compatibility.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is clearly better than the rest as a cluster representative because the set is not a true duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss are unrelated failures in different code paths." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "sys.modules access in modeling_utils and CLIP text-model structure inconsistency are different regressions." + }, + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Both are image-processor bugs, but one is Swin2SR over-padding and the other is Aria patch counting; different logic." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate handling and DAC latent reconstruction are separate bugs." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon caching allocator TypeError and CUDA float16 AMP crash are different backend-specific issues." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both involve processor/tokenizer loading, but one is Mistral regex patching and the other is Gemma4 missing _tokenizer; not the same defect." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A key error loading one model is not the same as the broad remote_code breakage described in the other issue." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure and PIL/torchvision dependency regression are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL rope index scaling are different model-specific problems." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both mention mistral_common, but one is an import error and the other is continue_final_message behavior; different root causes." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "They both report tied-weights warnings, but the models and warning cause differ; not the same underlying bug." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation error and CLIPTokenizer max_length issue are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to MXFP4 loading fallback in git-oss-20b." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video token helper variables and stale auto_mappings config references are different code paths." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output drift are not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file handling and tokenizer loading for EMBEDDIA/est-roberta are separate from-pretrained failures." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "A label-processing bug in BeitImageProcessorFast is unrelated to the MPS attention correctness issue." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed and multimodal token counting are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "GenerationMixin import failure and CustomTrainer compute_loss signature mismatch are different API breaks." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of an external issue and FP8 inference failure on B200 are not the same defect." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "GenerationMixin import error and Jupyter custom-model initialization breakage are different compatibility issues." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent/forward mismatch and DiaConfig token-ID persistence loss are unrelated model/config bugs." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding window logic and PIL backend torchvision dependency regression are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3 indexing failure with a non-pretrained Bert model and evaluate-after-train behavior are distinct trainer issues." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Torch-only install missing TensorFlow backend and kernels v0.15 import failure are different packaging/import problems." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading and OmDet-Turbo input-size mismatch are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "A trailing-slash local-path warning and Qwen3 MoE conversion failure do not share the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 20, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:29159", + "issue:34567", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43454", + "issue:43525", + "issue:43540", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43696", + "issue:43742", + "issue:43746", + "issue:43749", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44387", + "issue:44479", + "issue:44568", + "issue:44610", + "issue:44617", + "issue:44756", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44977", + "issue:44987", + "issue:45042", + "issue:45356", + "issue:45459", + "issue:45538", + "issue:45759", + "issue:45803", + "issue:45859", + "issue:46291", + "issue:46421", + "issue:46424", + "issue:46701" + ], + "soft_pairs": [ + "issue:39692|issue:44912", + "issue:43746|issue:44811", + "issue:38175|issue:44610", + "issue:43643|issue:45759", + "issue:43854|issue:44977", + "issue:44387|issue:44617", + "issue:43257|issue:43696", + "issue:43262|issue:44811", + "issue:43257|issue:43931", + "issue:46421|issue:46424", + "issue:45459|issue:45803", + "issue:42915|issue:43742", + "issue:43329|issue:44479", + "issue:43329|issue:43450", + "issue:43749|issue:44756", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:34567|issue:40990", + "issue:43606|issue:45859", + "issue:43650|issue:43906", + "issue:43540|issue:46701", + "issue:43927|issue:44912", + "issue:29159|issue:34567", + "issue:43525|issue:46291", + "issue:43454|issue:43577", + "issue:44987|issue:45538", + "issue:44568|issue:45356" + ], + "prepared_review_unit_hash": "f1a65433d519d521c25ec53332740f6db5151d39b529dc37267f0abad4e657f5", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35889, + "estimated_input_tokens": 8973, + "estimated_eval_tokens": 18202 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: it mixes tokenizer, model-loading, processor, cache, quantization, and offload bugs that do not describe the same underlying defect. No soft pair looks like a true duplicate/merge candidate.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 is the most prominent and broadly impactful item here (highest discussion/inbound reference signal), so it is the closest thing to a representative issue, but the set is too diverse for a real canonical duplicate anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Different problems: SigLIP2 doc example errors vs MXFP4 quantization loading/fallback behavior." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT local-checkpoint loading bug vs Whisper batch_decode skip_special_tokens handling; unrelated code paths." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 vs OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both are config-loading issues, but one is missing fields from trust_remote_code and the other is composite Qwen unwrapping; not the same bug." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7 unit-test load failure vs Qwen3.5 flash-attention generation regression; different failures." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is int4 reserved-memory regression and the other is a model-specific video OOM." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MOE weight conversion/loading issue vs a GPT-oss CUDA OOM; not the same underlying bug." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs Whisper batch_decode option handling; different processor behaviors." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MOE conversion bug vs Qwen3-VL shape mismatch on load; separate model/load paths." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both cache-related, but one is lazy-init recompilation and the other is max_cache_length scope; related area, different bugs." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both are error-swallowing patterns, but they affect different tensor/conversion paths and are not the same defect." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe vs key error loading MobileLLM-125M; unrelated model-load issues." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Undefined video token-counting helpers vs a broad v5.3.0 video input regression across models; not one concrete bug." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Video token counting code bug vs batched video processor output shape bug; separate failure modes." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU RAM efficient loading vs disabling mmap on Strix Halo; both memory-related but different fixes." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights argument incompatibility vs PIL image processors incorrectly requiring torchvision; unrelated subsystems." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3Moe FP8 loading problem vs AutoImageProcessor URL loading failure; no shared code-path bug." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "TrainerState token counter not updating vs high perplexity on GPT-oss; unrelated training/eval issues." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "CPU offload device mismatch for bark-small vs Qwen3_5Moe rotary embedding offload incompatibility; different models and code paths." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "'ADD THE DATA' is not a technical bug report, and the other is an isolated reproduction of a different issue." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError vs DiffusionGemma image-text pipeline failure; different models and surfaces." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig save/load token-ID loss vs MXFP4 quantization fallback; unrelated configuration vs quantization behavior." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency vs TrainerState input-token accounting; no shared bug." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Llama4Config missing pad_token_id vs kernels v0.15 import/version requirement; unrelated failures." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head tying bug vs BLIP2 dtype propagation issue; different model internals." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Physical-intelligence/fast loading regression vs CLIPTokenizer max_length bug; different tokenizer/load issues." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "mdeberta add_special_tokens regression vs Kimi-K2.5 codec/regex regression; separate tokenizer bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 21, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36331", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43317", + "issue:43404", + "issue:43425", + "issue:43452", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44568", + "issue:44821", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:44991", + "issue:45127", + "issue:45137", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45907", + "issue:46291", + "issue:46512", + "issue:46693" + ], + "soft_pairs": [ + "issue:28282|issue:34689", + "issue:43404|issue:45127", + "issue:43697|issue:45357", + "issue:43425|issue:46291", + "issue:42915|issue:45538", + "issue:45137|issue:45405", + "issue:33357|issue:34567", + "issue:42175|issue:43824", + "issue:43742|issue:44291", + "issue:28282|issue:29127", + "issue:44479|issue:46693", + "issue:43317|issue:45538", + "issue:44062|issue:44991", + "issue:44849|issue:45701", + "issue:42915|issue:43749", + "issue:43577|issue:46512", + "issue:43452|issue:43582", + "issue:41720|issue:43540", + "issue:27957|issue:28282", + "issue:24643|issue:29159", + "issue:43540|issue:44560", + "issue:42175|issue:44488", + "issue:44291|issue:44821", + "issue:44987|issue:45907", + "issue:43540|issue:44479", + "issue:44479|issue:45381", + "issue:44568|issue:45488", + "issue:38175|issue:44898", + "issue:34567|issue:36331" + ], + "prepared_review_unit_hash": "2ca7c5b4e7c785b01b128214d802d8af96eb4d2223dc90db3c5fae9caa10badc", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35839, + "estimated_input_tokens": 8960, + "estimated_eval_tokens": 18176 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly unrelated transformer issues with only broad subsystem overlap (tokenizers, Qwen, model loading, dtype, training). No soft pair looks like the same concrete bug/change, so I would keep them separate. No PRs are present.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 45488 is the most representative cluster anchor: it is detailed, currently open, and describes a broadly impactful tokenizer regression with inbound references, making it the best umbrella issue in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45488 best fits as the cluster\u2019s canonical issue because it is specific enough to be actionable while still broad in impact, and its discussion/inbound references suggest it is the most central/visible report here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Both are model-loading failures, but one is a PyTorch-missing import error and the other is a Llama 3.2 Vision loading regression; different causes and code paths." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both mention tied embeddings, but one is about Mistral3 lm_head weight tying and the other is LoRA merge collapse with extended vocab; not the same bug." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output mismatch and Qwen3.5 save_pretrained visual-key regression are separate model-specific failures." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch version incompatibility and kernels import requiring revision/version are unrelated issues." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and CLIPTokenizer max_length bug are unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error and a PEFT minimum-version bump are not the same underlying problem." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP loading and TrainerState token counter not updating are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging omission and missing Qwen2.5-VL class import are different failures." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM key-loading error and init_empty_weights unexpected argument error are distinct load-path bugs." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch import failure and LayoutLMv3 box-validation clarity issue are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Both involve vision/FA2-adjacent transformer regressions, but one is a video-input functional regression and the other is a long-context performance regression; not the same bug." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "device_map/offload loading failure and CLIPTokenizer max_length bug are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AddedToken argument duplication bug and the other is a model-specific tokenizer load failure; insufficient evidence of same root cause." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-state output bug and a general tokenization version-change report are not the same concrete issue." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "FP8 config failure and FSDP CPU RAM efficient loading breakage are separate training/inference paths." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both concern dtype handling, but one is Blip2/qformer weights staying float32 and the other is config.dtype being stale after from_config; different bugs." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "gguf_file loading breakage and Apple Silicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert and Qwen3OmniMoe video-input ValueError are different failure modes." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 with XLMRoberta is unrelated to a missing PyTorch import error." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training runtime error and tokenizer slow/fast inconsistency are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are Qwen video-input problems, but they involve different models and different errors (ValueError vs StopIteration), so they do not look like the same bug." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Packaging/backend omission and a specific model-loading failure for sleng-bert are not the same issue." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument during empty-weight loading and URL image-processor loading are separate code paths." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Model loading failure under transformers>=5.1.0 and a missing revision-forwarding bug in list_repo_files are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Same general Qwen video area, but different models and different regressions; too broad to treat as one duplicate." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": false, + "reason": "Both affect Qwen video inputs, but one is a broader regression across models and the other is specifically wrong vision_position_ids in qwen2.5-vl; not the same concrete bug." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but one is mdeberta special-token insertion and the other is LlamaTokenizer pre-tokenizer override; different root causes." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 and Perceiver interpolation failure are unrelated vision-model bugs." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token counting and custom compute_loss signature mismatch are separate Trainer API issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 22, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:32090", + "issue:43257", + "issue:43404", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43577", + "issue:43697", + "issue:43716", + "issue:43742", + "issue:43756", + "issue:43906", + "issue:44062", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44877", + "issue:44913", + "issue:44964", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45823", + "issue:45920", + "issue:45987", + "issue:46121", + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491", + "issue:46693", + "issue:46701" + ], + "soft_pairs": [ + "issue:43257|issue:43756", + "issue:44792|issue:44964", + "issue:44062|issue:45081", + "issue:43742|issue:44589", + "issue:43756|issue:46399", + "issue:45042|issue:45081", + "issue:43716|issue:46701", + "issue:44737|issue:46463", + "issue:43756|issue:45325", + "issue:44568|issue:46491", + "issue:43531|issue:44821", + "issue:45405|issue:45600", + "issue:43906|issue:46468", + "issue:43577|issue:45072", + "issue:45823|issue:46463", + "issue:43452|issue:44488", + "issue:43906|issue:46693", + "issue:43404|issue:45987", + "issue:44737|issue:45588", + "issue:44625|issue:44877", + "issue:43526|issue:45479", + "issue:45357|issue:45920", + "issue:44913|issue:46121", + "issue:43540|issue:46693", + "issue:28282|issue:32090", + "issue:24643|issue:29127", + "issue:43697|issue:45216" + ], + "prepared_review_unit_hash": "3912f437c7cc66de466c2ff82a5a357fdd76e164634b5d808510c69cdf948f85", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35594, + "estimated_input_tokens": 8899, + "estimated_eval_tokens": 18054 + }, + "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 different models, tokenizers, configs, and runtime paths. The soft-similarity links do not look like true duplicates, so I reject all of them and see no strong canonical duplicate target.", + "confidence": 0.97, + "canonical_issue_reason": "There is no single true canonical issue here because the items span unrelated subsystems and model families; if one representative must be chosen, issue:45920 is the broadest current regression report.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue:45920 is the most representative as a broad, active regression affecting multiple models/tokenizers; however, the set is not a coherent duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different models and different failure modes: Qwen3 MoE weight conversion vs Smollm3 RoPE layer dropping." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "A Janus test failure and a Phi-4 multimodal load regression are not the same underlying bug." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AddedToken argument conflict and the other is a Mistral regex/backend_tokenizer crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated issues: MobileLLM key error versus missing Float8 storage object." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Different Qwen/Smollm model-specific rotary/kernel problems; not the same code-path bug." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image-processor torchvision dependency regression is unrelated to Mistral tokenizer regex patching." + }, + { + "left": "issue:43716", + "right": "issue:46701", + "accept": false, + "reason": "Different multimodal models and different symptoms: dtype mismatch versus pipeline failure." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "CPU device placement bug in XLNet is unrelated to T5 decoder-input validation messaging." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both involve RoPE-adjacent topics, but they concern different Qwen variants and different logic errors." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Tokenizer regressions, but different models and different behaviors: BOS/EOS omission versus leading-space loss." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to loading AutoImageProcessor from URL." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "Both affect CI/runtime setup, but one is a dependency version bump and the other is stale auto-mapping references." + }, + { + "left": "issue:43906", + "right": "issue:46468", + "accept": false, + "reason": "A reproduction of an older issue and a Nemotron example breakage are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but they affect different models and different forward/inference paths." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "A device-map mismatch in Gemma4 PLE is unrelated to T5 decoder input validation." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading-related, but the evidence is too weak to treat them as the same gguf/model-loading bug." + }, + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "An isolated reproduction of one issue is unrelated to a long-context FA2 performance regression." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Weight tying in Mistral3 is unrelated to StaticCache sequence-length return type." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU tensor creation and flash_attention s_aux=None crashes are different code paths." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "A num_labels propagation bug and a strict-config loading failure are not the same issue." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "ImageProcessor label reduction and sequence-classification loss degeneracy are different bugs." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "save_pretrained visual-encoder key regression is unrelated to tokenizer ID corruption." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both mention rope/config handling, but one is silent default reversion on reload and the other is a TypeError in validation." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "Video input processing failure in Qwen3OmniMoe is unrelated to FA2 long-context performance regression." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import and _gpu_broadcast_one NoneType handling are completely different failures." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed training weight-shape runtime error is unrelated to a LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output changes and Qwen3.5 save_pretrained regression are different symptoms and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 23, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:33357", + "issue:35141", + "issue:38175", + "issue:41762", + "issue:42898", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43819", + "issue:43994", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44493", + "issue:44568", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44898", + "issue:44936", + "issue:45005", + "issue:45200", + "issue:45356", + "issue:45372", + "issue:45375", + "issue:45542", + "issue:45704", + "issue:45823", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46463", + "issue:46491", + "issue:46701" + ], + "soft_pairs": [ + "issue:43720|issue:44898", + "issue:28282|issue:35141", + "issue:43643|issue:45375", + "issue:44466|issue:45005", + "issue:45200|issue:46463", + "issue:41762|issue:44936", + "issue:44361|issue:44442", + "issue:43688|issue:43749", + "issue:43452|issue:44843", + "issue:44493|issue:46050", + "issue:43653|issue:44361", + "issue:43819|issue:44811", + "issue:43525|issue:46701", + "issue:43994|issue:44898", + "issue:38175|issue:43295", + "issue:45372|issue:46170", + "issue:44936|issue:45823", + "issue:43531|issue:43742", + "issue:45542|issue:46178", + "issue:42898|issue:46491", + "issue:43646|issue:43749", + "issue:43653|issue:44568", + "issue:43295|issue:45356", + "issue:43577|issue:43697", + "issue:29159|issue:33357", + "issue:44464|issue:44743", + "issue:43452|issue:43577", + "issue:43688|issue:45704" + ], + "prepared_review_unit_hash": "1343194a1e5e3b428bdf67fd01ae3da37889f62eb8747b2090cbd8811c236608", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35905, + "estimated_input_tokens": 8977, + "estimated_eval_tokens": 18210 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: most items are distinct tokenizer, model-loading, or regression reports, with only a weak tied-weights subtheme standing out. No pull requests are present in the cluster.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45005 is the broadest and most detailed issue around the only clearly recurring theme here: v5 tied-weights behavior. It can plausibly serve as the umbrella report for the related lm_head.weight serialization symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45005 has the richest scope and discussion among the partially related items, making it the strongest single issue to anchor triage for the tied-weights regression thread.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Different models and unrelated failure modes; no shared concrete bug." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "ImportError for missing PyTorch vs embedding resize/reinit regression are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config fields, but the underlying causes and affected paths differ." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is about lm_head.weight serialization by device and the other is a broader translation-model regression; not clearly the same bug." + }, + { + "left": "issue:45200", + "right": "issue:46463", + "accept": false, + "reason": "Gemma 4 multimodal input defaults vs T5 decoder-input validation are different issues." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 loading crash vs trainer.evaluate-after-train failure are unrelated." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizers and different errors; no evidence of the same code path." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization bug and FSDP loading breakage are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both hit from_pretrained flows, but gguf_file handling and offline model_info patching are different bugs." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Position-id key warnings and sinq save/load quantization breakage are different problems." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer-specific failures with no shared underlying defect." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC forward/latents mismatch and Whisper batch_decode behavior are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:46701", + "accept": false, + "reason": "Llama4 config attribute error and DiffusionGemma pipeline failure are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both involve vision models, but they describe different symptoms and likely different root causes." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 zero-probabilities and processor/tokenizer regression are different bugs." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Gemma processor import dependency breakage and Windows ROCm torch.distributed issue are unrelated." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer lifecycle failure and Gemma4 device-mismatch forward bug are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM loading key error are unrelated." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Different optional-backend dependency errors; not the same backend mapping bug." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "Generic v5 cleanup behavior change vs CodeLlama leading-space regression are not clearly the same bug." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage and FSDP loading breakage are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer special-token bugs affecting different models." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but the affected flows and described failure modes differ." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "BLIP2 dtype propagation and RTDetrV2 output divergence are unrelated." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency and MacOS CLIP bus error are unrelated." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation with compiled forward and Qwen3 recurrent-state reset are different generation bugs." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading breakage and BLIP2 dtype mismatch are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization and T5 apex RMSNorm memory leak are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 24, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43493", + "issue:43504", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43782", + "issue:43927", + "issue:44361", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44484", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44821", + "issue:44843", + "issue:44964", + "issue:44977", + "issue:44991", + "issue:45405", + "issue:45440", + "issue:45491", + "issue:45701", + "issue:46002", + "issue:46050", + "issue:46365", + "issue:46402", + "issue:46612", + "issue:46693" + ], + "soft_pairs": [ + "issue:43756|issue:46402", + "issue:44843|issue:44991", + "issue:43653|issue:43927", + "issue:43688|issue:45701", + "issue:44479|issue:44625", + "issue:45491|issue:46002", + "issue:43550|issue:44361", + "issue:43688|issue:45440", + "issue:44484|issue:44704", + "issue:33666|issue:36010", + "issue:43742|issue:44821", + "issue:32090|issue:33666", + "issue:44734|issue:46612", + "issue:44466|issue:46050", + "issue:41720|issue:44464", + "issue:44464|issue:44977", + "issue:43504|issue:43720", + "issue:43504|issue:43550", + "issue:44977|issue:46693", + "issue:43504|issue:44610", + "issue:43577|issue:44554", + "issue:43720|issue:43927", + "issue:33666|issue:36331", + "issue:43643|issue:44561", + "issue:43606|issue:43927", + "issue:43278|issue:43688", + "issue:43782|issue:46693", + "issue:44964|issue:45405", + "issue:39692|issue:43493", + "issue:43262|issue:43295", + "issue:44554|issue:46365", + "issue:43720|issue:44361" + ], + "prepared_review_unit_hash": "bb9b12373a7dd161aec76aca1ca8772431c32f93e4e39479391ab4e34a147209", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37428, + "estimated_input_tokens": 9357, + "estimated_eval_tokens": 18970 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is highly heterogeneous: the items cover unrelated bugs, regressions, model-specific issues, and docs/examples. No soft-edge pair looks like the same underlying change, so I reject all candidate duplicates and do not nominate a single canonical issue/PR.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Different model families and failure modes: Smollm3 RoPE layer mismatch vs Sam3 text encoder weights not loading." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but one is offline-model_info patching and the other is a specific tokenizer failure for EMBEDDIA/est-roberta; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug is unrelated to DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to a broad tokenization-version change report." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video input regression for Qwen VL models is unrelated to num_labels propagation in Qwen3.5 config." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Gemma3 mixed-length NaNs in attention are unrelated to classification-head NaNs during checkout/init." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE auxiliary-loss scaling bug is unrelated to DeepseekV3MoE implementation divergence." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "A save_pretrained shard-size question is unrelated to AutoProcessor not forwarding kwargs to cached_file." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training is unrelated to an import error for GenerationMixin." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM key-error loading is unrelated to AutoImageProcessor failing from URL." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error is a generic training crash, not the same as Qwen2-VL multi-GPU training setup issue." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "KV-cache continuation indexing in transformers serve is unrelated to beam-search cache reorder omissions for specific model classes." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Inconsistent lm_head.weight serialization is unrelated to quantized model save/load breakage." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto device mapping cudaErrorAssert is unrelated to chunked generation inconsistencies under compiled forward." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation under compiled forward is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "BEiT legacy-field loading failure is unrelated to BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "BEiT pretrained preset loading failure is unrelated to Bamba torch.compile SDPA behavior." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3.5 flash-attention correctness bug is unrelated to a general FA2 long-context performance regression." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "BEiT legacy-field load bug is unrelated to OmDet-Turbo processor/model image-size mismatch." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to MPS attention correctness when value head dim differs from query." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet accelerate-loading unpack bug is unrelated to DiaConfig losing custom token IDs." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Qwen2-VL multi-GPU training is unrelated to CustomTrainer.compute_loss signature changes." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is unrelated to DiaConfig token-ID loss." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "BF16-to-FP32 dtype drift in evaluation is unrelated to MoE auxiliary-loss normalization." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3VL weight_only load error is unrelated to FA2 long-context performance regression." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal load failure is unrelated to a PEFT version pin being ahead of PyPI release." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 docs example errors are unrelated to SigLIP2 implementation discrepancy with the original JAX version." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate default bug is unrelated to processor.tokenizer access / passing images to tokenizer regression." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness issue is unrelated to Trainer import failure caused by datasets installation." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading issue is unrelated to MLukeTokenizer task AttributeError." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 25, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:39692", + "issue:42796", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43701", + "issue:43824", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44112", + "issue:44361", + "issue:44451", + "issue:44466", + "issue:44488", + "issue:44514", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44792", + "issue:44843", + "issue:44849", + "issue:44977", + "issue:45290", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46489", + "issue:46710" + ], + "soft_pairs": [ + "issue:43493|issue:43825", + "issue:44514|issue:45290", + "issue:43525|issue:43572", + "issue:43452|issue:44451", + "issue:44625|issue:45584", + "issue:44977|issue:46209", + "issue:45823|issue:46050", + "issue:28218|issue:33357", + "issue:43994|issue:46192", + "issue:42796|issue:44977", + "issue:43278|issue:43844", + "issue:43493|issue:43844", + "issue:44849|issue:46209", + "issue:39692|issue:43295", + "issue:33357|issue:33666", + "issue:43646|issue:44466", + "issue:44568|issue:46489", + "issue:44849|issue:45357", + "issue:46209|issue:46710", + "issue:43278|issue:43701", + "issue:44792|issue:45405", + "issue:44488|issue:44843", + "issue:43575|issue:45405", + "issue:43927|issue:44361", + "issue:44589|issue:45520", + "issue:43257|issue:43299", + "issue:43824|issue:44589", + "issue:44568|issue:45812", + "issue:43504|issue:43606", + "issue:43606|issue:44112" + ], + "prepared_review_unit_hash": "3aa4dbfc9e7a287aeaf415f82604fffc341cadc71f9338846dd1a1edef16e08b", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous. The soft-similarity pairs are mostly false positives across different models, code paths, and failure modes; none looks like the same underlying bug/change closely enough to merge.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45812 is the broadest tokenizer-regression report in the set and has the strongest downstream impact among the tokenizer-related items, so it is the best umbrella representative if one issue must anchor the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "45812 is the most general and widely relevant issue here: it describes a silent tokenizer regression affecting all Granite models, which makes it the best candidate for duplicate triage within this otherwise mixed set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "Both mention model/pipeline correctness, but one is a SigLIP2 implementation mismatch and the other is a pipeline task-support error; different subsystems and symptoms." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both are apply_chat_template crashes, but the triggers differ: batched Qwen2.5-VL with padding=False versus assistant tool-call messages with no content." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "These are similar migration-style config errors, but they involve different models and missing fields (pad_token_id vs pad_token_idx), so they are not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "One is a gguf_file/from_pretrained breakage, the other is a model-specific load failure for ScandiBERT; same broad area, but not the same code-path problem." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Different models and unrelated failure modes: Qwen3.5 num_labels propagation versus Whisper generation after align_special_tokens." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "These are separate runtime issues: Qwen3.5 flash-attention generation versus FP8 inference failure on B200 hardware." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "One is a device-map forward mismatch, the other is quantized save/load breakage; different stages and likely different fixes." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "The tokenizer-added-space bug and the MacOS CLIP bus error are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 nonsense outputs and silently ignored return_tensors are distinct API bugs with different triggers." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer BPE-marker handling and Qwen3.5 flash-attention generation are unrelated issues." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype drift in evaluate and gradient explosion under ZeRO-3 are different training/runtime problems." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "A SigLIP2 discrepancy versus a ZeRO-3 gradient regression are clearly different bugs." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 output_hidden_states behavior and FP8 inference on B200 do not share the same underlying cause." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Documentation example errors are not the same as the processor/tokenizer API regression in custom model code." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "A MacOS bus error in CLIP loading and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is custom model initialization and the other is device-dependent serialization of tied weights." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different mechanisms; they are not the same concrete bug." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Output-hidden-states behavior and incorrect visual encoder key saving are separate Qwen3.5 issues." + }, + { + "left": "issue:46209", + "right": "issue:46710", + "accept": false, + "reason": "FP8 inference on B200 and a DeepSeek-R1-Distill-Llama accuracy regression after upgrade are different problems." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype changes and resume_from_checkpoint key mismatches do not share the same code path." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A failing Janus image-generation test and an unreleased peft version bump are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both involve model loading/tokenizers, but one is a specific model load failure and the other is an offline-mode regression in a regex patch." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to a dependency version bump." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig save/load token-ID loss and MLukeTokenizer task AttributeError are different failures." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "These are different backend compatibility errors: Float8 storage lookup versus flash_attn import metadata on Python 3.13." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both concern Qwen3 MoE models, but one is weight conversion under accelerate+deepspeed and the other is a loading regression for Qwen3VL MoE models." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "An import error for a model class and a Float8 storage object lookup failure are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different model families and different concrete behaviors." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "A legacy-field loading failure in BEiT semantic segmentation is unrelated to the Bark CPU offload device mismatch." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "These both mention devices, but one is a real forward-path mismatch and the other is a stale CI test; not the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 26, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:38175", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43572", + "issue:43653", + "issue:43696", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43937", + "issue:43994", + "issue:44291", + "issue:44423", + "issue:44464", + "issue:44493", + "issue:44554", + "issue:44734", + "issue:44964", + "issue:44977", + "issue:45084", + "issue:45200", + "issue:45229", + "issue:45245", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46121", + "issue:46285", + "issue:46369", + "issue:46424", + "issue:46701", + "issue:46710" + ], + "soft_pairs": [ + "issue:43334|issue:43572", + "issue:43334|issue:45759", + "issue:43572|issue:44291", + "issue:45245|issue:46710", + "issue:44493|issue:46285", + "issue:43504|issue:43653", + "issue:44464|issue:46424", + "issue:43937|issue:46701", + "issue:45200|issue:45229", + "issue:44464|issue:45584", + "issue:45229|issue:45823", + "issue:43844|issue:44964", + "issue:43873|issue:46050", + "issue:43994|issue:44493", + "issue:29942|issue:30333", + "issue:45588|issue:46050", + "issue:29159|issue:32090", + "issue:42490|issue:43696", + "issue:45084|issue:45520", + "issue:44977|issue:46701", + "issue:28218|issue:34567", + "issue:15354|issue:38175", + "issue:43065|issue:46369", + "issue:44554|issue:45446", + "issue:44423|issue:44734", + "issue:43493|issue:43701", + "issue:41720|issue:46710", + "issue:45406|issue:46121" + ], + "prepared_review_unit_hash": "0e37f31d798ce2663d66e87772e6398793b552ae114f72a7dd3c9dae91f6342f", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36157, + "estimated_input_tokens": 9040, + "estimated_eval_tokens": 18336 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are broadly unrelated, with only loose thematic overlap around model/config loading regressions. No soft pair looks like the same underlying bug, so I would not merge any of them as duplicates.", + "confidence": 0.41, + "canonical_issue_reason": "issue:43334 is the most representative of the small subset of load-time config compatibility reports (Qwen multimodal config missing a pad token attribute), but overall the cluster is too heterogeneous to indicate a true duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43334 is the clearest, most concrete bug report among the set and sits near the most plausible thematic center of the cluster (model/config loading regressions).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing-attribute config load failures, but they affect different models, different fields, and different code paths." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is incorrect text_config unwrapping; not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "One is a missing config field after a version update; the other is an init_empty_weights/_is_hf_initialized constructor mismatch. Different failures." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "Integer category-limit error and an accuracy regression are unrelated bugs." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected position-id key handling and CLIP text model structure inconsistency are separate issues." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field model loading failure and tokenizer special-token registration are different problems." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Compiled-forward chunked-generation inconsistency is not the same as cache_length being ignored except for quantized cache." + }, + { + "left": "issue:43937", + "right": "issue:46701", + "accept": false, + "reason": "Invalid GenerationConfig for GLM-5 and a diffusion pipeline failure are different code paths and symptoms." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Defaulting mm_token_type_ids and CUDA OOM in multi-GPU inference are unrelated." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "DeepSpeed/ZeRO gradient growth and Phi-4 multimodal loading failure are not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantized offloading behavior and sinq save/load breakage are different quantization-related issues, not one concrete defect." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention sink-less crash and sinq serialization breakage have no shared underlying code-path failure." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency and _gpu_broadcast_one NoneType errors are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLFlow reporting on failed jobs are separate concerns." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation errors and flash_attn import key errors are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:46701", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issues and DiffusionGemma pipeline failure are different models and mechanisms." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer added-space behavior and TrainerState token counting are unrelated." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure and SIGLIP2 zero-probability outputs are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Sam3PixelDecoder dummy convolution and sam3-litetext memory overflow are different issues." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and a PyTorch version guard for flex_attention imports are not the same bug." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve-related, but one crashes on multimodal tensor conversion and the other on KV-cache continuation indexing; distinct failures." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:46710", + "accept": false, + "reason": "CUDA assert under auto device mapping and an accuracy regression after upgrade are not the same underlying issue." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and rope-parameter JSON list handling are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 27, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43404", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43653", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:44112", + "issue:44186", + "issue:44466", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44849", + "issue:44987", + "issue:45072", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45356", + "issue:45381", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46710" + ], + "soft_pairs": [ + "issue:43404|issue:46710", + "issue:45072|issue:45198", + "issue:37428|issue:43957", + "issue:43295|issue:43450", + "issue:45588|issue:45759", + "issue:43540|issue:45200", + "issue:45200|issue:45823", + "issue:44186|issue:46192", + "issue:43295|issue:43873", + "issue:45823|issue:46192", + "issue:43295|issue:44610", + "issue:45520|issue:46178", + "issue:43232|issue:44849", + "issue:43572|issue:44987", + "issue:43450|issue:43575", + "issue:43550|issue:43844", + "issue:43572|issue:43957", + "issue:44554|issue:45356", + "issue:37428|issue:43824", + "issue:42490|issue:43493", + "issue:43653|issue:44112", + "issue:44466|issue:45127", + "issue:38175|issue:43493", + "issue:39692|issue:46192", + "issue:43278|issue:43856", + "issue:45381|issue:46209", + "issue:44521|issue:45290", + "issue:28218|issue:35141", + "issue:42915|issue:43824", + "issue:43950|issue:46285" + ], + "prepared_review_unit_hash": "7b35ca44567f1eee8f0fda422d471df6df07dbbf8104e50247e4ceb1bbefe7f1", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37041, + "estimated_input_tokens": 9261, + "estimated_eval_tokens": 18778 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All proposed soft edges look like false positives: the items span unrelated tokenizer, model-loading, generation, quantization, multimodal-processing, and documentation bugs with no clear shared code path.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: every soft pair appears to describe a different bug or regression, so there is no duplicate cluster to collapse.", + "canonical_pr_reason": "No PRs present in the provided items.", + "best_issue_reason": "No best issue: none of the issues is clearly representative of a merged duplicate set.", + "best_pr_reason": "No PRs present in the provided items.", + "soft_edge_verdicts": [ + { + "left": "issue:43404", + "right": "issue:46710", + "accept": false, + "reason": "Different bugs and model families: Mistral3 tie-weight issue vs DeepSeek accuracy regression." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "CI dtype mismatch in inference vs Wav2Vec2 save/tokenization failure; unrelated symptoms." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Missing import symbol vs meta-device loading regression; different code paths." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor.tokenizer regression with images vs batched video output shape bug." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "flash_attention crash on s_aux=None vs AutoModelForCausalLM config unwrapping; unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Qwen3Omni video ValueError vs Gemma4 mm_token_type_ids defaulting issue." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Missing mm_token_type_ids vs device_map auto forward mismatch; different root causes." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Tokenizer crash on NER/batched padding vs return_tensors ignored with text_kwargs; different behaviors." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer bug vs quantization offloading failure; not the same bug." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device mismatch vs return_tensors handling; different APIs and failures." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Images-to-tokenizer regression vs processor output size mismatch; unrelated preprocessing bugs." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn import KeyError on Python 3.13 vs BACKENDS_MAPPING error for tensorflow_text." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwarg update after sync_gpus vs Qwen3.5 hidden_states bug; different code paths." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in config vs model loading failure for a specific repo; unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape bug vs tensor-parallel OOM when loading a large model; not duplicates." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile+SDPA failure vs ZeRO-3 gradient increase; different training/runtime issues." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Config regression vs meta-device loading breakage; different underlying problems." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness bug vs Kimi-K2.5 tokenizer regression; unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Flash-attention helper import error vs missing Qwen2_5_VL export; different missing symbols." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Local-path import warning vs SigLIP2 implementation discrepancy; unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration vs GraniteSpeech CI device override test." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "lm_head serialization inconsistency vs LoRA merge collapse after vocab extension; related theme but not the same code-path bug." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one reports zero probabilities and the other a JAX/HF discrepancy; not clearly the same bug." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "Documentation example errors vs return_tensors/text_kwargs handling; different issues." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift on evaluate vs MoE memory inefficiency; unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "qwen2.5-vl video position-id bug vs FP8 inference failure on B200; different models and hardware paths." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "assistant_masks all zero vs crash on tool-call messages; same API but distinct bugs." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Added-token spacing bug vs embedding reinitialization after resize; unrelated tokenizer/model-init issues." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FineGrainedFP8Config failure on Qwen3Moe vs Qwen2_5_VL import error; unrelated." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption in from_pretrained vs CLIPTextModel structure inconsistency; different loading/model-shape bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 28, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:43262", + "issue:43334", + "issue:43335", + "issue:43388", + "issue:43425", + "issue:43450", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43901", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44792", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45584", + "issue:45593", + "issue:45759", + "issue:46170", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46512" + ], + "soft_pairs": [ + "issue:43452|issue:46512", + "issue:43824|issue:43975", + "issue:43334|issue:43525", + "issue:43262|issue:43450", + "issue:44442|issue:45198", + "issue:44792|issue:45563", + "issue:43819|issue:44112", + "issue:44479|issue:46192", + "issue:44898|issue:45412", + "issue:43827|issue:43901", + "issue:45200|issue:45325", + "issue:44561|issue:46234", + "issue:44610|issue:44857", + "issue:43388|issue:43526", + "issue:43335|issue:43550", + "issue:43819|issue:45083", + "issue:41720|issue:46209", + "issue:44857|issue:45593", + "issue:44206|issue:45584", + "issue:45759|issue:46285", + "issue:45520|issue:46170", + "issue:43606|issue:45083", + "issue:38175|issue:45412", + "issue:43335|issue:43653", + "issue:29942|issue:32090", + "issue:37428|issue:43425", + "issue:43550|issue:44112" + ], + "prepared_review_unit_hash": "b3cc25802611c3dc09ae59ecf397499a8a400ff4d512db57b698147ddd0cb765", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35515, + "estimated_input_tokens": 8879, + "estimated_eval_tokens": 18014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft-similarity pairs mostly share broad themes like import errors, model-specific regressions, or processor/config issues, but they do not look like the same underlying bug or change. I found no safe PR duplicate merge here; the strongest issue anchor is a highly discussed, concrete bug report, but it is not representative of the whole set.", + "confidence": 0.39, + "canonical_issue_reason": "issue:43824 has the strongest signal in the cluster (high discussion/inbound refs) and a concrete, reproducible import failure, making it the best single issue anchor despite the cluster\u2019s overall heterogeneity.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the clearest, most actionable issue in the set; it has strong engagement and a specific failure mode, while most other items are unrelated one-off regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "Different problems: gguf/from_pretrained loading breakage vs config.dtype not being updated by from_config." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Both involve Qwen-related behavior, but one is an import failure and the other is a detokenization bug in a different path." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Same symptom class, but different model configs and likely separate missing-attribute fixes." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and batched video shape handling are unrelated code paths." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Both are speech/audio tokenizer issues, but they affect different models and different failure points." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure and a stale generation warning are not the same bug." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent mismatch vs GraniteSpeech CI device override are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Video-input regression in specific Qwen models vs return_tensors being ignored in text_kwargs are different bugs." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution failure and RT-DETR memory release are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but they reference different deprecated/changed behaviors and are not the same change." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma multimodal token ids defaulting and Qwen2.5-VL rope indexing are separate model-specific regressions." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Both mention missing import_utils helpers, but they are different helper removals with different breakages." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Processor output size mismatch and AMP crash in a loss are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both concern labels, but one is gather_for_metrics batching and the other is BeitImageProcessorFast label reduction." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and Bamba torch.compile SDPA failure are different code paths." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent mismatch and qwen3_omni_moe length helper behavior are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Both are GPU inference failures, but on different models, hardware, and likely different root causes." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Loss AMP crash and missing auxiliary losses when denoising is off are separate training issues." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported feature-extractor arg and Whisper empty-transcription handling are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Composite Qwen config unwrapping and CLIP text-model structure inconsistency are distinct model-loading issues." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 flash_attn import key error and Windows ROCm torch.distributed dependency failure are unrelated platform issues." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch and qwen3_omni_moe feature length calculation are not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Zero probabilities in SigLIP2 and RT-DETR memory retention are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Sparse-layer config bug and tokenizer special-token registration bug are different subsystems." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Flash-attention import error and generic Torch 2.10 incompatibility are related only at a broad dependency level, not the same concrete fix." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba torch.compile SDPA failure and GraniteSpeech CI stale test are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 29, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43504", + "issue:43526", + "issue:43606", + "issue:43611", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:43931", + "issue:44112", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45859", + "issue:46234", + "issue:46326", + "issue:46431", + "issue:46512", + "issue:46566" + ], + "soft_pairs": [ + "issue:28282|issue:33357", + "issue:43611|issue:43931", + "issue:44448|issue:44743", + "issue:43335|issue:43504", + "issue:45759|issue:46512", + "issue:44855|issue:45083", + "issue:43720|issue:44112", + "issue:43278|issue:43825", + "issue:29127|issue:32090", + "issue:43335|issue:45083", + "issue:43335|issue:44112", + "issue:43335|issue:45593", + "issue:37428|issue:44351", + "issue:44514|issue:44521", + "issue:34689|issue:35141", + "issue:45588|issue:45859", + "issue:45127|issue:45245", + "issue:45290|issue:46326", + "issue:46431|issue:46566", + "issue:42490|issue:43854", + "issue:45200|issue:45588", + "issue:45310|issue:45478", + "issue:39692|issue:43526", + "issue:44448|issue:44779", + "issue:43606|issue:44857", + "issue:45459|issue:46234", + "issue:43819|issue:44493" + ], + "prepared_review_unit_hash": "bbf8110eabb273742de19bd6d339641a3372f4eb85dfad4a80cd846dece9b1ad", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35623, + "estimated_input_tokens": 8906, + "estimated_eval_tokens": 18068 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issue reports with one clear duplicate pair: the Qwen3.5 MoE from_pretrained error. All other soft pairs appear to describe different bugs or regressions, even when they touch nearby code paths.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is the earlier, more specific report for the Qwen3.5 MoE `from_pretrained` failure and matches 45478 exactly.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45310 is the strongest representative because it is the exact duplicate target of 45478 and has the clearest, most direct problem statement.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Both are loading/runtime failures, but one is a missing PyTorch import and the other is a MacOS bus error with a CLIP model; different underlying bugs." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Both are model-loading regressions, but they involve different failure modes and different affected models (`base_model_prefix` vs Qwen3-VL weight shape mismatch)." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "One is an output difference for Pegasus between v4 and v5; the other is recurrent states resetting in modular_qwen3_5 with cache. Not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers config creating sparse layers and BEiT pretrained preset loading are unrelated issues in different model families." + }, + { + "left": "issue:45759", + "right": "issue:46512", + "accept": false, + "reason": "Both concern `from_config`, but one is about unwrapping `text_config` in composite multimodal configs while the other is about stale `config.dtype` after dtype-aware initialization." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 import parsing failure in DebertaV2 is unrelated to a helper returning unexpected lengths in qwen3_omni_moe." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Both mention CI/device issues, but one is packed-weight unpacking during accelerate loading and the other is a stale device override test in GraniteSpeech." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Embedding dtype drift during evaluate is unrelated to a pipeline error message about translation tasks in v5." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 box-validation messaging and Trainer `_gpu_broadcast_one` NoneType handling are distinct error-handling bugs." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Different model components and symptoms: sparse-layer config bug vs feature-extraction length helper behavior." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Sparse-layer creation in SwitchTransformers and GraniteSpeech CI device override are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers config bug and D-FINE auxiliary-loss behavior are unrelated model-specific issues." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different breakages (`_flash_supports_window_size` vs `HybridCache`)." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both are Qwen2.5-VL chat-template issues, but one is a batching crash with `padding=False` while the other is incorrect all-zero assistant masks for multimodal inputs." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision model loading regression and output embedding reinitialization after resizing are different problems." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both involve offload/flash-related behavior, but one is an `s_aux=None` crash in flash attention integration and the other is a CPU-offload incompatibility in rotary embedding forward." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge collapse with extended vocab is unrelated to the category-count ceiling error." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both touch chat templating, but one crashes on assistant tool-call messages without content and the other is a `continue_final_message` bug for mistral_common tokenizers." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "Missing chat template support in `apply_chat_template` is unrelated to DiffusionGemma weights staying on meta with auto device map." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "A trailing-slash local-path import warning is not the same as a GLM-4.7-Flash model-loading failure in unit tests." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 `mm_token_type_ids` defaulting and flash attention `s_aux=None` are unrelated issues." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Exact same Qwen3.5 MoE `from_pretrained` error on the same version range; 45478 is a clear duplicate of 45310." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 documentation example errors and BEiT Fast image processor label reduction are different concerns." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output differences and DeepSeek tokenizer regressions are separate model/tokenizer bugs." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch in Bark-small is unrelated to mixed-precision CUDA crash in LwDetrImageLoss." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf error masking and a missing `is_falcon_mamba_ssm_available` utility function are unrelated." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC `from_latents` mismatch due to missing STE and unexpected position-id keys are different defects." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet, and the soft-edge verdicts are conservative. The only accepted pair, issue:45310 vs issue:45478, is strongly supported by the identical titles and close timing; the remaining pairs are reasonably treated as different bugs or regressions." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 30, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:26413", + "issue:27957", + "issue:28218", + "issue:29942", + "issue:31501", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:42175", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43701", + "issue:43927", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44617", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45290", + "issue:45381", + "issue:45588", + "issue:45698", + "issue:45823", + "issue:45920", + "issue:46192", + "issue:46399", + "issue:46655", + "issue:46697" + ], + "soft_pairs": [ + "issue:26413|issue:31501", + "issue:28218|issue:32090", + "issue:39692|issue:44493", + "issue:44617|issue:46697", + "issue:44206|issue:45823", + "issue:43701|issue:45083", + "issue:32090|issue:33357", + "issue:44857|issue:45072", + "issue:42490|issue:43575", + "issue:44360|issue:44561", + "issue:38175|issue:44493", + "issue:43550|issue:45083", + "issue:29942|issue:35141", + "issue:27957|issue:34689", + "issue:43278|issue:43329", + "issue:45920|issue:46192", + "issue:44448|issue:44977", + "issue:40444|issue:43526", + "issue:45290|issue:45381", + "issue:43278|issue:43381", + "issue:44977|issue:45229", + "issue:10000|issue:15354", + "issue:44855|issue:46399", + "issue:43927|issue:45698", + "issue:44857|issue:45588", + "issue:45290|issue:46655", + "issue:37428|issue:42175", + "issue:32090|issue:33290" + ], + "prepared_review_unit_hash": "57c08377205bdc8484916ba51a2dce8b43645d1c2f4e413ecd4777ce286569b1", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36110, + "estimated_input_tokens": 9028, + "estimated_eval_tokens": 18312 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly unrelated across different models, processors, training flows, and regressions. The soft-similarity links are superficial, so none of the candidate pairs should be merged.", + "confidence": 0.96, + "canonical_issue_reason": "No single issue cleanly represents the cluster: it mixes unrelated bugs in training, tokenization, Flash Attention, model loading, serialization, and image/video processors.", + "canonical_pr_reason": null, + "best_issue_reason": "No best issue among these; the cluster is too heterogeneous to have a meaningful canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Both are training-start failures, but one is a resume_from_checkpoint epoch-iterator problem and the other is a FineWeb SLM startup issue; different code paths and root causes." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing after an added token is unrelated to the Trainer _gpu_broadcast_one NoneType error." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "A SigLIP2 docs/example mismatch plus quantization failure is not the same bug as a position-id warning across many models." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "Sam3Video CUDA OOM and Swin2SR over-padding are different models and different failure modes." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported center argument in LasrFeatureExtractor is unrelated to Gemma4 device-map forward mismatches." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "A checkpoint key mismatch is not the same as unexpected feature-extraction length behavior in qwen3_omni_moe." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast of NoneType and MacOS bus error in a CLIP model are distinct runtime problems." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "float16 AMP crash in LwDetrImageLoss and dtype mismatches in SwitchTransformers/TimmWrapperModel are separate issues." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "A local-path import warning is unrelated to an out-of-memory failure when loading a very large Qwen2 model with tensor parallelism." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "Missing ReLU in the DSA indexer and removal of is_torch_fx_available breaking trust_remote_code models are unrelated changes." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in siglip2-base-patch16-224 is not the same as a generic position-id key regression." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Bamba-9B-v2 torch.compile/SDPA failure and qwen3_omni_moe feature-length behavior involve different subsystems and bugs." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to output embeddings being reinitialized during post_init after resize_token_embeddings." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 90B Vision loading regression are different concrete issues." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "A BF16-to-FP32 dtype change in evaluation is unrelated to undefined variables in the video branch of multimodal token counting." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Wrong token IDs across multiple tokenizers is a different tokenizer bug than silently ignoring return_tensors when text_kwargs is passed." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output differences and Qwen3.5 flash-attention generation errors are unrelated model-specific regressions." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "IterableDataset finetuning with multiple images and BeitImageProcessorFast reduce_labels returning one label are distinct bugs." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crashes and Qwen2.5-VL video vision_position_ids issues are different code paths." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Both mention eval-mode behavior, but BF16/FP32 drift is not the same as gradient checkpointing being disallowed in eval mode." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation failure in Qwen3.5 is unrelated to Gemma4 multi-GPU CUDA OOM." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "German DistilBertModel issue and GeneratorExp/JIT export failure are separate, old, unrelated problems." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Python 3.13 IndentationError/JIT parsing failure is unrelated to apply_rotary_pos_emb registration errors in Qwen3.5." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Losing custom token IDs after save/load and loading the wrong custom module after save_pretrained are both persistence-related, but they are different mechanisms and bugs." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP crashes in LwDetrImageLoss and flash_attention.py AttributeError on s_aux=None are different failures." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "A missing internal flash-attention import and missing TensorFlow backend in a torch install are unrelated packaging/import issues." + }, + { + "left": "issue:43329", + "right": "issue:46655", + "accept": false, + "reason": "Undefined helper variables in multimodal token counting are unrelated to incorrect padding calculation in SmolVLM's video processor." + }, + { + "left": "issue:44493", + "right": "issue:44561", + "accept": false, + "reason": "Unexpected position-id keys and removal of is_torch_fx_available breaking trust_remote_code models are not the same regression." + }, + { + "left": "issue:38175", + "right": "issue:46192", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to return_tensors being ignored when text_kwargs is explicitly passed." + }, + { + "left": "issue:43381", + "right": "issue:45823", + "accept": false, + "reason": "Gradient checkpointing in eval mode and Gemma4 PLE device mismatch are different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 31, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:34567", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43278", + "issue:43381", + "issue:43412", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43873", + "issue:43976", + "issue:44186", + "issue:44206", + "issue:44361", + "issue:44442", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45712", + "issue:45803", + "issue:45823", + "issue:46121" + ], + "soft_pairs": [ + "issue:25251|issue:32090", + "issue:44561|issue:44779", + "issue:43873|issue:45229", + "issue:44448|issue:44466", + "issue:42490|issue:43873", + "issue:43976|issue:45440", + "issue:43278|issue:43575", + "issue:45588|issue:45823", + "issue:41720|issue:45229", + "issue:33357|issue:39692", + "issue:43381|issue:43701", + "issue:44568|issue:45491", + "issue:29127|issue:34567", + "issue:44977|issue:45563", + "issue:44871|issue:44977", + "issue:45198|issue:45803", + "issue:45372|issue:45588", + "issue:43653|issue:44743", + "issue:44186|issue:45198", + "issue:44361|issue:45198", + "issue:43611|issue:46121", + "issue:42915|issue:42994", + "issue:43412|issue:45588", + "issue:45372|issue:45406", + "issue:29942|issue:30064", + "issue:45459|issue:45712", + "issue:43653|issue:44442", + "issue:43756|issue:45440", + "issue:44206|issue:44479" + ], + "prepared_review_unit_hash": "78a8056c17bf7a6b78895f2b31cd141c47b5138a42565e37266b2f72338fa992", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36545, + "estimated_input_tokens": 9137, + "estimated_eval_tokens": 18530 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set looks like a collection of unrelated Transformers bug reports across tokenization, model loading, training, inference, quantization, and docs. The suggested soft pairs share superficial wording only; none appear to describe the same concrete underlying bug or merge cleanly into one fix.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Both mention runtime/output problems, but one is about pipeline top_k changing list nesting and the other is a Trainer NoneType broadcast error. Different code paths and failure modes." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions involving remote or tokenizer-related breakage, but one is about removal of is_torch_fx_available and the other is a DeepSeek tokenizer correctness issue. Not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Both involve memory/offloading behavior, but one is offloading with quantization generally and the other is a Gemma4 31B multi-GPU CUDA OOM. Different models and likely different root causes." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both reference v5 behavior changes, but one is a generation/output difference for pegasus and the other is inconsistent lm_head.weight serialization. Distinct bugs." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "A local-path import warning is unrelated to quantization/offloading behavior. No shared underlying defect." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "One is a Python version compatibility regression, the other is a DeepseekV3MoE implementation divergence. Different layers of the stack." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Both concern dtype/OOM-like model loading behavior, but one is an embedding dtype mismatch during eval and the other is tensor-parallel OOM for a specific Qwen2 model. Not the same bug." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Both mention Gemma4/processor-related runtime issues, but one is a flash_attention crash on s_aux=None and the other is a PLE device mismatch during forward. Different failure points." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Both mention CUDA/device-map failures, but one is a Qwen3 auto device mapping cudaErrorAssert and the other is Gemma4 31B multi-GPU OOM. Different symptoms and likely causes." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "One is a MacOS bus error for a community CLIP model; the other is a SigLIP2 docs example with model/processor mismatch and quantization failure. Not the same bug." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Both involve checkpoint/resume/state handling, but one is about gradient checkpointing in eval mode and the other is a resume_from_checkpoint key mismatch. Different mechanisms." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Both involve tokenizer/model configuration issues, but one is missing BOS/EOS addition and the other is NaN embeddings from sliding-window attention with mixed-length batches. Different code paths." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "One is a LayoutLMv3 error-message clarity issue; the other is TrainerState num_input_tokens_seen not updating. No shared bug." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "One is a Qwen3.5 flash-attention generation regression; the other is a stale warning for paged generate num_return_sequences. Different behavior and fix surface." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "One is an eos_token_id configuration mismatch in Gemma-3; the other is a Qwen3.5 flash-attention generation issue. Unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Both are save/convert-related failures, but one is Wav2Vec2 save_pretrained/tokenization and the other is a bare except in FuyuBatchFeature.convert_to_tensors(). Different components and bug classes." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Both touch Gemma4 processing, but one is an ImportError from mistral_common while loading the processor and the other is a flash_attention AttributeError on s_aux=None. Separate defects." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Both involve model/tokenizer implementation correctness, but one is BigBirdTokenizer special-token registration and the other is Qwen3.5 recurrent state reset with cache. Different subsystems." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Both are tokenizer/input handling bugs, but LayoutLMv2 NER padding/truncation crashes are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError on tasks is a tokenizer runtime bug; Wav2Vec2 save_pretrained/tokenization failure is a different model and failure mode." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "One is a Transformers 5.0 base_model_prefix loading regression; the other is a rope-validation helper TypeError on JSON-loaded lists. Different code paths." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both involve quantization, but one is Qwen3Moe with FineGrainedFP8Config and the other is quantized model saving. Not the same underlying issue." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Gemma3n executorch export failures and a flash_attention s_aux=None crash are separate export/runtime problems." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both concern Gemma4 processor-related failures, but one is a missing dependency import and the other is a missing _tokenizer attribute in transformers serve. Distinct bugs." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to void segmentation map processing in an image processor." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Both involve error-handling or import-related behavior, but one is protobuf decode error handling in tokenizers and the other is dummy PT classes leaking into dir(transformers). Different issues." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer-loading bugs." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE layer-count mismatch and DeepseekV3MoE divergence are model architecture differences, not the same defect." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "The LasrFeatureExtractor center-arg regression and the Qwen2_5/Qwen3 video-input regression are different API breakages in different model families." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 32, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29942", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:42490", + "issue:42915", + "issue:43065", + "issue:43066", + "issue:43334", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43701", + "issue:43761", + "issue:43824", + "issue:43931", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:45072", + "issue:45198", + "issue:45245", + "issue:45250", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45593", + "issue:45834", + "issue:45859", + "issue:46121", + "issue:46369", + "issue:46399", + "issue:46640", + "issue:46693" + ], + "soft_pairs": [ + "issue:45491|issue:46693", + "issue:29127|issue:33357", + "issue:45372|issue:45459", + "issue:45245|issue:45563", + "issue:43065|issue:45412", + "issue:43065|issue:43994", + "issue:44265|issue:45198", + "issue:27957|issue:29942", + "issue:43575|issue:43701", + "issue:45440|issue:45593", + "issue:29127|issue:35141", + "issue:42490|issue:43381", + "issue:43066|issue:44462", + "issue:45412|issue:46369", + "issue:43994|issue:45412", + "issue:43575|issue:44387", + "issue:43421|issue:45491", + "issue:44442|issue:45072", + "issue:42915|issue:43931", + "issue:43526|issue:45834", + "issue:43526|issue:46693", + "issue:43824|issue:44279", + "issue:27957|issue:29127", + "issue:27957|issue:33290", + "issue:46121|issue:46399", + "issue:43526|issue:44186", + "issue:43994|issue:46369", + "issue:33290|issue:34689", + "issue:43334|issue:43531", + "issue:45250|issue:46640", + "issue:33290|issue:35141", + "issue:33290|issue:34567", + "issue:43761|issue:45859" + ], + "prepared_review_unit_hash": "de274acb319777c6826975ecd961608d3983e437a4c2256907cc5003332784a6", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36752, + "estimated_input_tokens": 9188, + "estimated_eval_tokens": 18632 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issues that only share broad keywords or subsystems. I reject all soft edges because they do not appear to be the same underlying bug or change.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:46693 is the clearest standalone report in the set: it is open, recent, and describes a concrete Flash Attention 2 long-context performance regression with discussion and inbound references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both mention FA2, but one is a Gemma3 NaN embedding bug on mixed-length batches and the other is a long-context performance regression; different symptoms and code paths." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 box-validation messaging and a MacOS CLIP bus error are unrelated failures." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "One is a mistral_common import error affecting Gemma 4 processor loading; the other hides tokenizer errors when protobuf is missing. Different bugs." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-limit runtime error and a stale paging warning are not the same problem." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and RT-DETR memory release behavior are different model issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to SigLIP2 producing bad outputs." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure via torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "XLMRoberta FA2 support and failing FA2 tests are not the same concrete bug; one is model support, the other is test/CI failure." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Qwen2 load OOM under tensor parallelism and a resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE remote divergence and D-FINE auxiliary-loss behavior are different code paths and symptoms." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer/error-message clarity and embedding reinitialization after resize are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Path-trailing-slash import warning and gradient-checkpointing eval-mode restriction are different issues." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are separate tokenizer-loading bugs." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both mention memory, but one is an RT-DETR GC leak while the other is a Sam3-LiteText overflow; not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 nonsensical results and RT-DETR memory retention are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both are OOM reports, but the model/context and root cause differ (tp load vs int4 quantization reserved memory)." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "TokenizersBackend post-processor updates and Gemma3 NaN embeddings are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and bfloat16 dtype mismatches are different failures." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are distinct model-loading issues." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and Kosmos2.5 long-ocr index errors are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:46693", + "accept": false, + "reason": "A Beit image-processor label bug is unrelated to a Flash Attention 2 performance regression." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "A missing Qwen2_5_VL symbol import and a generic dependency issue are not the same underlying bug." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "FA2 support for XLMRoberta and a LayoutLMv3 error-message complaint are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "Flash Attention 2 support and an Adafactor/DeepSpeed OOM are different problems." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "RoPE-parameter dict conversion TypeError and a Qwen3.5 kernel-registration error are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "Beit label reduction and LayoutLMv2Tokenizer NER/padding crashes are different tokenizer/image-processor bugs." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 bad outputs and Sam3-LiteText memory overflow are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Adafactor OOM in DeepSpeed and Llama 3.2 Vision loading breakage are different issues." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL missing pad_token_id and Qwen3-MoE sliding_window issues share a family name but not the same bug." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "These are both Flash Attention 2 topics, but the evidence is too broad to show the same concrete bug or mergeable change." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "DeepSpeed OOM and embedding reinitialization after resize are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM under Adafactor/DeepSpeed and num_input_tokens_seen not updating are unrelated training bugs." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 33, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:34689", + "issue:42175", + "issue:42490", + "issue:43262", + "issue:43316", + "issue:43381", + "issue:43493", + "issue:43525", + "issue:43611", + "issue:43643", + "issue:43646", + "issue:43720", + "issue:43723", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44206", + "issue:44442", + "issue:44479", + "issue:44514", + "issue:44743", + "issue:44811", + "issue:44855", + "issue:44871", + "issue:44898", + "issue:44913", + "issue:44977", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45216", + "issue:45245", + "issue:45325", + "issue:45381", + "issue:45440", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45759", + "issue:45907", + "issue:46178", + "issue:46234", + "issue:46697" + ], + "soft_pairs": [ + "issue:45588|issue:45600", + "issue:43381|issue:43873", + "issue:44514|issue:45381", + "issue:43646|issue:44077", + "issue:43262|issue:44206", + "issue:44977|issue:45245", + "issue:44743|issue:44811", + "issue:44442|issue:44898", + "issue:34689|issue:43611", + "issue:42490|issue:44743", + "issue:45003|issue:45440", + "issue:43720|issue:44442", + "issue:42175|issue:46234", + "issue:43316|issue:43525", + "issue:43723|issue:45479", + "issue:43493|issue:43827", + "issue:44077|issue:44479", + "issue:45325|issue:46697", + "issue:44062|issue:45907", + "issue:43873|issue:44871", + "issue:43761|issue:44079", + "issue:43749|issue:43931", + "issue:43931|issue:45216", + "issue:43643|issue:44913", + "issue:44855|issue:45137", + "issue:44987|issue:46178", + "issue:43316|issue:45759" + ], + "prepared_review_unit_hash": "be0a166c8899925b4101e13a82f941ccf92af3209faf1dbed73fdf900fd48b63", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36014, + "estimated_input_tokens": 9004, + "estimated_eval_tokens": 18264 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These soft links are overwhelmingly false positives: they connect issues from different models, features, and failure modes rather than the same underlying bug or change. I would not merge any of the proposed pairs.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the items span unrelated regressions across loading, processors, configs, quantization, docs, and model-specific behavior, so there isn\u2019t a cohesive duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue; every proposed pair is still about a different concrete code path, so none can serve as a representative duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "Different bugs: flash-attention runtime crash vs auto-mapping/CI reference cleanup." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Unrelated failures: gradient checkpointing in eval mode vs quantization/offloading behavior." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Same product family but different code paths: batched chat template padding crash vs video position-id computation." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Both initialization-related, but one is a broad v5 custom-model init regression and the other is a specific post_init policy issue." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Different audio preprocessing bugs: default sample rate handling vs unsupported center argument." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Completely unrelated: flash-attention generation bug vs category-cardinality runtime error." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems and symptoms: recurrent cache state reset vs Whisper batch_decode token handling." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure vs Perceiver image classification/interpolation bug; no shared underlying defect." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but for different versions and different causes; not the same concrete bug." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning and Qwen recurrent-state reset are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "sys.modules access bug vs DeepSeekV3MoE implementation divergence; different code paths." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading issue is unrelated to FastSpeech2Conformer tokenizer loading." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Package-backend install issue vs missing import_utils symbol; not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both config API inconsistencies, but for different models and different missing fields." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading regression vs sequence-classification loss bug; unrelated code paths." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch vs docs still referencing removed pipeline APIs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Specific post_init restriction vs Qwen video-input regression; different features and fixes." + }, + { + "left": "issue:45325", + "right": "issue:46697", + "accept": false, + "reason": "Qwen2.5-VL rope/temporal position bug vs Swin2SR over-padding bug." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken kwargs collision vs repo listing missing revision forwarding; unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading issue vs Gemma eos_token_id config mismatch." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVision hidden_states not returned vs generic ModelOutput key assignment bug; related area but different defects." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading breakage vs Qwen3-VL shape-mismatch loading error." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Model loading regression vs backend registry error; no shared underlying bug." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields with trust_remote_code vs rotary_pct not persisting on reload; different config behaviors." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 torch.jit parsing/import issue vs DeepSpeed ZeRO3 deque error." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Different failure modes and code paths; one is model loading, the other is backend availability resolution." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Config API discrepancy for Gemma3TextConfig vs composite Qwen multimodal from_config unwrapping." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 34, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42175", + "issue:42617", + "issue:42907", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:43825", + "issue:44077", + "issue:44162", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44496", + "issue:44534", + "issue:44568", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:45479", + "issue:45584", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46234", + "issue:46291", + "issue:46365", + "issue:46421", + "issue:46506" + ], + "soft_pairs": [ + "issue:43761|issue:44188", + "issue:44190|issue:44496", + "issue:43232|issue:43388", + "issue:44442|issue:45803", + "issue:42907|issue:45698", + "issue:42617|issue:43575", + "issue:43701|issue:46506", + "issue:43575|issue:43742", + "issue:27957|issue:32090", + "issue:25251|issue:28282", + "issue:44568|issue:45479", + "issue:43550|issue:43701", + "issue:46234|issue:46291", + "issue:43606|issue:43701", + "issue:44464|issue:46421", + "issue:16998|issue:22355", + "issue:43295|issue:43825", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:43316|issue:43716", + "issue:42175|issue:43316", + "issue:44077|issue:44534", + "issue:44871|issue:46050", + "issue:43646|issue:44913", + "issue:43452|issue:44496", + "issue:45803|issue:46365", + "issue:44162|issue:44792", + "issue:45803|issue:46170", + "issue:43404|issue:43441" + ], + "prepared_review_unit_hash": "30ba16acf0d29ecb2a8c0496f0b2a0ba25374e0047b414fa90aafd56c18f37ac", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36350, + "estimated_input_tokens": 9088, + "estimated_eval_tokens": 18432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated Transformers issues across tokenization, generation, model loading, training, and backend/runtime regressions. None of the soft-edge pairs look like the same underlying bug or change, so all are rejected as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Different failures: CLIPVisionModel hidden_states regression vs attention-kernel divergence in a separate torch.compile branch." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an example script is unrelated to an unrecognized-model/config.json error." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs after sync_gpus and metric gathering in evaluation are different code paths and symptoms." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer load failure and a bare-except bug in Fuyu tensor conversion are unrelated issues." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but one is dequantized model saving and the other is custom module resolution after save_pretrained." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "A 3d_parallel example failure is not the same bug as tensor-parallel OOM loading Qwen2-57B." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to CPU/CUDA output divergence in PPDocLayoutV3." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "Tensor-parallel OOM and a key error loading MobileLLM are different loading failures." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "FlashAttention support for XLMRoberta is unrelated to a Trainer broadcast NoneType error." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output-shape behavior has nothing to do with missing PyTorch at import time." + }, + { + "left": "issue:44568", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer special-token regression and degenerate sequence-classification loss are separate bugs." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile SDPA failure for Bamba is unrelated to resume_from_checkpoint key mismatches." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "A missing import_utils helper and a kernels dependency/version import error are different availability/import problems." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to checkpoint resume key handling." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Inconsistent generated outputs and spurious recompilation are distinct chunked-generation issues." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "model_max_length behavior for DeBERTa-V3 is unrelated to the transformers.onnx import error." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Custom model/processor regression and a pipeline error-message regression are different bugs." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to Gemma3 eos_token_id configuration mismatch." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma3 EOS config inconsistency is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image-preprocessor dtype mismatch are separate model issues." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Backend install packaging and Gemma3TextConfig API mismatch are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "patchtsmixer post_init allowance and junk non-persistent buffers are different v5 regressions." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma3 EOS token config mismatch is unrelated to quantized-model save/load breakage." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model initialization breakage is not the same as GPTNeoXConfig rotary_pct not persisting." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "gguf_file loading regression and missing model_type config are different model-loading failures." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Fuyu bare-except behavior is unrelated to importing Trainer with datasets installed." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and Janus image-generation test failure are unrelated model-specific problems." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Fuyu exception handling bug is unrelated to a Windows ROCm torch.distributed import issue." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Mistral-3 weight tying and Ministral-3 FlashAttention failures affect different code paths and symptoms." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 35, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43065", + "issue:43352", + "issue:43388", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43572", + "issue:43650", + "issue:43673", + "issue:43688", + "issue:43761", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44190", + "issue:44479", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45406", + "issue:45542", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:46082", + "issue:46178", + "issue:46202", + "issue:46234", + "issue:46369", + "issue:46424", + "issue:46617", + "issue:46640", + "issue:46697" + ], + "soft_pairs": [ + "issue:45632|issue:46424", + "issue:43650|issue:46202", + "issue:44610|issue:46369", + "issue:43761|issue:44479", + "issue:43572|issue:46178", + "issue:45542|issue:46234", + "issue:43540|issue:44514", + "issue:43504|issue:45698", + "issue:43388|issue:43688", + "issue:45127|issue:45200", + "issue:43065|issue:44617", + "issue:43352|issue:46640", + "issue:44792|issue:44977", + "issue:44792|issue:45600", + "issue:43673|issue:44743", + "issue:41950|issue:43475", + "issue:43525|issue:43531", + "issue:44190|issue:44964", + "issue:35141|issue:42490", + "issue:44534|issue:44913", + "issue:44734|issue:45406", + "issue:46082|issue:46697", + "issue:45632|issue:46617", + "issue:43931|issue:44060", + "issue:43827|issue:43994" + ], + "prepared_review_unit_hash": "9ff6ecd7fd3ac383d155a2b4cd150d51b52b83b32430ed2f30f7b8ce669e26e8", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35141, + "estimated_input_tokens": 8786, + "estimated_eval_tokens": 17828 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set does not form one coherent duplicate cluster; the soft pairs are mostly unrelated issues that only share broad subsystem keywords. I rejected all soft edges as distinct bugs/changes.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical duplicate exists across the full set. If a representative issue must be chosen, issue 45698 is the broadest and most active open loading bug in the group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45698 is the best overall representative because it is open, broadly phrased, and has the highest discussion/reference activity among the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both involve loading/cache wording, but one is a trust_remote_code cache-path collision and the other is max_cache_length behavior; different bugs." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Unrelated topics: a placeholder issue versus a Gemma pipeline roundtrip/tokenization bug." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Different models and failure modes: input-size mismatch in OmDet-Turbo versus OOM in sam3-litetext." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to Qwen video-input regression." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLmConfig missing pad_token_idx is unrelated to a tensorflow_text backend mapping error." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Tensorboard/TensorFlow backend availability issue is unrelated to a missing import_utils helper." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are video/input processing bugs, but they affect different Qwen processors and different code paths." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy preset loading in Beit is a different bug from incorrect custom-module loading after save_pretrained." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Distributed metrics label truncation is unrelated to auxiliary-loss normalization." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA/extended-vocab collapse with tied embeddings is not the same as missing mm_token_type_ids defaults." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "A dummy Conv2d implementation issue is not the same as a SAM3 video memory overflow." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "Both mention Flash Attention 2, but the specific unsupported-model bug is not clearly the same as the generic issue." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus test failure and Qwen3.5 flash-attention generation bug are different models and symptoms." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "A failing image-generation test is unrelated to stale auto_mappings references to removed configs." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Chunked-prefill cache missing and recurrent states resetting are both cache-related but describe different mechanisms and code paths." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor mismatch is unrelated to a missing fpn_position_embeddings attribute in SAM3 video output." + }, + { + "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:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in a training script is unrelated to loading a multimodal model preset." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinitialization on resize is unrelated to import warnings from a trailing-slash local path." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization corruption is unrelated to GPTNeoX rotary_pct persistence." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation indexing and missing _tokenizer on Gemma4Processor are unrelated serve failures." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "LlamaConfig head_dim validation is unrelated to Swin2SR image over-padding." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both mention trust_remote_code/cache, but one is a leaf-directory collision and the other is a FileNotFoundError from cache loading; not the same bug." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch is unrelated to an incorrect tied-weights warning in Qwen3-Next." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is unrelated to a SigLIP2 model producing nonsensical outputs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 36, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43526", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43673", + "issue:43761", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44242", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44779", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44938", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45803", + "issue:45907", + "issue:46369", + "issue:46421", + "issue:46489", + "issue:46655", + "issue:46697" + ], + "soft_pairs": [ + "issue:44855|issue:45588", + "issue:43673|issue:46421", + "issue:43761|issue:44242", + "issue:44617|issue:45584", + "issue:45563|issue:45584", + "issue:43644|issue:44913", + "issue:25251|issue:27957", + "issue:44933|issue:45362", + "issue:43335|issue:44855", + "issue:45381|issue:46369", + "issue:44077|issue:44877", + "issue:44479|issue:44734", + "issue:43550|issue:44190", + "issue:41720|issue:45245", + "issue:43335|issue:44190", + "issue:44464|issue:45563", + "issue:43065|issue:44610", + "issue:45127|issue:45357", + "issue:39692|issue:44190", + "issue:44112|issue:45600", + "issue:45381|issue:46655", + "issue:43526|issue:46697", + "issue:44779|issue:46489", + "issue:44617|issue:46369", + "issue:43295|issue:45276", + "issue:44737|issue:45803", + "issue:43540|issue:44734", + "issue:44938|issue:45542", + "issue:45276|issue:45584", + "issue:44462|issue:45907" + ], + "prepared_review_unit_hash": "1d1f31db77022d49c8f36c38cab2f7385aac50e6054c80c46e21848f2c5fabcf", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36660, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a mix of unrelated issues from different models/features. The only strong duplicate-style match is the DeepSeek tokenizer regression pair (44779/46489); the rest look like distinct bugs or follow-up symptoms in different code paths.", + "confidence": 0.71, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:44779 is the best issue anchor because issue:46489 reads like a follow-up to the same DeepSeek tokenizer regression; the rest of the cluster is too heterogeneous to support a single canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Different subsystems and failures: Python 3.13 import/torch.jit parsing vs flash-attention sink handling." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill/cache, but one is a missing cache in v5.0 and the other is lazy-init causing recompilation; too different to treat as the same bug." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression and MoE load-balancing loss omission are unrelated issues." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is not the same bug as Whisper empty-transcription generation failure." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale warning for num_return_sequences is unrelated to Whisper failing on empty transcription." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer initialization junk and GPTNeoXConfig reload behavior are different code paths." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline top_k nesting bug and XLMRoberta Flash Attention 2 support are unrelated." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing image_utils import and Qwen3.5 chat crash are different problems." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "Sparse-layer config creation and Python 3.13 import parsing are unrelated." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen video vision-position-id regression and Sam3-lite memory overflow are different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Strict config loading for granite_speech and optional post_init on patchtsmixer are separate config issues." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both are generation regressions, but one is video-input position ids and the other is KV-cache continuation tensor indexing." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to local dataset loading in the example script." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Auto device mapping on Qwen3 and category cardinality limits are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Sparse-layer config bug and dataset loading failure are unrelated." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Chunked generation inconsistency and a stale warning are different issues." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor resolution mismatch are different bugs." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge with tied embeddings and incorrect visual encoder keys on save_pretrained are not the same problem." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "SigLIP2 documentation errors and local dataset loading are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Stale GraniteSpeech CI test and removed Sam3LiteText auto_mappings references are different CI breakages." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video/image processor bugs, but the model, symptom, and likely fix differ (vision_position_ids vs padding calculation)." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "Beit reduce_labels returning one label and Swin2SR over-padding are different image processor bugs." + }, + { + "left": "issue:44779", + "right": "issue:46489", + "accept": true, + "reason": "Both describe the same DeepSeek tokenizer regression on v5+, with 46489 explicitly reading like a follow-up gap in the earlier fix." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Sam3Video CUDA OOM and Sam3-lite memory overflow are related by model family only, not the same concrete bug." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer regression and resize_token_embeddings not updating Gemma4 embedding fields are different bugs." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet device placement bug and a bare-except in Fuyu tensor conversion are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video input validation and serve KV-cache indexing failure are unrelated." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import failure and tensorboard-without-TensorFlow backend error are different problems." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resize bug and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files missing revision forwarding are unrelated repository/API bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted soft edge, issue:44779 \u2194 issue:46489, is well supported by the titles as the same DeepSeek tokenizer regression with the later issue reading like a follow-up gap. The rest of the pairwise rejections are reasonable given the different models, subsystems, or symptoms described in the issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 37, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:43295", + "issue:43404", + "issue:43479", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44295", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44734", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45042", + "issue:45216", + "issue:45245", + "issue:45335", + "issue:45375", + "issue:45381", + "issue:45446", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:46082", + "issue:46121", + "issue:46234", + "issue:46424", + "issue:46506" + ], + "soft_pairs": [ + "issue:44521|issue:45381", + "issue:45632|issue:46234", + "issue:44246|issue:45003", + "issue:43854|issue:44964", + "issue:44964|issue:45600", + "issue:43643|issue:44496", + "issue:44112|issue:44964", + "issue:46082|issue:46121", + "issue:27957|issue:35141", + "issue:42175|issue:44279", + "issue:45245|issue:45584", + "issue:43404|issue:45216", + "issue:36246|issue:41628", + "issue:42796|issue:44521", + "issue:44186|issue:44734", + "issue:43650|issue:44220", + "issue:36246|issue:38617", + "issue:32090|issue:35141", + "issue:43479|issue:45375", + "issue:41950|issue:44016", + "issue:43531|issue:45600", + "issue:43673|issue:46424", + "issue:44295|issue:46506", + "issue:44016|issue:44220", + "issue:43976|issue:44987", + "issue:43295|issue:45335", + "issue:44683|issue:45446", + "issue:44496|issue:45042" + ], + "prepared_review_unit_hash": "5e6409a9076183c36c4eb08f35f3f7f4ec5e0be8edbe4df865451768d2b7e1ed", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35502, + "estimated_input_tokens": 8876, + "estimated_eval_tokens": 18008 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All soft-edge candidates are false positives: the items cover unrelated bugs, regressions, and feature requests across different models, pipelines, and import paths rather than the same underlying change or defect.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue stands out because none of the soft pairs look like true duplicates of the same bug or change.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue for deduplication purposes: the set is heterogeneous, so any single issue would be arbitrary rather than representative of a duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/Qwen-related, but one is about assistant masks in apply_chat_template and the other is wrong vision position ids; different bugs." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Both involve imports, but one is a trust_remote_code cache-path collision and the other is a missing utility symbol; not the same failure." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import-related symptoms differ: intermittent slow import vs unsafe sys.modules access; no shared code-path bug is evident." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-loading failures, but they concern different models and different root causes." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "One is a model load regression; the other is stale auto_mappings references breaking CI. Different problems." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Both are config/model-loading issues, but one is missing fields under trust_remote_code and the other is an unrecognized model_type error." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI stale device override test failure is unrelated to the Phi-4 multimodal model loading issue." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "One is LlamaConfig head_dim validation; the other is rope-validation list conversion. Different validation paths." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to token-embedding resize reinitialization." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "One is a missing TensorFlow backend in a torch install extra; the other is a general dependency issue without the same root cause." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Runtime category-limit error and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: untied lm_head weights in Mistral3 vs incorrect save_pretrained checkpointing in Qwen3.5." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they reference different missing symbols and different modules." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer BPE-marker handling is unrelated to multimodal assistant_mask generation." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/crash in NER inputs is not the same code-path as serve KV-cache continuation indexing." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "ADD THE DATA is not a substantive bug report and does not match the fbank feature issue." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different import failures: Qwen2.5 VL image processor vs layer_type_validation missing from configuration_utils." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer broadcast NoneType error is unrelated to embedding reinitialization after resize_token_embeddings." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "One is default vision/audio config initialization from None; the other is a missing deepstack_visual_indexes field dropped by strict parsing." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline using image processors is unrelated to a notebook syntax error." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and removed Sam3LiteText auto_mappings entries are separate issues." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Chunked-prefill cache-mixin regression is different from max_cache_length only applying to quantized cache." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "Registering position_ids as a buffer causing read errors is unrelated to CPU/CUDA output divergence." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and _torch_extract_fbank_features() bug are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version compatibility problem and a specific model load failure are not the same defect." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Regression in processor.tokenizer/image passing is unrelated to t5gemma resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and an incorrect version check for AuxRequest import are different code paths." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/model_type and PIL backend requiring torchvision are separate import/load issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 38, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:39692", + "issue:41950", + "issue:42796", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43526", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45020", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:46082", + "issue:46123", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46506", + "issue:46512", + "issue:46515", + "issue:46693" + ], + "soft_pairs": [ + "issue:43526|issue:43901", + "issue:43381|issue:45593", + "issue:45083|issue:46506", + "issue:43901|issue:44496", + "issue:42907|issue:43441", + "issue:43976|issue:44279", + "issue:43673|issue:44336", + "issue:43441|issue:45588", + "issue:44496|issue:45759", + "issue:44016|issue:44062", + "issue:44792|issue:46424", + "issue:43824|issue:44933", + "issue:43299|issue:45216", + "issue:46123|issue:46369", + "issue:44279|issue:46515", + "issue:44521|issue:46693", + "issue:45335|issue:45584", + "issue:46082|issue:46399", + "issue:44279|issue:45405", + "issue:43577|issue:45216", + "issue:39692|issue:43901", + "issue:43295|issue:44779", + "issue:44075|issue:44368", + "issue:43976|issue:45405", + "issue:46082|issue:46512", + "issue:44351|issue:45278", + "issue:43299|issue:45357", + "issue:42796|issue:43650", + "issue:41950|issue:45020", + "issue:44779|issue:45812" + ], + "prepared_review_unit_hash": "1795f04033fdc3935b413467df27000dcdc197108fe61a9fa051ebe5f4fda8ef", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36628, + "estimated_input_tokens": 9157, + "estimated_eval_tokens": 18570 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are clearly different bugs touching different models, APIs, or subsystems. The only plausible duplicate is the import/export regression pair around `HybridCache` and the broader 4.57.0\u21925.5.0 import failures.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45278 is the broadest representative of the import/export regression cluster, and it plausibly subsumes specific missing-symbol reports like 44351.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the strongest umbrella issue in this set because it describes the wider import breakage rather than a single symbol failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Different surfaces: Beit image processor label reduction vs pipeline docs mentioning `return_all_scores`." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated failures: gradient checkpointing in eval mode vs D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "Different code paths and symptoms: Qwen3 Omni feature-length helper vs CPU/CUDA discrepancy in PPDocLayoutV3." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs/API mismatch in text classification pipeline vs model-loading config error for an OLMo checkpoint." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family name, but different bugs: dequantized save failure vs FlashAttention crash." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "General dependency/version complaint is too broad to equate with a specific Python support regression." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Completely different areas: generation cache behavior vs ANSI output in loading report." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Ministral model-specific failure and the other is an integration crash on `s_aux=None`." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Both are multimodal/model-loading related, but the concrete bugs differ: unrecognized model type vs `text_config` unwrapping in `from_config`." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and tokenizer constructor argument conflict are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to cache-length configuration handling." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but for different missing symbols/modules and no clear shared code-path." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Same Qwen3.5/Qwen3VL family, but one is model loading and the other is `save_pretrained` checkpoint serialization." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Partial-batch mask generation bug and SAM memory overflow are different failures." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "Generic dependency issue is not the same as a specific torch 2.4.1 incompatibility." + }, + { + "left": "issue:44521", + "right": "issue:46693", + "accept": false, + "reason": "Assistant-mask generation for multimodal chat templates is unrelated to FA2 long-context performance regression." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Different models and behaviors: T5Gemma embedding resize vs Whisper empty transcription generation." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "Llama config validation and Qwen3.5 kernel registration crash are different bugs." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A broad dependency problem is not the same as the specific unreleased PEFT version bump." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Dtype handling in Blip2 loading and Qwen3.5 save_pretrained serialization are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Different documentation problems with different root causes: SigLIP2 example errors vs stale pipeline docs." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor/tokenizer API regression and Deepseek tokenizer output regression are different issues." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args being ignored is unrelated to a tied-embeddings warning." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support mismatch is not the same as a PEFT minimum-version packaging issue." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "One is config validation for `head_dim`; the other is stale `model.config.dtype` after `from_config`." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": true, + "reason": "`HybridCache` import failure looks like a specific instance of the broader 4.57.0\u21925.5.0 import/export breakage described in 45278." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5, but loading a composite config and saving incorrect visual encoder keys are separate bugs." + }, + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE-marker handling and a placeholder issue titled `ADD THE DATA` are not comparable duplicate candidates." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup and remote-code breakage are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Two tokenizer regressions, but for different model families and likely different code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45278 is clearly the broad umbrella import-regression issue, and 44351 is a specific HybridCache import failure that plausibly fits it. The other soft-edge rejections are conservative and tied to distinct titles/symptoms. No overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 39, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:33290", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43761", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44242", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44291", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45081", + "issue:45103", + "issue:45200", + "issue:45276", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45907", + "issue:46209" + ], + "soft_pairs": [ + "issue:43299|issue:43854", + "issue:44242|issue:45200", + "issue:43479|issue:43525", + "issue:44466|issue:45276", + "issue:41950|issue:45823", + "issue:43653|issue:45593", + "issue:44683|issue:45542", + "issue:43650|issue:46209", + "issue:44016|issue:44291", + "issue:44462|issue:45081", + "issue:43441|issue:45834", + "issue:44361|issue:45593", + "issue:45103|issue:45372", + "issue:42994|issue:43317", + "issue:33290|issue:45656", + "issue:45081|issue:45907", + "issue:44261|issue:44263", + "issue:45600|issue:45712", + "issue:44623|issue:44749", + "issue:42175|issue:43976", + "issue:44671|issue:44779", + "issue:43901|issue:43994", + "issue:44623|issue:45397", + "issue:43441|issue:44683", + "issue:43901|issue:44246", + "issue:45542|issue:45800", + "issue:43673|issue:43761" + ], + "prepared_review_unit_hash": "3068fc3fc8b799cb04769cd0da0a7013dc9c1b0915d8bf0c5cdcd4d3d49a2df4", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35201, + "estimated_input_tokens": 8801, + "estimated_eval_tokens": 17858 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No duplicate cluster emerges here: the soft pairs are mostly superficial matches across version regressions, model families, or broad subsystems rather than the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:44779 is the strongest standalone bug report in the set: it is a concrete regression with a clear v4\u2192v5 behavior change and a specific tokenizer/model impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different failures: Qwen3VL MoE loading regression vs a GLM-4.7-Flash unit-test/model-load issue; no shared code path is evident." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "One is missing load-balancing loss when router logits are off; the other is Gemma 4 text-only fine-tuning needing default mm_token_type_ids. Different mechanisms." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Both are config-related, but one is bad default vision/audio config initialization and the other is missing pad_token_id on Llama4Config. Not the same bug." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of tied lm_head weights vs resize_token_embeddings not updating Gemma4 embeddings/output layers. Different operations and symptoms." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video-classification pipeline processor lookup and Gemma4 PLE device mismatch are unrelated issues." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration/empty decode is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 vs tensorboard-only backend import issue; no common code path." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "A vague data request is unrelated to FP8 inference failing on B200." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error is unrelated to init_empty_weights passing unexpected _is_hf_initialized." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are both tokenizer-related but distinct bugs in different paths." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Ministral FlashAttention failure and Kosmos2.5 OCR long-input index error are different model-specific problems." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation handling and a mistral_common import error affect different subsystems." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized-model saving failure and dequantized offload loading failure are both quantization-adjacent but not the same concrete bug." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "AdaFactor OOM with DeepSpeed and optimizer.step being called twice with DeepSpeed are different optimizer/engine issues." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regex crash and list_repo_files not forwarding revision are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA q_a_layernorm config precision issue and GlmMoeDsaIndexer torch.split return values are different model internals." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "auto_mappings cleanup/CI breakage and leftover dummy_pt_objects classes are separate maintenance issues." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "processor.save_pretrained missing files and a Chinese performance regression report do not describe the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow in the torch extra and Python version support problems are unrelated packaging/environment issues." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions, but CamemBERT masked-LM output issues and Deepseek tokenizer correctness are different code paths and models." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs drift around return_all_scores is unrelated to SigLIP2 producing nonsensical outputs." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor save_pretrained missing files and Gemma-4 zero3 from_pretrained are different behaviors." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both mention attention, but Ministral FlashAttention failure and compiled flex_attention failure on torch>=2.9 are distinct problems." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Documentation mismatch and slow import behavior are unrelated." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "Tensorboard/TensorFlow backend mis-installation is unrelated to torch 2.4.1 vs transformers 5.8.0 incompatibility." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Cache-missing during chunked_prefill and CLIPVision hidden_states=None are different regression reports." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 40, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42886", + "issue:43299", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43643", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43976", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44589", + "issue:44871", + "issue:45003", + "issue:45103", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:45823", + "issue:46132", + "issue:46291", + "issue:46399", + "issue:46424", + "issue:46506", + "issue:46614", + "issue:46655", + "issue:46665", + "issue:46726" + ], + "soft_pairs": [ + "issue:43723|issue:44162", + "issue:46291|issue:46399", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:44016|issue:45823", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:44188|issue:44336", + "issue:44336|issue:44373", + "issue:46655|issue:46726", + "issue:41628|issue:43976", + "issue:43784|issue:45542", + "issue:44246|issue:44496", + "issue:44297|issue:45397", + "issue:45561|issue:45563", + "issue:46614|issue:46665", + "issue:45103|issue:46399", + "issue:43526|issue:45003", + "issue:44242|issue:45593", + "issue:42491|issue:43299", + "issue:43531|issue:45362", + "issue:44246|issue:44871", + "issue:44336|issue:44485", + "issue:44295|issue:45397", + "issue:44336|issue:46424", + "issue:42886|issue:46132", + "issue:43550|issue:46506", + "issue:41950|issue:44220" + ], + "prepared_review_unit_hash": "6052b48cf71ae98994277d8f4c4e10836ee45e6fa0388e8e02671e9555d010f0", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35355, + "estimated_input_tokens": 8839, + "estimated_eval_tokens": 17934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely unrelated Transformers bugs across different models, tokenizers, processors, and runtime paths. None of the soft pairs look like the same underlying bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.83, + "canonical_issue_reason": "41628 is the cleanest, most concrete regression report in this set: a straightforward top-level import failure with a clear user-facing symptom. It\u2019s a good anchor for the processor/import-related subset, even though the cluster is not truly duplicate-heavy.", + "canonical_pr_reason": null, + "best_issue_reason": "41628 is the best single issue candidate because it is specific, actionable, and easy to triage; the failure mode is crisp and well-scoped.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading regression vs a broad ESM2 workflow breakage; different model/path and no shared concrete bug." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both mention kernels, but one is a pipelines import/version constraint and the other is a Qwen3.5 registration crash; different code paths." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Specific missing model export vs a Python-version compatibility problem; not the same bug." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Llama4Config missing field vs glm-4v loading failure; different models and failure mechanisms." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device-mapping runtime mismatch; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id on a specific config vs missing fields from trust_remote_code AutoConfig; related in theme only, not the same defect." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Storage class import/serialization issue vs Qwen3.5 chat crash; different bug surfaces." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile attention-kernel divergence vs terminal ANSI output in loading_report; unrelated." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI escape generation bug vs a wrong docstring; not the same code issue." + }, + { + "left": "issue:46655", + "right": "issue:46726", + "accept": false, + "reason": "SmolVLM video padding calculation bug vs AutoProcessor backend-missing import/reporting issue; different processor failures." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Top-level import failure for AutoImageProcessor vs Python version support issue; unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Missing nn import in sentence-transformers integration vs undefined tf backend from tensorboard-only install; different packages and causes." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slow-path performance issue vs unrecognized model/config loading failure; not the same bug." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "tokenizer.save_pretrained metadata mismatch vs Gemma-4 zero3 from_pretrained failure; different behaviors." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "pytest-xdist race condition vs stale generate() warning; unrelated test/runtime issues." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained behavior inside Python classes; too different to be duplicates." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "auto_docstring crash under future annotations vs kernels registration error in Qwen3.5; unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access; different modules and symptoms." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "MoE load-balancing loss not added vs D-FINE auxiliary losses disabled; both are loss-related but on different models/code paths." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both are Qwen MoE loading regressions, but they involve different model families/version contexts and not clearly the same concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue vs Qwen3.5 chat crash; not the same failure mode." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import performance issue vs Gemma-3 eos_token_id inconsistency; unrelated." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output bug vs GLM-5 RoPE implementation discussion; different areas." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "position_ids buffer read error vs Gemma-4 zero3 loading problem; not the same defect." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "ANSI codes in loading_report vs cache max_cache_length behavior; unrelated." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline tokenizer cache loading failure vs AttentionInterface.register behavior change; different mechanisms." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "torch.compile+SDPA failure in Bamba vs CPU/CUDA result mismatch in PPDocLayoutV3; different models and bugs." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification processor lookup bug vs _torch_extract_fbank_features issue; different pipelines and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 41, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36246", + "issue:42757", + "issue:43257", + "issue:43262", + "issue:43295", + "issue:43352", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43673", + "issue:43901", + "issue:44038", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44351", + "issue:44466", + "issue:44479", + "issue:44496", + "issue:44561", + "issue:44623", + "issue:44734", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44877", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45459", + "issue:45561", + "issue:45584", + "issue:45600", + "issue:45759", + "issue:45859", + "issue:46170", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46506", + "issue:46655", + "issue:46726" + ], + "soft_pairs": [ + "issue:44779|issue:45292", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43901|issue:45003", + "issue:43257|issue:45137", + "issue:45003|issue:45561", + "issue:44734|issue:46424", + "issue:44496|issue:45003", + "issue:43540|issue:46655", + "issue:45859|issue:46399", + "issue:43673|issue:44561", + "issue:44162|issue:45200", + "issue:44479|issue:46655", + "issue:43262|issue:44246", + "issue:45375|issue:45759", + "issue:43475|issue:46369", + "issue:44242|issue:44857", + "issue:44295|issue:44749", + "issue:45137|issue:45600", + "issue:44877|issue:45375", + "issue:36246|issue:43596", + "issue:45292|issue:45584", + "issue:43295|issue:45292", + "issue:44561|issue:46399", + "issue:44466|issue:45335", + "issue:44623|issue:46726", + "issue:44987|issue:45459", + "issue:44038|issue:46506", + "issue:44479|issue:46170", + "issue:43352|issue:45250" + ], + "prepared_review_unit_hash": "28aed8c2dcde9c5b715a50887977c35d92a2418aed66cbbb9fa2b176ce54160a", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36758, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly heterogeneous and do not form a strong duplicate cluster. The soft matches are mostly superficial topical overlap (imports, cache, Qwen, DeepSpeed, video, embeddings) rather than the same concrete bug.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 36246 is the most concrete, self-contained bug report in the set: a specific missing export/import path with a clear failure mode and likely fix surface. The rest of the issues are broader or on different code paths.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 36246 is the best representative issue because it is narrowly scoped, reproducible, and actionable compared with the more ambiguous or broad reports in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Both mention transformer behavior regressions, but one is a DeepSeek tokenizer correctness issue and the other is resize_token_embeddings propagation; different code paths and failures." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for unrelated symbols and dependencies: a missing transformers vision processor export vs huggingface_hub.is_offline_mode." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing-name import errors, but they concern different exported objects and likely different fixes." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed, but one is Qwen3 MoE weight conversion and the other is a ZeRO3 deque/indexing failure; not the same bug." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and pytest-xdist file races are distinct issues in different subsystems." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "Both mention KV/cache behavior, but one is a serve continuation indexing crash and the other is cache length config handling; different concrete failures." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-processing bugs, but they affect different models and different failure points." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is CPU offload incompatibility and the other is a registration/decorator crash with use_kernels=True." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "GenerationMixin chunked_prefill cache missing and removal of is_torch_fx_available are unrelated regressions." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids handling are unrelated model-specific bugs." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-input regressions, but they target different models and likely different preprocessing code paths." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat template sample-rate defaulting and import performance have no substantive overlap." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both concern Qwen multimodal config handling, but one is a missing config field and the other is failure to unwrap text_config; different root causes." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "SAM 3 Video output attribute loss and SAM3 LiteText memory overflow are different model bugs." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Load-balancing loss logic and an AMP/CUDA crash in LwDetrImageLoss are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "A position_ids buffer access error and a Chinese performance regression are not the same bug." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque underflow and broken auto_mappings references are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict config loading for granite_speech and missing Qwen visual indexes are different config-validation issues." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Importing a missing processor class and a DeepSpeed zero3/BertModel index error are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "resize_token_embeddings propagation and Whisper empty-transcription handling are different model behaviors." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "A regression in processor.tokenizer/image passing is unrelated to token embedding resizing behavior." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "Removal of an FX helper breaking trust_remote_code models is unrelated to Qwen3.5 rotary embedding registration." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both involve embeddings/weights, but one is lm_head serialization and the other is decoder.embed_tokens not updating after resize." + }, + { + "left": "issue:44623", + "right": "issue:46726", + "accept": false, + "reason": "processor.save_pretrained file omissions and AutoProcessor failing without a vision backend are different processor-loading/saving issues." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "A model-loading failure for physical-intelligence/fast and tokenizer error masking when protobuf is missing are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug and CPU/CUDA result divergence in PPDocLayoutV3 are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Video-input regression in Qwen/VL models and Windows ROCm torch.distributed import issues are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A model-specific Flash Attention 2.0 unsupported error is not the same as a generic Flash Attention 2.0 issue title." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 42, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:42886", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43540", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43701", + "issue:43723", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44871", + "issue:44877", + "issue:44977", + "issue:44995", + "issue:45200", + "issue:45229", + "issue:45397", + "issue:45412", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45656", + "issue:45685", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421", + "issue:46431", + "issue:46710", + "issue:46726" + ], + "soft_pairs": [ + "issue:44060|issue:44077", + "issue:45412|issue:46123", + "issue:43653|issue:44162", + "issue:44368|issue:44977", + "issue:44877|issue:46170", + "issue:44995|issue:45200", + "issue:43475|issue:43540", + "issue:42886|issue:45229", + "issue:44368|issue:45563", + "issue:43931|issue:44368", + "issue:43065|issue:43408", + "issue:45397|issue:45656", + "issue:44995|issue:45588", + "issue:45987|issue:46082", + "issue:43723|issue:46710", + "issue:44188|issue:46202", + "issue:43906|issue:46202", + "issue:44483|issue:46202", + "issue:44188|issue:46421", + "issue:45584|issue:46170", + "issue:45685|issue:46431", + "issue:43065|issue:46123", + "issue:44361|issue:44749", + "issue:44297|issue:44623", + "issue:43994|issue:46726", + "issue:43646|issue:44162", + "issue:44749|issue:45656", + "issue:44871|issue:46132", + "issue:43232|issue:43673", + "issue:43701|issue:44749" + ], + "prepared_review_unit_hash": "ad5ee9157271483980af2ee70f48fff82238388b826dfd5d20d956caf81bcef3", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36117, + "estimated_input_tokens": 9030, + "estimated_eval_tokens": 18316 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: tokenizer/serialization, generation/cache, model-loading, vision, and deepspeed issues are all mixed together. None of the soft pairs look like the same underlying bug or change, so they should not be merged.", + "confidence": 0.82, + "canonical_issue_reason": "issue:44297 is the most concrete, self-contained serialization bug in the set, and it is the closest to another save_pretrained-style report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44297 has the clearest failure mode and the most focused code path, making it the best representative issue here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init policy are unrelated subsystems and failure modes." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory retention and MaskGenerationPipeline partial-batch dropping are different runtime behaviors." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated tokenizer/model bugs." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tie_word_embeddings warning for Qwen3.5 is not the same bug as flash-attention generation failure." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading for granite_speech and Windows ROCm import errors have no shared code path." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale indexer cache crash and a Gemma4 mm_token_type_ids default issue are different defects." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Missing fpn_position_embeddings in SAM3 and Qwen3OmniMoe video input validation are distinct bugs." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Tokenizer offline cache loading and multi-GPU CUDA OOM are unrelated problems." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A model config warning and a paged-generate stale warning for num_return_sequences are not the same issue." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch is unrelated to the Qwen3.5 tie-word-embeddings warning." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and SAM3 model-type mismatch warning are different SAM3 issues." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "gemma-4 from_pretrained/zero3 loading and double optimizer-step behavior under deepspeed are separate bugs." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "GlmMoeDsa stale cache crash and flash_attention s_aux=None AttributeError affect different code paths." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache return-type contract and LlamaConfig head_dim validation are unrelated API changes." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading in v5 and an accuracy regression after upgrade are too broad and not the same defect." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile and Gemma pipeline roundtrip on \u2581 are different issues." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of another issue and a Gemma roundtrip failure are not clearly the same bug." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "chat/completions request rejection and Gemma pipeline roundtrip failure involve different APIs." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "torch.compile attention branching and chunked-prefill cache lazy init are separate generation/runtime problems." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty transcription failure and Windows ROCm import dependency issues are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram on Int in moe code and missing chat template in a processor are unrelated errors." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer and partial-batch result dropping are different code paths." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError and a performance slowdown after upgrade are not the same underlying bug." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained mismatch and processor.save_pretrained missing files are similar on surface but different serialization failures." + }, + { + "left": "issue:43994", + "right": "issue:46726", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and AutoProcessor backend-missing detection are different vision-stack issues." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage in v5 and ESM2 breakage are broad but distinct regressions." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Training slowdown after upgrade and a deepspeed double-step bug do not share a concrete code path." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "sync_gpus generation kwargs update and chunked_prefill cache-missing regression are different generation bugs." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "resume_from_checkpoint key mismatch and a filtering slowdown after upgrade are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 43, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:36010", + "issue:39692", + "issue:41628", + "issue:42371", + "issue:42673", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43986", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44351", + "issue:44623", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45406", + "issue:45478", + "issue:45656", + "issue:45753", + "issue:46202", + "issue:46326", + "issue:46399", + "issue:46701", + "issue:46726" + ], + "soft_pairs": [ + "issue:42994|issue:43482", + "issue:44623|issue:45656", + "issue:43475|issue:43643", + "issue:45375|issue:45478", + "issue:28282|issue:30990", + "issue:41628|issue:44933", + "issue:44869|issue:45356", + "issue:42371|issue:45092", + "issue:44964|issue:45137", + "issue:45310|issue:45375", + "issue:43986|issue:46726", + "issue:42994|issue:43525", + "issue:44038|issue:44623", + "issue:45230|issue:45397", + "issue:39692|issue:46726", + "issue:43475|issue:45381", + "issue:43723|issue:46326", + "issue:43784|issue:46399", + "issue:43232|issue:44188", + "issue:45406|issue:46202", + "issue:43873|issue:44945", + "issue:44945|issue:45753", + "issue:41628|issue:44351", + "issue:46202|issue:46701", + "issue:45081|issue:45103", + "issue:42994|issue:43824", + "issue:42673|issue:43704", + "issue:36010|issue:41628", + "issue:43278|issue:45137", + "issue:44038|issue:45310", + "issue:44222|issue:45310" + ], + "prepared_review_unit_hash": "343973e152ad19dece89e723bbf5d999649f7029d393734f14efe0b5fa5e5ad4", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36791, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issue set, but there are two clear duplicate pockets: the Qwen3ForCausalLM multi-thread VRAM leak and the Qwen3.5 MoE from_pretrained/loading regression. The Qwen3.5 config-field report is the best canonical issue.", + "confidence": 0.63, + "canonical_issue_reason": "issue:45375 is the most specific/root-cause report in the Qwen3.5 MoE loading-error cluster; the other matching reports are more generic symptom variants.", + "canonical_pr_reason": null, + "best_issue_reason": "Most representative item overall: it captures the clearest diagnosis in the strongest duplicate cluster in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized save failure vs GGUF loading failure; related theme but different operations and failure modes." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "processor.save_pretrained file omission vs DeepSpeed optimizer double-step; unrelated code paths." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM3 missing attribute vs trust_remote_code returning missing fields; both are attribute errors but not the same bug." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": true, + "reason": "Same Qwen3.5 MoE from_pretrained regression; one report is the specific config-field cause, the other is the generic failure symptom." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch-missing ImportError vs sentence-transformers loading hang; different root causes." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Different missing-import symbols (AutoImageProcessor vs image_utils export); same area but not the same defect." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash vs tokenizer codec regression; unrelated." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API warning vs InternVL2 meta-init incompatibility; different subsystems." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal loading failure vs DeepSpeed ZeRO3 deque pop; unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": true, + "reason": "Same Qwen3.5 MoE from_pretrained failure; the detailed config-field report and the generic loading error describe the same regression." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both involve AutoProcessor/backend handling, but one is video-model/no-torchvision and the other is image-processor recognition; too broad to merge." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model save bug vs Llama4 config missing pad_token_id; different issues." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe transformers v5 bug vs processor.save_pretrained missing files; unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Placeholder 'bug report' provides no concrete linkage to the gemma-4 ZeRO3 from_pretrained error." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "SigLIP2 doc mistakes vs missing-vision-backend AutoProcessor error; different problems." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "SAM3 attribute error vs qwen2.5-vl video position-id bug; different models and code paths." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "AutoTokenizer loading issue in v5 vs mistral_common continue_final_message behavior; different tokenizer bugs." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "sentence-transformers import NameError vs Qwen3.5 kernel registration failure; unrelated." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "_update_model_kwargs_for_generation after sync_gpus vs attention-kernel divergence under torch.compile; different generation bugs." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer vs Gemma pipeline roundtrip on \u2581; related model family but distinct failures." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading issue vs pipeline-parallelism wrong output; different runtime behaviors." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallelism incorrect output vs Qwen3_5 infinite loop on a specific image; distinct bugs." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports (AutoImageProcessor vs HybridCache); same symptom class but not the same defect." + }, + { + "left": "issue:46202", + "right": "issue:46701", + "accept": false, + "reason": "Gemma pipeline token roundtrip vs DiffusionGemma image-text-to-text pipeline failure; related family but different path." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Mistral regex tokenizer crash vs auto_docstring future-annotations crash; unrelated utility bugs." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized save failure vs missing Qwen2.5-VL export; different import/load problems." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact duplicate Qwen3ForCausalLM VRAM leak when used across multiple dataloader threads." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import error vs AutoImageProcessor import error; different missing symbols." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype regression in evaluate vs ZeRO3 deque underflow; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Different Qwen model families (Qwen3-VL-Moe vs Qwen3.5 MoE); similar shape but not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained issue vs Qwen3.5 from_pretrained error; different code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering summary is grounded in the packet, and the soft-edge calls are generally conservative. The two accepted pockets are supported by the titles: the exact duplicate Qwen3ForCausalLM VRAM leak pair, and the Qwen3.5 MoE from_pretrained regression cluster where 45375 is a more specific config-field report. The rejected pairs mostly involve clearly different operations or subsystems, so I don\u2019t see overstatement here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 44, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:27957", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42222", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:43116", + "issue:43232", + "issue:43295", + "issue:43377", + "issue:43479", + "issue:43504", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44016", + "issue:44079", + "issue:44190", + "issue:44222", + "issue:44242", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45561", + "issue:45600", + "issue:45712", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46231", + "issue:46693" + ], + "soft_pairs": [ + "issue:43232|issue:44792", + "issue:42831|issue:44222", + "issue:30333|issue:33290", + "issue:43504|issue:45561", + "issue:42175|issue:44933", + "issue:45561|issue:45600", + "issue:44496|issue:45561", + "issue:43976|issue:44933", + "issue:30990|issue:35141", + "issue:42175|issue:45712", + "issue:42898|issue:44568", + "issue:46077|issue:46123", + "issue:43377|issue:44242", + "issue:44623|issue:45310", + "issue:43479|issue:44869", + "issue:42673|issue:46693", + "issue:44016|issue:46231", + "issue:42898|issue:45812", + "issue:43723|issue:44568", + "issue:42222|issue:45200", + "issue:43874|issue:45092", + "issue:42898|issue:44779", + "issue:43479|issue:44079", + "issue:43784|issue:43824", + "issue:44734|issue:45103", + "issue:27957|issue:30333", + "issue:30333|issue:34689", + "issue:42898|issue:45356", + "issue:42898|issue:43295", + "issue:43116|issue:44190" + ], + "prepared_review_unit_hash": "e1a4149fbe6d645710d689d1bb895aefdc395395a9f16ea037884cafed739f62", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the issues span unrelated bugs in tokenization, model loading, training, packaging, tests, and inference. None of the soft-edge pairs look like the same underlying defect, so this is not a true duplicate cluster.", + "confidence": 0.93, + "canonical_issue_reason": "No obvious duplicate centerpiece exists; if one issue must represent the cluster, issue 46693 is the most actionable open regression with a concrete performance problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46693 is the strongest standalone bug report here: it is open, specific, and describes a reproducible performance regression. The rest are mostly unrelated closed issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Different failures: generation kwargs handling vs a janus image-generation test regression." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both mention FP8, but one is an accuracy issue and the other is save_pretrained/moe serialization." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLFlow reporting on failed jobs is unrelated to a deepspeed adafactor OOM." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Model preset loading via a legacy field is unrelated to a pytest-xdist file race." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Packaging/backend install behavior is unrelated to a missing import symbol in image_utils." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "A test-file race is not the same bug as stale auto_mappings references to removed configs." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Unrecognized model/config loading failure is unrelated to a testing-utils race condition." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility is a different issue from a nonexistent import in image_utils." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence-transformers loading hangs vs output embedding reinitialization during resize are different code paths." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Install extra backend packaging differs from dummy class cleanup and repo-check failures." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Tokenization cleanup behavior in v5 is unrelated to BOS/EOS tokens not being added." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Gemma4 attention-mask construction and partial-batch pipeline result dropping are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask output mismatch is unrelated to missing load-balancing loss for router logits." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Missing processor files on save is a different bug from Qwen3.5 MoE from_pretrained loading failure." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Default multimodal config initialization is unrelated to Whisper word-timestamp decoding crash." + }, + { + "left": "issue:42673", + "right": "issue:46693", + "accept": false, + "reason": "VRAM leak in multi-threaded dataloading is distinct from a long-context FA2 performance regression." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "A notebook syntax error is not the same as a documentation parameter-name mismatch." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer cleanup behavior changes are unrelated to Granite token-id regression." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading in v5 is not the same defect as add_special_tokens failing to add BOS/EOS." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Broken vitpose model files and Gemma4 token-type defaults concern different models and code paths." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "A missing image-processor method is unrelated to old InternVL2 checkpoints failing meta initialization." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Tokenization cleanup behavior is unrelated to Deepseek tokenizer codec handling." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Default config initialization and ModelOutput key assignment are separate bugs." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Import-time NameError in sentence-transformers is unrelated to a missing Qwen2.5-VL export." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation indexing in serve has nothing to do with auto_docstring annotations handling." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to MLFlow reporting after failed training jobs." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Failed-job MLFlow reporting is unrelated to a Llama 3.2 vision model-loading regression." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Tokenization-space cleanup changes are unrelated to Kimi-K2.5 codec handling regression." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenization cleanup behavior is unrelated to processor.tokenizer/image-token passing regression." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Empty multi-label classification outputs in the example script are not the same as local dataset loading failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 45, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:35141", + "issue:36331", + "issue:41628", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:43122", + "issue:43232", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:44060", + "issue:44075", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44448", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45245", + "issue:45310", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45834", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46231", + "issue:46421", + "issue:46424", + "issue:46665", + "issue:46710" + ], + "soft_pairs": [ + "issue:45405|issue:45712", + "issue:44060|issue:45125", + "issue:44295|issue:45656", + "issue:30333|issue:35141", + "issue:30333|issue:36331", + "issue:22355|issue:28282", + "issue:44933|issue:45712", + "issue:44075|issue:46231", + "issue:44222|issue:44297", + "issue:44448|issue:46132", + "issue:42831|issue:44623", + "issue:42796|issue:43723", + "issue:44297|issue:45656", + "issue:43479|issue:43644", + "issue:45245|issue:46132", + "issue:44222|issue:44623", + "issue:43479|issue:46665", + "issue:41628|issue:42757", + "issue:45071|issue:46082", + "issue:44373|issue:44485", + "issue:46132|issue:46710", + "issue:44297|issue:45310", + "issue:44263|issue:45092", + "issue:43723|issue:44977", + "issue:44222|issue:44484", + "issue:44190|issue:46082", + "issue:43232|issue:46424", + "issue:44206|issue:45103", + "issue:43122|issue:46209", + "issue:43232|issue:45834", + "issue:44734|issue:46421" + ], + "prepared_review_unit_hash": "95f36a4ddca1d29977f44bfb891f7bb7f2f37f120ab5ddfddd058ab9f1c3a3ef", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No clear duplicate cluster here: the items span many unrelated Transformers issues, so all soft edges should be rejected.", + "confidence": 0.86, + "canonical_issue_reason": "No single canonical issue; the cluster is heterogeneous (imports, tokenizer serialization, training, generation, config, FP8/cache, and docs bugs) rather than one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44297 is the most concrete and self-contained bug report, but it only covers a narrow tokenizer save_pretrained mismatch and does not represent the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Dependency version bump vs leftover dummy-class cleanup; different problems." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both involve Qwen MoE, but one is a tied-weights warning and the other is missing tensor-parallel metadata." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer read error vs optimizer step duplication under deepspeed; unrelated." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLflow job reporting failure vs embedding reinitialization after resize; different code paths." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting bug vs custom trainer loss signature mismatch; not the same issue." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Both are import errors, but for different modules and causes." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Nonexistent image_utils import vs dummy class cleanup; different fixes." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer arguments ignored vs docstring parameter-name mismatch." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained bug vs tokenizer save_pretrained metadata mismatch." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Generation output regression vs attention registration behavior change; not the same defect." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy issue vs processor.save_pretrained missing files." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "Tokenizer BPE marker handling vs AutoTokenizer loading failure in v5." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config mismatch vs optimizer step behavior; unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Phi4 multimodal config defaults vs non-persistent buffer junk; different underlying bugs." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category limit error vs attention registration behavior." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "FP8 save_pretrained issue vs processor save_pretrained missing files." + }, + { + "left": "issue:43479", + "right": "issue:46665", + "accept": false, + "reason": "Config default initialization bug vs from_pretrained behavior inside Python classes." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Different import failures from different packages/modules." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking regression vs LlamaConfig head_dim validation." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo vs RoPE implementation discussion; not the same bug." + }, + { + "left": "issue:46132", + "right": "issue:46710", + "accept": false, + "reason": "Attention registration behavior vs model accuracy regression; too broad to merge." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save_pretrained metadata bug vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return values in an indexer vs remote-code/meta-init incompatibility." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading issue in v5 vs Qwen3.5 flash-attention generation bug." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained bug vs max_shard_size question." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Example script dataset loading issue vs LlamaConfig head_dim validation." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update after sync_gpus vs cache max_cache_length behavior." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Feature extractor center-arg regression vs auto-docstring AttributeError." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer output drift across versions vs FP8 inference failure on B200." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs update bug vs Kosmos2.5 OCR index error." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "Serve KV-cache indexing crash vs lazy cache initialization causing recompilation." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 46, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43644", + "issue:43646", + "issue:43824", + "issue:44038", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44568", + "issue:44623", + "issue:44829", + "issue:45092", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46369", + "issue:46402", + "issue:46431", + "issue:46655" + ], + "soft_pairs": [ + "issue:43475|issue:46305", + "issue:46209|issue:46431", + "issue:46305|issue:46655", + "issue:44263|issue:44360", + "issue:44295|issue:44623", + "issue:30990|issue:38175", + "issue:43644|issue:46305", + "issue:44261|issue:45092", + "issue:45092|issue:45357", + "issue:44263|issue:44373", + "issue:45381|issue:46305", + "issue:42907|issue:46050", + "issue:44279|issue:46328", + "issue:42831|issue:45310", + "issue:44829|issue:46002", + "issue:22355|issue:27957", + "issue:44483|issue:45406", + "issue:41628|issue:43824", + "issue:42831|issue:44038", + "issue:43408|issue:46402", + "issue:22355|issue:28218", + "issue:42831|issue:45753", + "issue:39401|issue:44568", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:44188|issue:44373", + "issue:43232|issue:45593", + "issue:43408|issue:46369", + "issue:44261|issue:44373", + "issue:43646|issue:46002" + ], + "prepared_review_unit_hash": "dd1816e8c7cee2b7b2bf961a90fc05f0abbbe35c893822db821607f24d56e65e", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36817, + "estimated_input_tokens": 9205, + "estimated_eval_tokens": 18666 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All candidate soft pairs are rejects: they only share broad themes like tokenizers, loading, quantization, or SAM3, but they describe different concrete bugs and code paths. No single issue stands out as a canonical representative for this mixed set.", + "confidence": 0.94, + "canonical_issue_reason": "No canonical issue exists here: the items are a mixed bag of unrelated bug reports, and the superficially similar pairs still point to different models, features, or failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "None of the issues is a strong global representative because the set is heterogeneous and the closest-looking pairs are not the same underlying defect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "Different SAM3 vision-encoder attribute error vs MoonshineStreaming mask off-by-one; unrelated code paths." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 B200 inference failure is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Both involve video processing, but they are different bugs in different models and functions." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Different DSA indexer problems: torch.split return values vs missing ReLU." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Registering position_ids as a buffer is a different issue from processor.save_pretrained missing files." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and SigLIP2 zero probabilities are different failures." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk in Transformers 5.0 is unrelated to MoonshineStreaming masking." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA RMS eps precision issue is unrelated to InternVL2 meta-initialization incompatibility." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Different models and bugs: InternVL2 meta init vs Qwen3.5 save_pretrained key handling." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Indexer torch.split behavior is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both concern quantized save/load, but they affect different quantization flows and concrete failure modes." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Generic dependency issue is not the same as SequenceFeatureExtractor.pad inefficiency." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue and Qwen3.5 Moe from_pretrained error are different bugs." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Flash Attention 3 degenerate training is not the same as NaNs after checkout/init." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "onnx import missing and XLMRoberta Flash Attention 2 are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat completions request rejection differs from Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different causes." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FineGrainedFP8 accuracy and Qwen3-VL-Moe are not the same underlying defect." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Wrong model-type warning and missing text encoder weights are distinct SAM3 issues." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing transformers.onnx module and extra whitespace after added token are different tokenizer/import problems." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy issue is unrelated to a Qwen3_5 infinite loop on a specific image." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Wrong offset_mapping and missing BOS/EOS insertion are separate tokenizer bugs." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "A specific is_offline_mode import failure is only one of many unrelated import regressions." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 on XLMRoberta and Sentence Transformers loading hang are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer problems." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving failure are unrelated." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to a docstring issue." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and D-FINE auxiliary-loss behavior are different code paths." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "SAM3 model-type warning and SAM3-litetext memory overflow are different failures." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA layernorm precision bug is unrelated to the position_ids docstring." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads are not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 47, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:22355", + "issue:29159", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42796", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43576", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44488", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:46326", + "issue:46614" + ], + "soft_pairs": [ + "issue:42796|issue:43122", + "issue:42371|issue:44263", + "issue:42907|issue:44945", + "issue:30990|issue:33357", + "issue:43576|issue:43784", + "issue:43756|issue:44060", + "issue:39401|issue:44488", + "issue:42907|issue:44964", + "issue:43976|issue:44945", + "issue:41628|issue:45278", + "issue:44295|issue:44297", + "issue:43784|issue:45600", + "issue:42907|issue:43937", + "issue:44117|issue:44704", + "issue:44625|issue:44829", + "issue:44360|issue:44485", + "issue:39401|issue:44521", + "issue:44704|issue:45290", + "issue:43408|issue:43475", + "issue:43122|issue:44521", + "issue:43122|issue:46326", + "issue:42907|issue:43452", + "issue:22355|issue:29159", + "issue:44038|issue:44297", + "issue:43116|issue:45200", + "issue:39401|issue:46614", + "issue:45216|issue:45478", + "issue:44038|issue:44222", + "issue:44038|issue:44295", + "issue:43824|issue:44351", + "issue:43122|issue:44779", + "issue:36296|issue:43116", + "issue:42757|issue:44351" + ], + "prepared_review_unit_hash": "0979ba55038e42903adc8151920e6ca61cdb2660ff6815b777106a32cb926892", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37218, + "estimated_input_tokens": 9305, + "estimated_eval_tokens": 18866 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is loosely related by v5 regressions and import/tokenizer breakages, but most pairs are not true duplicates. The broadest representative is the v5 upgrade/import-regression issue.", + "confidence": 0.39, + "canonical_issue_reason": "issue:45278 is the broadest umbrella here: it describes multiple import failures after upgrading to 5.5.0, which matches the cluster\u2019s strongest recurring theme better than the more specific model- or tokenizer-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best single issue to represent this cluster because it is the most general and central regression report among a set of mostly version-related import/tokenizer problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer-related, but one is a model-specific BPE-marker bug and the other is a broader v4\u2192v5 tokenization regression; not the same underlying defect." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "Unrelated: TF32 settings vs. a DSA indexer ReLU/math bug." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Different failures in different code paths: dequantized save/load vs pipeline-parallel output correctness." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Both mention loading, but they concern different models and failure modes (stuck load vs Mac bus error)." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Both are version-5 breakages, but one is a broken CLI command and the other is a missing import during sentence-transformers import; not the same bug." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Completely different model-architecture bugs: RoPE layer count vs tied-weights warning." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer/model issues: wrong offset mappings vs failure to load a specific model." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Save/dequantization bug is unrelated to a multimodal model-loading failure." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "A Python-version compatibility issue is not the same as a pipeline-parallelism output bug." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but one is a missing symbol and the other is a broad set of upgrade import errors; too different to merge." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Unrelated: buffer access/readback bug vs tokenizer save_pretrained metadata mismatch." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Different subsystems and failures: missing nn import vs stale auto_mappings references." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Different model families and error types: save_pretrained dequantization vs invalid GenerationConfig." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Both involve from_pretrained APIs, but one is tokenizer mapping None handling and the other is AutoProcessor kwargs propagation; not the same bug." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Different issues: config propagation for Qwen3.5 vs degenerate training with flash_attention_3." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA math/activation discussion is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Both are tokenizer-adjacent, but one is offset_mapping correctness and the other is assistant mask generation for multimodal chat templates." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Unrelated API bugs: cached_file kwargs forwarding vs chat-template crashing on tool-call messages." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Same model family, but different concrete bugs: model-type warning vs missing output attribute." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Both concern tokenization behavior, but one is a general regression between versions and the other is multimodal assistant-mask generation." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Different tokenizer regressions in different functions/models; not the same underlying bug." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Unrelated failures: dequantized save and gguf loading behavior are different code paths." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Different topics: missing transformers.onnx module vs slow/fast tokenizer inconsistency." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Different bugs: Qwen3-VL-Moe model issue vs tokenizer_config save metadata mismatch." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script classification output bug vs Gemma 4 multimodal token-type-id requirement; unrelated." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Both are tokenizer-related, but they concern different tokenizers and distinct state bugs." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Related only at a high level (Qwen3.5 and v5.4.0), but one is save_pretrained checkpoint correctness and the other is from_pretrained loading; not enough evidence of the same concrete bug." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Different MoE/model-save problems; no shared concrete failure path." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Unrelated: Qwen3-VL-Moe loading bug vs position_ids buffer access bug." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing exports and likely different root causes." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizers/models and are not the same bug." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Completely unrelated: tensor-parallel training vs example-script classification results." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different packages; no shared 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": 1, + "component_count": 23, + "review_unit_index": 48, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:22355", + "issue:28348", + "issue:29127", + "issue:38617", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42898", + "issue:43066", + "issue:43116", + "issue:43278", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44188", + "issue:44355", + "issue:44485", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44938", + "issue:44945", + "issue:44995", + "issue:45072", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45753", + "issue:46002", + "issue:46231", + "issue:46328", + "issue:46614", + "issue:46640", + "issue:46697", + "issue:46710" + ], + "soft_pairs": [ + "issue:43278|issue:44945", + "issue:22355|issue:29127", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:44188|issue:46328", + "issue:43937|issue:44485", + "issue:38617|issue:45278", + "issue:42898|issue:46614", + "issue:43611|issue:46002", + "issue:43116|issue:43493", + "issue:44355|issue:46231", + "issue:43066|issue:44779", + "issue:43976|issue:45446", + "issue:44038|issue:45478", + "issue:44625|issue:46002", + "issue:42371|issue:43874", + "issue:42673|issue:45072", + "issue:45357|issue:45478", + "issue:45083|issue:45446", + "issue:43116|issue:43825", + "issue:42222|issue:43441", + "issue:44075|issue:44936", + "issue:43992|issue:44704", + "issue:44995|issue:46697", + "issue:28348|issue:46640", + "issue:43066|issue:46710", + "issue:43827|issue:45446" + ], + "prepared_review_unit_hash": "c5d9d1d27830db28ecfb03f1f6cc7a559abbad2a2416c44fa46cb0892c6701dd", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 34822, + "estimated_input_tokens": 8706, + "estimated_eval_tokens": 17668 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated closed issues with superficial title similarity around Transformers v5 regressions; all soft edges should be rejected. No PRs are present.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across tokenization, model loading, training, vision, packaging, and docs; the soft matches are mostly broad v5-era similarity rather than the same underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best umbrella representative because it is broad, upgrade-related, and closest to the cluster\u2019s common theme of v5 breakage/import errors, even though it is not a true duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Both mention training/evaluation behavior, but one is a dtype mismatch between train and eval while the other is incorrect output under pipeline parallelism; different code paths and symptoms." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Importing transformers.onnx failing is unrelated to LayoutLMv3 error-message clarity for bounding boxes; different subsystem and bug type." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Missing TensorFlow in a torch extra install is a packaging/backend issue, while Python 3.14 load failure is an interpreter-compatibility issue; not the same defect." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 HF/JAX numerical discrepancy and Qwen3_5 infinite-loop-on-image are different models and different failure modes." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to a performance issue in SequenceFeatureExtractor.pad." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig invalid for GLM-5 and RoPE implementation discussion are distinct model issues; not the same concrete bug." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "A missing layer_type_validation import is a specific symbol breakage, while 45278 is a broad multi-import regression after upgrade; similar theme but not the same underlying issue." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer cleanup behavior change and stale phonemizer backend after language switching are unrelated text-processing vs phonemizer state bugs." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "base_model_prefix loading breakage and NaNs after checkout/init affect different code paths and have different root causes." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Multi-label example script returning empty results is unrelated to a SigLIP2 implementation discrepancy." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors and a documentation parameter-name typo are not the same bug." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but one is decoder-type selection and the other is incorrect DeepSeek tokenization results; too different to merge." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility and a wrong PyTorch version check for AuxRequest are separate environment-check problems." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL-Moe and Qwen3.5 Moe from_pretrained errors are related by model family, but the concrete failure paths are different." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 config is unrelated to NaNs in classification heads." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 behavior settings and a missing image-patch-count method in GLM46VImageProcessorFast are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "A VRAM leak in multi-threaded dataloader use and dtype mismatches in bfloat16 inference are different bugs." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Incorrect visual encoder keys on save_pretrained and from_pretrained failure are both Qwen3.5-related, but they are opposite sides of different serialization/loading bugs." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Feature-extractor length calculation in qwen3_omni_moe is unrelated to a PyTorch version gate for AuxRequest import." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example-script multi-label classification failure and pipeline() translation messaging after v5 removals are different issues." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Vitpose model breakage and Ministral-3 flash-attention failure are unrelated model-specific regressions." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Ignored SGD optimizer args and trainer.evaluate() failing after train are distinct training-loop bugs." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight during loading and AutoProcessor not forwarding kwargs to cached_file are different loading/processor paths." + }, + { + "left": "issue:44995", + "right": "issue:46697", + "accept": false, + "reason": "Stale indexer cache crash on second forward pass and Swin2SR over-padding are unrelated model/runtime vs image-processing bugs." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "FlashAttention 2 support request for GPT2LMHeadModel and a vague flash-attention issue are too underspecified and not the same concrete bug." + }, + { + "left": "issue:43066", + "right": "issue:46710", + "accept": false, + "reason": "Wrong tokenizer decoder type and DeepSeek-R1-Distill-Llama-8B accuracy regression are both v5 regressions, but they involve different models and failure modes." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Stale docs referencing pipeline() and a PyTorch version check bug are unrelated documentation vs runtime-import issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 49, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:33453", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:42673", + "issue:42796", + "issue:43064", + "issue:43122", + "issue:43408", + "issue:43450", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44261", + "issue:44291", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44779", + "issue:44829", + "issue:44877", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45310", + "issue:45341", + "issue:45356", + "issue:45362", + "issue:45405", + "issue:45446", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46178", + "issue:46231", + "issue:46298", + "issue:46328", + "issue:46512", + "issue:46614", + "issue:46655" + ], + "soft_pairs": [ + "issue:43064|issue:43673", + "issue:43937|issue:45563", + "issue:33453|issue:39401", + "issue:42673|issue:42796", + "issue:44380|issue:46077", + "issue:45356|issue:46614", + "issue:43673|issue:46328", + "issue:42673|issue:46655", + "issue:44291|issue:45071", + "issue:45071|issue:46614", + "issue:45405|issue:45446", + "issue:44466|issue:46002", + "issue:44117|issue:45588", + "issue:44410|issue:45310", + "issue:39401|issue:44779", + "issue:43937|issue:44977", + "issue:45071|issue:46512", + "issue:45341|issue:45561", + "issue:45341|issue:46328", + "issue:44829|issue:45216", + "issue:43450|issue:43986", + "issue:44117|issue:45081", + "issue:38617|issue:41628", + "issue:42617|issue:45341", + "issue:43122|issue:46298", + "issue:45362|issue:46328", + "issue:43408|issue:43742", + "issue:42371|issue:44261", + "issue:42222|issue:46178", + "issue:44877|issue:46231" + ], + "prepared_review_unit_hash": "8dfa9f4c947dd0ff97e5471775fa53d4981daa0d7f641a4df88f216fd1ad2eef", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36721, + "estimated_input_tokens": 9181, + "estimated_eval_tokens": 18618 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of tokenizer regressions, model-loading failures, attention/backend bugs, and test-utils issues; the soft links mostly reflect broad topical similarity rather than true duplicates.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43122 is the broadest umbrella for the tokenizer-regression subset: it explicitly compares v4.57.3 vs v5 behavior and can cover several related tokenizer reports better than the more model-specific issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44779 is the most concrete standalone bug report here: it names a specific model, a version regression, and a user-visible incorrect tokenization result, making it the strongest actionable issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Different failures: optimizer state placement in FSDP/PEFT vs generation cache missing." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Unrelated generation bugs: invalid GenerationConfig vs a stale warning for num_return_sequences." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both are tokenizer-related, but one is loading regression and the other is offset_mapping correctness for Qwen3." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in multi-threaded dataloading is unrelated to tokenizer BPE marker handling." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both involve attention behavior, but they are different models/backends and different bugs." + }, + { + "left": "issue:45356", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer codec/fix_mistral_regex regression is unrelated to stale phonemizer backend state." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache missing and SequenceFeatureExtractor.pad inefficiency are unrelated components." + }, + { + "left": "issue:42673", + "right": "issue:46655", + "accept": false, + "reason": "VRAM leak in dataloader threads is unrelated to video processor padding calculation." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Both are v5-loading/type issues, but one is an unexpected init arg and the other is PretrainedConfig type checking." + }, + { + "left": "issue:45071", + "right": "issue:46614", + "accept": false, + "reason": "PretrainedConfig type-checking regression is unrelated to phonemizer backend persistence." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version bump problem is unrelated to a PyTorch version gate in flex_attention." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization of tied weights is unrelated to NaNs in classification heads." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES fallback bug is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different Qwen3-related issues: missing projections in one model versus a from_pretrained error in another." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different symptoms." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation and flash-attention generation failure are different code paths." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "PretrainedConfig type checking and stale config.dtype after from_config are separate bugs." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Generic testing_utils bug and xdist race on captured_info.txt are not clearly the same defect." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing utils and SequenceFeatureExtractor.pad inefficiency are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention training degeneration is unrelated to Qwen3.5 checkpoint serialization." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Video processor shape bug and AutoProcessor crash without torchvision are different problems." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Mapping-name None handling and Mistral regex patch crashing are unrelated tokenizer-loading failures." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different missing exports and different APIs." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "3d_parallel.py runtime failure is unrelated to a testing_utils bug." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer regression is unrelated to a noisy logging message." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Qwen3.5 chat crash is unrelated to SequenceFeatureExtractor.pad performance." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "SAM3 model-type warning is unrelated to a MobileLLM loading key error." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API guidance is unrelated to MLA rms_norm_eps precision behavior." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "vitpose breakage and tensorflow_text backend mapping are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading for granite_speech is unrelated to a documentation parameter-name mismatch." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 50, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43577", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43937", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44360", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:45003", + "issue:45071", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45656", + "issue:46172", + "issue:46209", + "issue:46231", + "issue:46617" + ], + "soft_pairs": [ + "issue:44484|issue:45310", + "issue:42371|issue:46617", + "issue:43937|issue:44496", + "issue:39401|issue:46209", + "issue:44410|issue:45478", + "issue:44933|issue:45278", + "issue:38617|issue:44220", + "issue:44075|issue:45656", + "issue:44263|issue:44485", + "issue:43066|issue:43122", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:42222|issue:45542", + "issue:41628|issue:45071", + "issue:43874|issue:44261", + "issue:45071|issue:45542", + "issue:43931|issue:44410", + "issue:42175|issue:42222", + "issue:44483|issue:45464", + "issue:44484|issue:45478", + "issue:44360|issue:46172", + "issue:44263|issue:46172", + "issue:43012|issue:43577", + "issue:27957|issue:43352", + "issue:43931|issue:44038", + "issue:43441|issue:44829", + "issue:43097|issue:43452", + "issue:45003|issue:45341", + "issue:43986|issue:45341", + "issue:43874|issue:44263", + "issue:44933|issue:46231", + "issue:44261|issue:44485" + ], + "prepared_review_unit_hash": "8d1a542a68d7018b1c4bbc65484b490e501b17848a470e9eaf4d27b37c11bd39", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36184, + "estimated_input_tokens": 9046, + "estimated_eval_tokens": 18348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose similarity cluster, not a true duplicate set: it mixes unrelated model bugs, tokenizer regressions, backend/install issues, and API failures. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.79, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the broadest representative of the cluster\u2019s recurring upgrade/import-breakage theme, even though the cluster itself is heterogeneous.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Different problems: save_pretrained shard sizing vs Qwen3.5 MoE loading failure." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 settings warning and trust_remote_code cache FileNotFoundError are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GLM generation config validation vs Olmo model type/config loading error." + }, + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset mapping bug and FP8 inference on B200 are different code paths." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen-family models, but the reported failures are distinct loading/projection bugs." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A missing/nonexistent import and broad upgrade import errors are not the same concrete bug." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "Configuration import error and fbank feature extraction issue are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Optimizer args ignored vs double optimizer.step under deepspeed are different defects." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer split-return issue and GLM-5 RoPE implementation bug are different." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer-related v5 regressions, but not the same underlying failure." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings API and unexpected position-id keys are unrelated regressions." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache/chunked_prefill issue and a testing_utils bug are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Vitpose model breakage and TF/tensorboard backend packaging error do not share a code path." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "AutoImageProcessor import failure and PreTrainedConfig type-checking breakage are different issues." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Missing image patch helper and missing rms_norm_eps in MLA are unrelated model bugs." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Config type checking and TF backend dependency handling are different regressions." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and qwen3next projection omission are distinct model-loading issues." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging and vitpose breakage are not the same bug." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both concern chat/completions, but one is API acceptance and the other is streaming inference failure." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "save_pretrained shard sizing and Qwen3.5 MoE from_pretrained error are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "Missing ReLU in the DSA indexer and deepseek-v4 comb.transpose bug are different implementation errors." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "torch.split return handling and deepseek-v4 tensor transpose are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "bfloat16 precision warning and BLIP2 dtype propagation bug are different symptoms and code paths." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and NemotronHForCausalLM Flash Attention 2 support are model-specific, separate requests." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Qwen3-VL shape mismatch and Qwen3-VL-Moe loading bug are not the same failure." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "FlashAttention load failure in Ministral-3 and flash_attention_3 training degeneration are different concrete problems." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Tie-embeddings API removal and gguf_file/from_pretrained behavior are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "sys.modules access bug and testing_utils bug are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/video-model torchvision dependency crash and testing_utils bug are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "These are separate GLM-family bugs affecting different helpers." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Nonexistent image_utils import and a doc mismatch in set_training() are not duplicates." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Missing rms_norm_eps precision issue and GLM-5 RoPE implementation issue are distinct." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 51, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42890", + "issue:42907", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43502", + "issue:43531", + "issue:43550", + "issue:43673", + "issue:43825", + "issue:43873", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44315", + "issue:44351", + "issue:44380", + "issue:44410", + "issue:44484", + "issue:44655", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:44987", + "issue:45103", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45588", + "issue:45712", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46369", + "issue:46421", + "issue:46655", + "issue:46664", + "issue:46701", + "issue:46710" + ], + "soft_pairs": [ + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43531|issue:45341", + "issue:43673|issue:46123", + "issue:45440|issue:46172", + "issue:43097|issue:45588", + "issue:38617|issue:42757", + "issue:43986|issue:45372", + "issue:44410|issue:45375", + "issue:43064|issue:44315", + "issue:43097|issue:46050", + "issue:44038|issue:44410", + "issue:46655|issue:46664", + "issue:43986|issue:44484", + "issue:38617|issue:44351", + "issue:42890|issue:46298", + "issue:46172|issue:46710", + "issue:43097|issue:43873", + "issue:44062|issue:44117", + "issue:44117|issue:46701", + "issue:44936|issue:46365", + "issue:43502|issue:43992", + "issue:44704|issue:45103", + "issue:44380|issue:46421", + "issue:44938|issue:46178", + "issue:43986|issue:44117", + "issue:43825|issue:44655", + "issue:42907|issue:44655", + "issue:42907|issue:45753", + "issue:43208|issue:43550", + "issue:45712|issue:46298" + ], + "prepared_review_unit_hash": "c8b44b7b6b692f29c06cb852952be38d76c784a8378dda4266b3590a90c64b74", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36763, + "estimated_input_tokens": 9191, + "estimated_eval_tokens": 18638 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated Transformers bug reports with topical overlap only; none of the soft pairs look like true duplicates or mergeable single-change PRs.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43097 is the best umbrella representative: it is a concrete, high-signal v5 regression in a core code path and broader than the more model-specific or incidental reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the strongest standalone issue to keep: clear impact, core compatibility regression, and enough context to guide a fix without being overly narrow.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different bugs: one is a Qwen3_5 image-loop issue, the other is a Trainer import failure tied to datasets." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Both involve loading/tokenizer paths, but the failures are distinct: mapping lookup returning None vs a model load break in v5.1.0." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Unrelated: a Qwen3-MoE sliding_window bug versus a testing_utils bug." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Both touch generation/pipeline behavior, but one is chunked_prefill cache handling and the other is final-batch flag logic in MaskGenerationPipeline." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Different DeepSeek-related code paths and symptoms; not the same underlying change or bug." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Core issues differ: a removed config attribute in v5 versus a flash-attention crash on sink-less models." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are ImportErrors, but for different symbols from different modules; no shared root cause shown." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Different dependency/loading failures: missing torchvision for video models versus missing ReasoningEffort from mistral_common." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen-family config/model issues, but they concern different models and different missing fields/projections." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Trainer/FSDP optimizer-state corruption is unrelated to Liger Kernel application in model_init." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "A v5 config/API removal issue is not the same as quantized save/load breakage." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Related family name, but the concrete failures are different Qwen variants and different missing components." + }, + { + "left": "issue:46655", + "right": "issue:46664", + "accept": false, + "reason": "Both are SmolVLM video-processor bugs, but one is padding math and the other is num_frames semantics." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Different areas: missing torchvision during processor loading versus save_pretrained shard-size behavior." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-name errors, but they involve different exported symbols and likely different fixes." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "A test seeding problem is unrelated to a noisy logging message." + }, + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "Different DeepSeek issues: a specific tensor operation bug versus an accuracy regression after upgrade." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "One is about an API/config removal; the other is an offloading/quantization malfunction." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer constructor argument handling is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:44117", + "right": "issue:46701", + "accept": false, + "reason": "Tokenizer mapping fallback and an image-text-to-text pipeline failure are separate code paths." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Trainer.evaluate failing after train is not the same as failing to import Trainer because datasets is installed." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network leakage and a missing embed_tokens.weight during UMT5 loading are different defects." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and an auto_docstring annotation crash are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to cache lazy initialization and recompilation." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 import/load failure is not the same as an unsupported tensorflow_text backend mapping." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Different failure modes and layers: video dependency loading versus tokenizer mapping resolution." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task-support messaging and saving Pipeline objects are separate problems." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Dequantized model save failure is different from inability to save Pipeline objects." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Quantized save/load behavior and a Qwen3_5 image infinite loop are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure do not share a concrete code path." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Dummy-class cleanup/check_repo issues are unrelated to a noisy logging message." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 52, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43867", + "issue:43906", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44077", + "issue:44162", + "issue:44222", + "issue:44230", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44625", + "issue:44655", + "issue:44683", + "issue:45478", + "issue:45479", + "issue:45584", + "issue:45600", + "issue:46132", + "issue:46172", + "issue:46298", + "issue:46365" + ], + "soft_pairs": [ + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:43992|issue:44484", + "issue:43066|issue:43906", + "issue:43352|issue:43976", + "issue:43097|issue:46298", + "issue:43352|issue:44380", + "issue:44162|issue:46172", + "issue:44380|issue:46132", + "issue:44077|issue:46365", + "issue:45584|issue:46365", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:44380|issue:45479", + "issue:42175|issue:46365", + "issue:42490|issue:43012", + "issue:43064|issue:44380", + "issue:42757|issue:43824", + "issue:43012|issue:44683", + "issue:43208|issue:43526", + "issue:45600|issue:46298", + "issue:44483|issue:44514", + "issue:43618|issue:44380", + "issue:43618|issue:43761", + "issue:44038|issue:44230", + "issue:38617|issue:43502", + "issue:43756|issue:44485", + "issue:43299|issue:43867", + "issue:42994|issue:44655", + "issue:44625|issue:46365" + ], + "prepared_review_unit_hash": "f0fe2aca0d8d784aeb5a96374e4bd4699a278d508d6c7179cf3a636666459c5f", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35774, + "estimated_input_tokens": 8944, + "estimated_eval_tokens": 18144 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a mixed bag of unrelated Transformers issues; the only strong duplicate-like pair is the CLIP v5 output regression pair. No PRs are present.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43761 is the cleanest representative of the CLIP v5 regression cluster: it states the concrete forward-pass symptom, includes the explicit regression tag, and is broader than the single attentions-only symptom in issue:43618.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43761 is the best overall anchor among the cluster candidates because it has the clearest concrete reproduction of the shared CLIP output regression and is the most mergeable as a single bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Different failures: a bfloat16 precision warning vs a torch.compile/SDPA model error." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Unrelated symptoms: precision warning vs a tie_word_embeddings warning during LoRA fine-tuning." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Both involve saving, but one is a missing weight during load/encode and the other is max shard size behavior in save_pretrained()." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Both are loading/import related, but they point to different symbols and different underlying code paths." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Different problems: Flash Attention support for a model vs a Python version compatibility break." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "API removal/regression vs a generic noisy logging message; not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Different attention-related issues: model support gap vs GPT2 attention scaling being ignored." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "Different model families and failure modes; no shared code-path is evident." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Config handling in attention backends vs AttentionInterface.register changing callback behavior." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "PatchTSMixer post_init validation is unrelated to Trainer import conflicts with datasets." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper generation after align_special_tokens is unrelated to Trainer import resolution." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both mention MoE/save/load, but one is FP8 save_pretrained and the other is a from_pretrained load regression." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma config attribute error and Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention backend config bug vs degenerate zero-loss classification behavior." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Package-install backend selection issue vs importing Trainer being blocked by datasets." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-path trailing slash warning vs bfloat16/float32 precision warning." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "Optimizer state mismatch under FSDP2/PEFT is not the same as attention scaling being ignored." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are ImportErrors, but for different missing symbols and different dependency surfaces." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "A runtime precision warning is not the same as a torch >= 2.9 compiled flex_attention failure." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training blockers and BeitImageProcessorFast label reduction are unrelated." + }, + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "auto_mappings CI breakage is unrelated to a generic logging complaint." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat completions API rejection and batched processor padding crash are different request paths." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both involve model internals, but CLIP output fields and GPT2 attention scaling are distinct bugs." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": true, + "reason": "Same underlying CLIP v5 regression: optional outputs are no longer propagated in forward results, affecting both attentions and hidden_states." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "A Qwen3-VL-Moe regression and an fp8 inference support request are not the same issue." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Different import/dependency failures: a missing internal symbol vs unexpected API requests under local_files_only." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "RoPE layer-dropping in Smollm3 and GLM-5 RoPE implementation discussion are not the same concrete bug." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL-Moe loading regression and a generic sorted-state_dict load error are not clearly the same failure." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Saving quantized models and saving Pipeline objects fail in different save_pretrained paths." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Trainer import failures are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the packet supports issue:43761 as the best canonical CLIP regression anchor, and the single accepted soft pair (issue:43618 vs issue:43761) is a conservative duplicate-like call. The remaining soft-edge rejections are appropriately cautious and do not overclaim shared bugs. No PR pairs are present." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 53, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36010", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42757", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43575", + "issue:43611", + "issue:43618", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44075", + "issue:44077", + "issue:44246", + "issue:44355", + "issue:44483", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:45071", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45478", + "issue:45561", + "issue:45850", + "issue:46077", + "issue:46082", + "issue:46202", + "issue:46285", + "issue:46365", + "issue:46369" + ], + "soft_pairs": [ + "issue:45276|issue:46077", + "issue:36683|issue:43475", + "issue:36010|issue:42757", + "issue:43352|issue:43937", + "issue:44483|issue:46369", + "issue:43611|issue:44655", + "issue:43901|issue:44655", + "issue:36683|issue:44077", + "issue:43867|issue:45216", + "issue:43976|issue:44938", + "issue:44625|issue:46077", + "issue:46077|issue:46082", + "issue:43618|issue:45216", + "issue:36683|issue:43525", + "issue:36683|issue:44625", + "issue:43992|issue:45310", + "issue:43010|issue:45278", + "issue:43064|issue:44492", + "issue:44075|issue:46365", + "issue:44355|issue:44855", + "issue:44938|issue:45071", + "issue:43867|issue:45375", + "issue:43867|issue:46202", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:41669|issue:44246", + "issue:43010|issue:45850", + "issue:36296|issue:43575", + "issue:43867|issue:45357", + "issue:43010|issue:43976", + "issue:43618|issue:44077", + "issue:43992|issue:45478", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:43618|issue:44857" + ], + "prepared_review_unit_hash": "71dfd700f43ef09f444a2b97ca67b85df7c73f204f8db7ecd15ca2aeb42c7c45", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37786, + "estimated_input_tokens": 9447, + "estimated_eval_tokens": 19150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: it mixes unrelated import/load/save/config regressions across different models and subsystems. None of the soft pairs look like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set, covering import regressions after upgrading; if one issue must represent this mixed cluster, it is the closest fit, though the cluster is not a clean duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best single representative because it is the most general import-regression report and overlaps with several other import-time complaints more than the model-specific issues do.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Same model family, but different bugs: token embedding/output embedding resize vs causal attention mask construction." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different models and different missing attributes; not the same failure path." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different packages and likely different fixes." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support rejection and invalid GenerationConfig are distinct validation/code-path issues." + }, + { + "left": "issue:44483", + "right": "issue:46369", + "accept": false, + "reason": "One is a chat/completions API rejection; the other is an import conflict around Trainer and datasets." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Loading models with base_model_prefix and saving Pipeline objects are separate code paths." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mentioning return_all_scores do not describe the same bug as Pipeline save_pretrained failure." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Missing Gemma3 config attribute and ignored SGD optimizer args are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Sorted state_dict load failure and incorrect save_pretrained output are different regressions." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are version/platform compatibility reports, but the actual breakage and environment constraints differ." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Config propagation for num_labels is unrelated to attention mask generation." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Bidirectional attention mask logic and head_dim validation are separate issues." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions being unset is unrelated to Qwen3.5 checkpoint saving." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different config attributes missing on different model families; not a shared root cause." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Both are config-related, but one is a missing Gemma3 attribute and the other is num_labels propagation." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Missing embed_tokens.weight and Qwen3.5 from_pretrained errors are not clearly the same bug." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache update no_grad behavior and broad import failures after upgrade are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Distributed optimizer-state corruption and a typo in cache strategies do not share the same code path." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Ignored SGD args and Trainer import failure are different subsystems and fixes." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled Python import errors and a torch.jit parser/IndentationError are not the same underlying defect." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python version incompatibility and PretrainedConfig type-checking regression are distinct issues." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Load failure from state_dict ordering and silent dropping of a config field are different problems." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "State_dict load errors and Gemma pipeline roundtrip failure are not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3 config attribute loss and FSDP2/PEFT optimizer-state corruption are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Test flakiness from missing set_seed and xdist race on captured_info.txt have different causes." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Import-* slowdown and intermittent import slowness are performance complaints, but not the same concrete regression." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "Cache update decoration and a broad 'breaks too many models' complaint are too different and too vague to merge." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and tp-related OOM are different failure modes." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Sorted state_dict loading and incorrect visual encoder keys on save are separate checkpoint bugs." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Both concern CLIP model structure, but one is attentions assignment and the other is .text_model structure inconsistency." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIPOutput attentions and an AMP/CUDA loss crash are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch missing import and tokenizer loading regression are different failure classes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 54, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:27957", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:42548", + "issue:42757", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43576", + "issue:43723", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:44062", + "issue:44112", + "issue:44222", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44655", + "issue:44734", + "issue:44749", + "issue:44961", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45375", + "issue:45440", + "issue:45468", + "issue:45478", + "issue:45600", + "issue:46132", + "issue:46617", + "issue:46664", + "issue:46726" + ], + "soft_pairs": [ + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:42898|issue:46664", + "issue:43502|issue:44062", + "issue:44734|issue:46617", + "issue:43723|issue:44556", + "issue:44222|issue:45125", + "issue:44387|issue:44556", + "issue:43874|issue:46617", + "issue:43824|issue:46617", + "issue:43010|issue:43065", + "issue:43502|issue:44485", + "issue:43576|issue:46132", + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:44355|issue:44749", + "issue:42548|issue:44351", + "issue:45125|issue:45375", + "issue:46664|issue:46726", + "issue:43010|issue:43232", + "issue:30990|issue:31795", + "issue:43502|issue:44961", + "issue:43502|issue:44222", + "issue:43827|issue:44961", + "issue:30333|issue:36296", + "issue:43299|issue:44556", + "issue:44355|issue:45071", + "issue:33453|issue:34689", + "issue:44355|issue:44655", + "issue:27957|issue:33453", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:44315|issue:44360", + "issue:31795|issue:38617", + "issue:43576|issue:43824", + "issue:44961|issue:45468", + "issue:44351|issue:46617", + "issue:44485|issue:44961", + "issue:38617|issue:43576" + ], + "prepared_review_unit_hash": "6ca9abc0b6ebab76143d67070eed7600aa8976a194f602bfdf53feb1683085f0", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39507, + "estimated_input_tokens": 9877, + "estimated_eval_tokens": 20010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is highly heterogeneous: most items are unrelated model-loading, import, CI, docs, or performance bugs. None of the soft pairs look like safe duplicates under a strict reading.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43576 is the broadest, most representative v5 regression in the set and a reasonable cluster anchor; it is user-facing and version-wide rather than tied to one model or test case.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43576 best fits as the cluster representative because it captures the generic v5 breakage theme seen across several nearby issues without being overly model-specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "Different bugs: a DSA indexer/ReLU discussion versus a DeepSeekV3 implementation divergence." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated symptoms and subsystems: SamHQ test seeding failures versus Phi-4 multimodal model loading." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are test/CI issues, but they target different models and failure causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "SamHQ integration flakiness is unrelated to removed Sam3LiteText configs in auto_mappings." + }, + { + "left": "issue:42898", + "right": "issue:46664", + "accept": false, + "reason": "One is a tokenizer cleanup behavior change; the other is a video processor num_frames semantic issue." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Different code paths: local_files_only network access versus AddedToken keyword duplication." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "Different loading/runtime failures: KV-cache continuation indexing in serve versus cache loading FileNotFoundError." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both involve loading, but one is AutoTokenizer loading in v5 and the other is checkpoint reload compatibility after upgrade." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained behavior and missing _tp_plan for tensor parallelism are separate issues." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Different regressions: int4 memory growth/OOM versus checkpoint reload breakage after version upgrade." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Missing get_number_of_image_patches is unrelated to trust_remote_code cache loading." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "ImportError for a specific Qwen VL class is not the same bug as cache FileNotFoundError." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "No shared underlying change: no_grad decoration for cache updates versus a dummy Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only network requests and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "A broken CLI env command is not the same as AttentionInterface.register changing function behavior." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Closely themed, but the titles point to different problems: missing _tp_plan versus a broader from_pretrained error." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting on failure and tokenizer loading regression are different issues." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors and a Chinese-language performance regression report are unrelated." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and likely different fixes." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Missing _tp_plan for tensor parallelism is unrelated to a missing vision-config field." + }, + { + "left": "issue:46664", + "right": "issue:46726", + "accept": false, + "reason": "Video num_frames semantics and missing vision-backend handling are distinct processor issues." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache no_grad decoration and generation kwargs updates after sync_gpus are different code paths." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence-transformers loading hang and model.forward documentation confusion are not the same bug." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "local_files_only networking and an unrelated racoon placeholder issue are not duplicates." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "No overlap beyond being in the same library; one is a download flag bug, the other is FP8 save_pretrained." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs referencing pipeline() after removals is unrelated to the racoon issue." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLFlow status reporting and tensor parallel training are different problems." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are version-related loading regressions, but they affect different model families and failure modes." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file errors do not match a PretrainedConfig type-checking regression." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both are loading regressions, but they concern different models and likely different root causes." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file failures and Pipeline save_pretrained problems are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to tokenizer-loading regression." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Hugging Face Hub offline-mode import failure and the transformers env command are different issues." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI env command breakage and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application via model_init and a DSA indexer/ReLU discussion are different code paths." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation wording and a missing internal import are not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken env command and Qwen2_5_VL import failure do not share an underlying fix." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "The placeholder racoon issue is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "HybridCache import failure and trust_remote_code cache loading failure are different problems." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and the racoon placeholder issue are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "Missing layer_type_validation import and the broken env command are different regressions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 55, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:36296", + "issue:41628", + "issue:41669", + "issue:42371", + "issue:42548", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43475", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43901", + "issue:43976", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44246", + "issue:44273", + "issue:44315", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44908", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45362", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46218", + "issue:46617", + "issue:46664" + ], + "soft_pairs": [ + "issue:44908|issue:44961", + "issue:36296|issue:43901", + "issue:46143|issue:46144", + "issue:43475|issue:46617", + "issue:36296|issue:43526", + "issue:44222|issue:44230", + "issue:41628|issue:42548", + "issue:44315|issue:46209", + "issue:44485|issue:46664", + "issue:45362|issue:45850", + "issue:43673|issue:44492", + "issue:44273|issue:45542", + "issue:44360|issue:46664", + "issue:44623|issue:45230", + "issue:44038|issue:45125", + "issue:44273|issue:45003", + "issue:44393|issue:45071", + "issue:44393|issue:45753", + "issue:41669|issue:43673", + "issue:43576|issue:44556", + "issue:43525|issue:45850", + "issue:44556|issue:45092", + "issue:43572|issue:45850", + "issue:45278|issue:45850", + "issue:44492|issue:45850", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:43408|issue:44315", + "issue:44230|issue:46095", + "issue:42890|issue:44908", + "issue:44246|issue:44273", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:43976|issue:44393", + "issue:44704|issue:44908", + "issue:44908|issue:45071" + ], + "prepared_review_unit_hash": "5a6dc41540fb8a230e42a119abd36942ce4218229f5319de6febb1e1a8c4e48a", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37671, + "estimated_input_tokens": 9418, + "estimated_eval_tokens": 19092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of Transformers v5 import/lazy-loading regressions and unrelated model-specific bugs; none of the soft pairs look like clear duplicates.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45850 is the broadest umbrella complaint in the set about v5 breakage, so it best anchors this mixed cluster despite being vague.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the most central representative of the repeated 'v5 breaks things' theme and is the most suitable single issue to stand in for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Completely unrelated: scheduler kwargs bug vs a placeholder issue with no matching code path." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug and a pipeline docs mismatch are unrelated." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "Both are RoFormer-related, but one is kwargs propagation and the other is attention-interface usage; different bugs." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "SAM3 vision encoder attribute error vs trust-remote-code cache FileNotFoundError; no shared failure path." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor parallel training and BEiT image-processor label reduction are separate subsystems." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both mention fp8/moe, but one is save_pretrained serialization and the other is inference support." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different export issues; too broad to merge." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Model_init/liger-kernel application and B200 fp8 inference failure are unrelated problems." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "RoPE implementation discussion vs video num_frames semantics; different model components." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A specific Qwen3.5 chat crash is not the same concrete bug as a broad 'v5 breaks too many models' complaint." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Cache-missing during chunked prefill and a typo in cache strategies are different issues." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading regression and TensorFlow backend detection/installation behavior are related only at a high level." + }, + { + "left": "issue:44360", + "right": "issue:46664", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to the SmolVLM video processor bug." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "One is a concrete missing-files bug in processor.save_pretrained; the other is an empty generic report." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Same model family, but generic Qwen3-VL-Moe issues and a missing tensor-parallel plan are different code paths." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both touch loading mechanics, but lazy loading and unsafe sys.modules access are distinct bugs." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issues are unrelated to PretrainedConfig type-checking regressions." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but bbox hallucinations and an infinite loop on a specific image are not the same bug." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star performance regression vs generation cache missing during chunked prefill; different failure modes." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "v5 env-command breakage and old-checkpoint reload incompatibility are separate regressions." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a specific config bug, not the same as a broad v5 complaint." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are upgrade/reload compatibility issues, but they target different checkpoint types and initialization paths." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is a specific missing field, not the same issue as the broad v5 breakage report." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Import errors after upgrade are related in theme, but this is not the same concrete bug as the generic v5 breakage complaint." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A cache-strategy typo is not the same as a general 'v5 breaks models' report." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API guidance and cache-strategy typo are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import-star slowdown and a specific module import failure do not look like the same root cause." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "SAM3 model-type warning and Liger Kernel model_init behavior are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support for Qwen models and a DeepSeek weight conversion script are different requests." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test flakiness from missing set_seed and an inverse_sqrt scheduler kwargs bug are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import slowness and lazy-loading failure may be adjacent, but the evidence is too weak to call them duplicates." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Nemotron Flash Attention support and Qwen3-VL bounding-box errors are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command and inverse_sqrt scheduler kwargs are different features with no shared code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 56, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28348", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43704", + "issue:43827", + "issue:43976", + "issue:44077", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44655", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45230", + "issue:45250", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46172", + "issue:46218", + "issue:46328", + "issue:46655" + ], + "soft_pairs": [ + "issue:44295|issue:45230", + "issue:44297|issue:45230", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:43976|issue:46218", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:43576|issue:44273", + "issue:44315|issue:44829", + "issue:44360|issue:45468", + "issue:43519|issue:46655", + "issue:44273|issue:44908", + "issue:44655|issue:44861", + "issue:43576|issue:46218", + "issue:45230|issue:45310", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:31515|issue:31795", + "issue:46018|issue:46172", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:45850|issue:46218", + "issue:43519|issue:43704", + "issue:43606|issue:45468", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:43298|issue:46095", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:28348|issue:45250", + "issue:44908|issue:44998", + "issue:44492|issue:44998" + ], + "prepared_review_unit_hash": "e7f15fb5f187def6f11fec4b59f96b7f7725ef12c589ed42475fb1e2bee0caaf", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39145, + "estimated_input_tokens": 9787, + "estimated_eval_tokens": 19830 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose set of unrelated Transformers bug reports. None of the soft pairs appear to describe the same concrete bug/change, so I would reject all candidate duplicate edges. The strongest standalone issue is the specific Qwen3.5 MoE `from_pretrained` regression.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 45310 is the clearest concrete regression report in the set: a specific `from_pretrained` failure for Qwen3.5 MoE on Transformers >=5.4.0, with a narrowly defined code path and likely reproducible behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45310 is the most triage-ready issue here because it is specific, actionable, and version-scoped; the other candidates are either too generic, documentation-only, or too loosely related.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "`44295` is a specific `position_ids` buffer bug; `45230` is a placeholder-level generic report, with no evidence they share the same root cause." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save metadata mismatch is a different failure mode from the generic `45230` report." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "One is about removing `import *` for startup/perf; the other is an import error for `HybridCache`. Related area, but not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "A cache-strategy typo and a `SequenceFeatureExtractor.pad` performance issue are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Model import cleanup versus a testing utility bug; no shared underlying defect." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Both mention broad v5 breakage, but they are different reports with no concrete shared code path." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different models and different failures: Qwen3-VL bounding-box output versus Qwen3.5 MoE loading." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python version compatibility and a missing Qwen3_5 module are different problems." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "`post_init` validation and `_get_tied_weight_keys` crashing on a list are separate code paths." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing `pipeline()` and `inverse_sqrt` scheduler kwargs being ignored are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion around forward args versus a tokenizer loading regression are not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "A broken `transformers env` command and lazy loading not working are different issues." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "`model_init` not applying Liger Kernel and degenerate training with flash-attn-3 are different failures." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "A DSA indexer math discussion and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Both are processor bugs, but one is Qwen3-VL timestamp calculation and the other is SmolVLM video padding; not the same root cause." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading versus scheduler kwargs are unrelated." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving pipeline objects and tied-weight-key handling are different subsystems and failures." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "CLI env command breakage is not the same as a missing `Qwen3_5ForCausalLM` import." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "`45230` is too generic to establish the same underlying bug as the specific Qwen3.5 MoE `from_pretrained` regression." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Different models and different implementation areas: timestamp handling versus RoPE implementation." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp math and scheduler kwargs are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import cleanup/startup cost and VRAM leakage in dataloader threads are not the same bug." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward-doc confusion and a model-loading regression for Llama 3.2 are distinct issues." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Slow `from_pretrained` checkpoint loading and confusing forward documentation are unrelated." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Shared-expert gating and a dtype/transpose bug in DeepSeek-V4 are different concrete problems." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leakage and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Different model families and different components; no shared bug." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "`local_files_only=True` network leakage and Qwen3-VL timestamp calculation are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 configuration behavior and VRAM leakage are different topics." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A broad complaint that v5 breaks many models is not the same as a specific missing module import." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Processor timestamp math and dataloader-thread VRAM leakage are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device-mismatch in CI and Gemma-4 audio positional encoding are different bugs." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "`torch.split()` return-value handling in a DSA indexer is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Layer-norm epsilon precision discussion and Gemma-4 audio positional encoding are not the same issue." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching audio-model support and DeepSeek-V4 weight conversion are different requests." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Missing `HybridCache` import and missing `Qwen3_5ForCausalLM` import are different symbols with no shared bug evidence." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention 2.0 support for NemotronH is unrelated to a missing Qwen3_5 module import." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both concern tokenizer changes in v5, but the reports are not specific enough to establish the same underlying bug." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "A GPT2-specific Flash Attention 2.0 support request is not the same as a generic Flash Attention 2.0 issue." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs handling and an unrelated placeholder title are not duplicates." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Cache-strategy typo and an unrelated placeholder title are not duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 23, + "review_unit_index": 57, + "review_unit_count": 57, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-479", + "nodes": [ + "issue:28348", + "issue:30333", + "issue:31515", + "issue:33453", + "issue:36296", + "issue:37934", + "issue:38617", + "issue:42936", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43525", + "issue:43572", + "issue:43618", + "issue:43668", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44623", + "issue:44861", + "issue:44938", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45850", + "issue:46018", + "issue:46144" + ], + "soft_pairs": [ + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:31515|issue:38617", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:28348|issue:43352", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:33453", + "issue:44230|issue:45310", + "issue:43298|issue:44230", + "issue:43717|issue:46144", + "issue:42936|issue:46018", + "issue:43668|issue:44031", + "issue:42936|issue:43240", + "issue:43240|issue:46144", + "issue:42936|issue:44031", + "issue:42936|issue:46144", + "issue:42936|issue:43813", + "issue:43240|issue:43668", + "issue:43668|issue:43813", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "7f3d1a44c3df2bd9e96cc6bb3717d2a7adbf339299cb6b984c5ebeb1b9182764", + "component_budget": { + "node_count": 479, + "item_count": 479, + "soft_pair_count": 1763, + "serialized_chars": 913995, + "estimated_input_tokens": 228499, + "estimated_eval_tokens": 457254 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 26, + "serialized_chars": 23352, + "estimated_input_tokens": 5838, + "estimated_eval_tokens": 11932 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab-bag of unrelated issues: model-specific feature requests, loading/version regressions, tokenizer warnings, and a few training/runtime bugs. None of the soft pairs look like the same underlying bug, so there is no real duplicate hub here.", + "confidence": 0.88, + "canonical_issue_reason": "No true canonical issue emerges because the set is heterogeneous. If forced to nominate a representative, issue 45850 is the broadest umbrella complaint, but it is too vague and not a credible duplicate anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the most general/high-level problem statement in the set, but it still does not match the rest closely enough to serve as a strong canonical duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "Different bugs: CLIP attentions regression vs an unrelated issue titled \"Unemployment\"." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "Unrelated issues; one is nonsensical/irrelevant, the other is a broad model-breakage complaint." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 import/load failure is not the same problem as the unrelated \"Unemployment\" issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Different code paths: tied-weight key handling vs chat template tokenization with tool calls." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading is unrelated to the layer_type_validation ImportError." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a different bug from _get_tied_weight_keys crashing." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLm pad_token_idx regression is unrelated to tied-weight-key handling." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but they are different model-specific support requests and not the same concrete fix." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support for Qwen models is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow job-status reporting failure is unrelated to slow checkpoint loading." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "from_pretrained performance regression and tensor-parallel training bug are different problems." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowness is not the same as a tokenizer-loading regression." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is fp8 support work and the other is a from_pretrained error in a different context." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Audio continuous batching support is unrelated to Qwen fp8 inference support." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention interface usage are separate model implementations." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to DeepSeek-V4 shared-expert gating." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A ModernBERT type-hint issue is not the same as the tokenizer regex warning regression." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value handling and kwargs propagation to loss calculation are different training bugs." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "kwargs not reaching the loss function is unrelated to RoFormer attention implementation details." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value logic is unrelated to tokenizer regex warnings." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former loss/ignore_value behavior is not the same as RoFormer attention interface usage." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to the quantization-operation typo." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "kwargs-to-loss propagation is a runtime bug, not the ModernBERT norm_eps type-hint issue." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERT type-hint mismatch and a typo in quantization attribute name are unrelated." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "A Llama4 norm implementation question is not the same as a ModernBERT type-hint bug." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value usage and ModernBERT norm_eps typing are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10519-3", + "nodes": [ + "issue:10519", + "issue:11005", + "issue:11307" + ], + "soft_pairs": [ + "issue:10519|issue:11307", + "issue:10519|issue:11005", + "issue:11005|issue:11307" + ], + "prepared_review_unit_hash": "003e4e9864bc2a213041457192b75ef445739dceafb0a721637f2c6df10ff8ba", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2878, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2878, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three items are unrelated feature requests in different areas, so this cluster should not be merged.", + "confidence": 0.99, + "canonical_issue_reason": "None of the issues are duplicates; if one had to be picked as a representative, 10519 is the clearest, self-contained product feature request, but it is not a true canonical duplicate target for the others.", + "canonical_pr_reason": null, + "best_issue_reason": "10519 is the most concrete and well-scoped issue, with a specific tokenizer behavior request; the other two are separate feature asks in unrelated domains.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Truncation-direction behavior for tokenizers is unrelated to ReduceLROnPlateau-like scheduler functionality." + }, + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Tokenizer truncation behavior and Wav2Vec2 word time offsets are different problems in different subsystems." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Learning-rate scheduling behavior is unrelated to speech model word-boundary timing offsets." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 23, + "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": "712cf169da8475617e702327fb998d8f38b3e4d28cdd2ff0d3b42a7626be73d5", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues do not look like true duplicates: one is about Tapas tokenization parity, and the other two are about model-loading/download failures with different causes and user-facing symptoms.", + "confidence": 0.28, + "canonical_issue_reason": "Issue #16618 is the broadest of the three and sits in the model-loading error area, so it is the best representative if one must be chosen for this noisy cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the most central issue for the cluster because it concerns a common model-load failure path, whereas #13244 is clearly about tokenization behavior and #15980 is specifically about private-model auth messaging.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different problem domains: Tapas tokenization mismatch vs. model loading failure. No indication they share the same underlying bug or fix." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both mention model download/load errors, but the concrete scenarios differ: private-model login/error messaging vs. a failure loading 'bert-base-uncased'. Not enough evidence of 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": 4, + "component_count": 23, + "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": "13e72cd1d3a00e2c96a8e8039d7ecc460aac681c1caa3156312c6f73f69b4901", + "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 two issues are both community contribution requests, but they target different features and code paths: Flash Attention 2 support vs GGUF support. They are not duplicates and should stay separate.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Reject: the issues request different capabilities (Flash Attention 2 vs GGUF) for different architectures and underlying implementations. They share only a generic contribution framing, not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-8", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:38435", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:38435", + "issue:28936|issue:44370", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46222", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:46222", + "issue:43834|issue:46226", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:46222|issue:46226" + ], + "prepared_review_unit_hash": "7d7c835ea1936c504d0b5a71695758e20b2cd20045d8714dd297a088c9206e10", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13529, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13529, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is noisy and mixes unrelated translation-request issues by language. The only clear duplicates are the placeholder-title translation issues (43834/44365/44370) and the exact Thai duplicate pair (46222/46226).", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43834 is the earliest of the largest near-identical placeholder-title subcluster, so it is the best representative issue for that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the cleanest representative of the dominant duplicate-like subcluster; the other language-specific translation requests are separate issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different languages/topics; only generic i18n wording overlaps." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language translation requests, not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish translation request vs generic placeholder-title issue; not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different language translation requests; similarity is only broad i18n template wording." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish vs Spanish translation requests are distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish translation request vs generic placeholder-title issue; no clear duplicate evidence." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language translation requests, not the same underlying issue." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different language translation requests, not the same underlying issue." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish vs Romanian translation requests are distinct." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian translation request vs generic placeholder-title issue; not the same issue." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Romanian translation request vs unrelated Thai-titled issue; no same underlying change." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation request vs unrelated Thai-titled issue; no same underlying change." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Placeholder-title translation issue vs different Thai-titled issue; not the same underlying request." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Placeholder-title translation issue vs different Thai-titled issue; not the same underlying request." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Same placeholder title and near-identical body pattern; likely duplicate issue." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder-title issue vs different Thai-titled issue; not the same underlying issue." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Identical Thai title and likely identical body; clear duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate calls on 43834/44365/44370 are supported by identical placeholder titles plus very high body similarity, and 46222/46226 is a clear exact pair. The language-specific issues are correctly treated as separate rather than collapsed based on broad i18n similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-18", + "nodes": [ + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:42738|issue:46653", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:43909|issue:46645", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133", + "issue:42503|issue:46653", + "issue:45027|issue:46653" + ], + "prepared_review_unit_hash": "ac1e94686d7e52323433fe8a06199721c81c58eed909a3aa61cda15a24be0e9c", + "component_budget": { + "node_count": 18, + "item_count": 18, + "soft_pair_count": 39, + "serialized_chars": 23784, + "estimated_input_tokens": 5946, + "estimated_eval_tokens": 12148 + }, + "budget": { + "node_count": 18, + "item_count": 18, + "soft_pair_count": 39, + "serialized_chars": 23784, + "estimated_input_tokens": 5946, + "estimated_eval_tokens": 12148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a loose collection of model-support request issues, not true duplicates. The soft-similarity links mostly reflect the shared \u201cadd model/support\u201d wording, but each item targets a different model or capability, so all soft edges should be rejected.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 43671 is a representative, well-scoped model-support request and sits near the center of the similarity graph, but it is still a specific Qwen3-TTS request rather than a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43671 is the best canonical issue only as a representative of the recurring feature-request pattern; it is clear and specific, but still not interchangeable with the other tickets.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert config support vs DEIMv2 support; different models and different requests." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config support vs Qwen3-TTS support; unrelated model families and tasks." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert config vs RF-DETR; same broad theme but clearly different model additions." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are distinct model-add requests, not the same change." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR detection model vs LFM2.5 Audio support; different domains and code paths." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR model add vs weight-sparse transformer request; unrelated concrete changes." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR vs Qwen3-ASR; different models and feature requests." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR vs Qwen3-TTS; distinct model-support tickets." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT vs Qwen3-ASR; different model families and functionality." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT addition vs sparse-transformer support; not the same underlying change." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT vs LFM2.5 Audio; unrelated model requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are separate model-add requests." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT vs ModernVBERT; similar naming but different models and requests." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support vs Qwen3-ASR support; different model targets." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 architecture vs sparse transformer; different concrete features." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 vs ModernVBERT; distinct model support requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture vs Voxtral-TTS support; different models and modalities." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model additions." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT vs Qwen3-ASR; different model support requests." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT models vs Qwen3-TTS; unrelated requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT model support vs BERT-like RoPE support; one is a model request, the other is an architectural feature request." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support vs Voxtral TTS support; different change types." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "Architecture feature request vs unrelated person-named issue; no shared underlying bug/change." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS vs Qwen3-ASR; related brand but different modalities and implementation work." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS vs Voxtral TTS; both audio-related, but different models and support paths." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS vs DeepSeek-OCR2; different models and tasks." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support vs sparse transformer request; different code paths." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR vs LFM2.5 Audio; both audio-related, but distinct model support issues." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR vs EUPE; different model/product targets." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR model support vs sparse transformer request; not the same change." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio vs SAM3-LiteText; different model additions." + }, + { + "left": "issue:43909", + "right": "issue:46645", + "accept": false, + "reason": "Audio model support vs Granite SWA+sink option; one is a model request, the other a model option tweak." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse transformer model request vs EUPE model request; unrelated additions." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR vs Voxtral-4B-TTS; different audio models and support work." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT vs Voxtral TTS; different model families and tasks." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR vs DeepSeek-OCR2; different models and modalities." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE by Meta AI vs TIPSv2 by Google DeepMind; separate model requests." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "ModernVBERT model request vs unrelated name-only issue; no shared change." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "Voxtral TTS support vs unrelated name-only issue; not a duplicate." + } + ] + }, + "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": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-38740-2", + "nodes": [ + "issue:38740", + "issue:46443" + ], + "soft_pairs": [ + "issue:38740|issue:46443" + ], + "prepared_review_unit_hash": "c0727bdfe6f78664f8ddedf6565280f285b6789fbed058b6f47da09605d3b963", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are unrelated: one is a documentation/request item about adding Pruna as an optimization framework, and the other is a Vision RoPE refactor. They do not describe the same bug or change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No duplicate cluster here; if forced to pick a representative, #46443 is the more substantive code-change issue, while #38740 is a docs request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different topics and outcomes: docs framework addition vs. Vision RoPE refactor. No shared code path, target, or change request." + } + ] + }, + "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": 23, + "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": "c618097dc32aa465692d8dff4aece3bfb3e21d8afa347f6f846060a9d2f275fd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are unrelated: one asks for generic block decoding behavior, while the other requests DeepSeek v3.2 support.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No true canonical duplicate exists; if one issue must represent the set, 43037 is the more concrete, actionable request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different underlying concerns: block decoding mechanics vs model/version support for DeepSeek v3.2." + } + ] + }, + "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": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41306-2", + "nodes": [ + "issue:41306", + "issue:41826" + ], + "soft_pairs": [ + "issue:41306|issue:41826" + ], + "prepared_review_unit_hash": "4e5a026c557c0d44084188830cb16b66ab61ec9101b33a8f5dffd329672c07c4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1585, + "estimated_input_tokens": 397, + "estimated_eval_tokens": 1050 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are only loosely related by the general theme of memory-efficient model components, but they describe different features: one is about exposing a TiledFusedLogitsLoss/Liger FLCE flag in `from_pretrained()`, the other is about integrating TiledMLP for lower memory usage. They do not look like duplicates.", + "confidence": 0.89, + "canonical_issue_reason": "Issue #41306 is the more concrete and actionable request, tied to a specific API flag and code path (`from_pretrained()`); #41826 is broader and more exploratory.", + "canonical_pr_reason": null, + "best_issue_reason": "#41306 is the best representative issue because it has a clear, specific implementation target, whereas #41826 is a higher-level feature idea.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41306", + "right": "issue:41826", + "accept": false, + "reason": "Different requested changes: one is an API flag for TiledFusedLogitsLoss/FLCE in `from_pretrained()`, the other is integrating TiledMLP. Same broad memory-efficiency theme is not enough to treat them as the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 23, + "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": "1b5ded1d467fb7607f8d993eb9ef6eddbe693f1d92c460a6c9cd56da09ddbca5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are related only at a high level (custom trainers / Trainer internals), but they describe different problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They do not look like duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue fits both: the first targets input-column handling in Trainer, while the second targets FLOPS accounting for custom trainers.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43598 is the more concrete and narrowly scoped report, but it should not be treated as a duplicate of 43600.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Same broad Trainer/custom-trainer theme, but the underlying bugs are different code paths and different expected fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 23, + "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": "d6580d97390f98d5547d9ae4e6af3df283d43a48317e1ebedad321b9078b1b19", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "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 level: both ask for additional AutoModel support for Qwen variants, but they target different models and different auto classes. They do not look like duplicates of the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader and more foundational request: it asks for Qwen3-Omni registration in AutoModel and AutoModelForConditionalGeneration, which is a generic model-registration gap rather than a specific task head request.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the better canonical issue because it covers the general registration work and is more likely to subsume related follow-up support requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying requests: #43684 is about registering Qwen3-Omni in AutoModel/AutoModelForConditionalGeneration, while #44405 is specifically about AutoModelForSequenceClassification support for Qwen3.5/Qwen3_5Config. Same ecosystem, but not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 23, + "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": "e67092b1830081399685bd921ec0a35725cf8e6f9692f7caebdc89eb3789c4c8", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three issues share the same generic 'huggingface' title and appear to be near-duplicates; 'Fett' looks unrelated. Among the huggingface issues, 43845 is the best representative/canonical candidate because it is the earliest of the closely matching set.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 43845 is the earliest and most representative of the near-identical 'huggingface' issues (43845/43846/43859), with slightly more context than 43859 and strong textual overlap with 43846.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the best global issue candidate because it anchors the cluster of highly similar huggingface reports and is the most plausible single dedup target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no substantive body evidence tying 'Fett' to the huggingface reports." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no concrete evidence of the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no concrete evidence of the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same title, very high textual similarity, and no visible evidence of distinct bugs; these look like duplicate reports." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same generic title and strong similarity suggest the same underlying issue despite the shorter body." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same generic title and strong similarity make these likely duplicate reports of the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded in the packet: the three 'huggingface' issues have strong soft-similarity, and 43845 is the earliest among that cluster. The 43835 links are conservatively rejected based on only weak similarity and unrelated title." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 23, + "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": "adbd4a222bbdc71b9919b24174fd3e0df6f6ae46d7ef0fd395ee059daede2359", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are low-information spam/vague issues with no concrete shared bug or change. The apparent similarity is only from generic titles, so none of the soft pairs should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are generic spam placeholders with no body, targets, or filenames; no concrete shared issue is visible." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same generic spam-only pattern; no evidence they describe the same underlying bug or request." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam-like and the other is a vague documentation suggestion; too little specificity to treat as the same issue." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "The titles are unrelated and there is no issue detail tying them to a common bug or task." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A vague documentation enhancement is not the same as a spam placeholder; no shared concrete change request." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Only broad low-signal similarity; no explicit overlap in problem statement or requested change." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders, but there is no evidence they refer to the same underlying report rather than separate junk submissions." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "No shared topic beyond generic similarity; the titles do not indicate the same bug or request." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete common issue is present; these look like unrelated reports." + } + ] + }, + "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": 23, + "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": "db1bc3ea61423c719476563c174d36dba7f3f6193a245800423d5b712b8604a2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues have identical titles, close creation times, and very similar body sizes, so they look like duplicate reports of the same underlying request. Issue #45145 should be the canonical thread because it was opened first and already has an inbound reference.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #45145 is the better canonical issue because it was created first and already has an inbound reference, making it the more established thread.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45145 is the strongest cluster representative since it predates #45151 and is the more established issue thread.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Identical titles, close timestamps, and highly similar body lengths with soft-similarity support treating these as duplicate reports of the same issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Reject: the pair looks plausibly duplicate based on identical titles, close timestamps, similar body lengths, and soft-similarity, but the analyst overstates the evidence by claiming the bodies have the \"same wording\" and that #45145 has \"slightly more engagement.\" Those specifics are not directly grounded in the packet. The canonical issue choice is fine, but the cluster summary and best-issue rationale should stay more conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 23, + "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": "07aab97c34af37d97bb703c8aeeff089d607bff3e48f62da58df57ea0acf8aff", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate 4-bit inference bugs: one in Gemma4 audio modules hitting a uint8/torch.finfo TypeError, and one in BitsAndBytesConfig where llm_int8_skip_modules drops the default lm_head exclusion and triggers an AssertionError. They are related only at the broad feature level, not the same underlying defect.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45674 is the more general, configuration-focused report with a clearer failure mode and broader applicability across 4-bit inference.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45674 best represents the cluster because it describes a concrete regression in BitsAndBytesConfig rather than a model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Different root causes and code paths: Gemma4 audio-module dtype handling vs. BitsAndBytesConfig skip-module exclusion logic. They share 4-bit inference context but are 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": 17, + "component_count": 23, + "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": "665531712120e22ef4ee9ac8fc2c898d4b840e3265f95f6a50c44c8d91361548", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues look unrelated: both are bare name/date titles with no body, no shared targets, and no evidence of the same underlying bug or request. The soft similarity is not enough to treat them 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:45762", + "right": "issue:45763", + "accept": false, + "reason": "Titles refer to different people and different dates; there is no issue body, shared target, or other evidence of the same underlying problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 23, + "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": "02540700f82477fae749111438a1843e60d6d625f7645aa410efd2b2c20aedc5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear unrelated; both are just person-name/date titles with no body, links, or shared targets, so there is no evidence they describe the same underlying bug or request.", + "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": "No substantive overlap beyond both being empty closed issues; titles point to different people/dates and there is no shared body, target, or code context." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45810-2", + "nodes": [ + "issue:45810", + "issue:46295" + ], + "soft_pairs": [ + "issue:45810|issue:46295" + ], + "prepared_review_unit_hash": "2b1fe1dd88b7166227b5e43a9b205589eea44af95dcf9bb0c8a4d9c719fec55c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one requests a new Qwen token-classification model class, while the other asks for FP8 training support. They do not describe the same bug or change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45810 is the more concrete, narrowly scoped request and is easier to use as a representative item, but it is not a duplicate of the FP8 training issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different features in different areas: adding a model class for token classification versus training/runtime support for FP8. No shared code path or concrete duplicate behavior is evident." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 23, + "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": "c2a695b21afc2d5d46a0dafca58f5de3b7b36ad35b276a385a98602fa6d3d348", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5726, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "There are two clear duplicate pairs: 45870 \u2194 45871 and 45872 \u2194 45873. Issue 45869 is a different dataset/promo request and should not be grouped with either pair. The two DataForge-themed pairs are superficially similar but describe different scopes, so they should not be merged across pairs.", + "confidence": 0.96, + "canonical_issue_reason": "45872 is the strongest canonical representative for the cluster because it is the more concrete DataForge request and is an exact duplicate of 45873; the other exact-duplicate pair (45870/45871) is a separate, more generic example issue.", + "canonical_pr_reason": null, + "best_issue_reason": "45872 is the most actionable issue in the set: it clearly states the DataForge time-series + scraping-service request, while 45869 is unrelated and 45870/45871 are a separate duplicate pair about a generic example.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different requests: dataset bundle promo/sample vs. a DataForge example." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different requests: dataset bundle promo/sample vs. a DataForge example." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Same title/body pattern and effectively the same example issue; this looks like a direct duplicate." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Related theme, but not the same change: one is an example, the other is a scraping-service request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Same title/body pattern and effectively the same DataForge request; this looks like a direct duplicate." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45870\u219445871 and 45872\u219445873 are the only clearly identical pairs (matching titles/body lengths and jaccard 1.0), while 45869 is a separate dataset bundle request. The soft-edge rejections between the example issues and the DataForge service issues are conservative and consistent with the titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46439-2", + "nodes": [ + "issue:46439", + "issue:46441" + ], + "soft_pairs": [ + "issue:46439|issue:46441" + ], + "prepared_review_unit_hash": "b3b9a949e9f8d77070222ac37fd32260fbc11d12e471ccc9c825c74e3188bcc6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both issues concern `cache_implementation=\"static\"` and linear-attention, but they describe different failure modes: one is `early_initialization` producing wrong shapes and a `RuntimeError`, the other is a Qwen3.5 `KeyError: 'linear_attention'`. I would not merge them as the same underlying bug from the available evidence.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 46439 is the better representative because it names the broader root path (`early_initialization` + linear-attention shape corruption) rather than a model-specific crash symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46439 is the strongest standalone report: it is more general, more diagnostic, and describes the concrete corruption mechanism rather than just a downstream exception.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Same broad static-cache/linear-attention area, but different observed failures (`RuntimeError` from shape corruption vs `KeyError` for Qwen3.5); not enough to treat as 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": 22, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46551-2", + "nodes": [ + "issue:46551", + "issue:46552" + ], + "soft_pairs": [ + "issue:46551|issue:46552" + ], + "prepared_review_unit_hash": "e336edbe9e1faab0498ee7abb05e6d0c3c963c494c05ebe71f930afcc0267cd5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues appear to be the same bug report for Wav2Vec2CTCTokenizer v5 treating the word delimiter as a special token, affecting special-tokens masking and token conversion. They differ only by issue number/state, so they should be clustered together.", + "confidence": 0.99, + "canonical_issue_reason": "Issue #46552 is the better canonical because it is open, has inbound references, and appears to be the active report for the same exact problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #46552 is the strongest representative since it is the active open instance of the identical title/body and is more likely to collect ongoing discussion and resolution.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "Exact same title and same reported tokenizer bug; no meaningful difference in scope or content beyond issue metadata." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analyst\u2019s clustering is grounded in the packet: both issues have the exact same title and identical body length, and the open issue (#46552) has inbound references while the other is closed. The soft-edge verdict is conservative and supported by the evidence, so accepting the pair is reasonable." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 23, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46657-2", + "nodes": [ + "issue:46657", + "issue:46692" + ], + "soft_pairs": [ + "issue:46657|issue:46692" + ], + "prepared_review_unit_hash": "98e4d0399c058e384b25bc0ba7abde8ef44ec27987bfefbd245c091be2f9840d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1566, + "estimated_input_tokens": 392, + "estimated_eval_tokens": 1040 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1566, + "estimated_input_tokens": 392, + "estimated_eval_tokens": 1040 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one reports a tensor shape mismatch during object-detection training, while the other reports an out-of-memory problem after several epochs in RT-DETR v2. They may both occur in training workflows, but they describe different failure modes and likely different fixes.", + "confidence": 0.91, + "canonical_issue_reason": "Issue #46657 is the better representative because it describes a concrete, actionable code-path failure (tensor size mismatch) rather than a broad resource symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "#46657 is the most suitable issue to anchor the cluster because it is specific and reproducible enough to guide debugging; #46692 appears to be a separate training-memory regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46657", + "right": "issue:46692", + "accept": false, + "reason": "Different underlying problems: shape mismatch vs OOM after epochs. Same training context is not enough to treat them as the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-14", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267" + ], + "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:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "498fdf428c0bd5df3d06fce3c155652f2f420e5c21d2add8f0ba171e6714005f", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38153, + "estimated_input_tokens": 9539, + "estimated_eval_tokens": 19334 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38153, + "estimated_input_tokens": 9539, + "estimated_eval_tokens": 19334 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: model additions, docs, tokenizer/processor tweaks, refactors, and tests that do not describe the same underlying change. The only true duplicate-like pair is the two `fixed_cross_entropy` PRs, which both implement the same cross-entropy kwargs support for issue 43240.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best representative of the only real duplicate pair because it explicitly targets issue 43240 and presents the same `fixed_cross_entropy` kwargs expansion with clearer issue context.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest cluster representative for the duplicate loss change; the rest of the cluster is unrelated, and 43251 is the cleaner, issue-linked version of the same fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "VideoPrism model/docs addition vs. an ExecuTorch dynamic-shapes export test; unrelated changes." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Mistral tokenizer helper methods vs. adding MobileLLM; different subsystems and no shared bug." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer API addition vs. ExecuTorch export testing; not the same underlying change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor-auto robustness fix vs. MobileLLM model addition; unrelated code paths." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Cross-entropy kwargs fix vs. MobileLLM model addition; no overlap in underlying bug." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Loss utility kwargs fix vs. Mistral4 test/model work; different bug and code path." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Both touch tests, but one is ExecuTorch export dynamic-shape coverage and the other is Mistral4 test/model fixes; not the same change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "ExecuTorch dynamic-shapes export test vs. MobileLLM model addition; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New MobileLLM model support vs. a `cls.config` resolution fix; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM model addition vs. Falcon output-collection refactor; different model families and goals." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor vs. model-creation skill docs/workflow; not the same code change." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output-interface refactor vs. Mistral4 test fixes; different models and bug targets." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes vs. `layrnorm` renaming across CLIP-like models; unrelated fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test fixes vs. serve chat parsing update; different subsystems and changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes vs. TDT loss kernel integration; unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test/model fixes vs. class config resolution order fix; different code paths." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 test fixes vs. model-creation skill docs/workflow; not duplicates." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition vs. model-creation skill; unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs change `fixed_cross_entropy` to pass additional `cross_entropy` kwargs for issue 43240; same underlying bug and mergeable into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The summary correctly treats the cluster as heterogeneous and only elevates the 43251/43254 pair, which is supported by the packet: both touch `src/transformers/loss/loss_utils.py`, both target issue 43240, and both expand `fixed_cross_entropy` to forward `weight`/`label_smoothing` to `nn.functional.cross_entropy`. The other soft pairs are reasonably rejected as unrelated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "4265ff26508be1eb13fc57e155ae2191bf7c6187969ac8b018e57aefd7e93e9c", + "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": "Both PRs are about VibeVoice, but they are not the same change: #40546 is the core VibeVoice implementation, while #42781 adds a separate realtime/acoustic-tokenizer path and generation changes. The filename overlap is expected for a follow-on feature, not a duplicate fix.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "#40546 is the broader, original VibeVoice implementation and is the better anchor for the cluster; #42781 looks like a later extension focused on realtime support rather than the same mergeable change.", + "best_issue_reason": null, + "best_pr_reason": "#40546 is the most representative and complete PR for the VibeVoice model introduction, with far more review activity and broader coverage than the draft realtime add-on.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "They share the VibeVoice codebase, but #42781 is a distinct realtime feature branch (including acoustic tokenizer and generation tweaks), not the same concrete bug/change as #40546." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "fc563cab6cdbf2abe23c03b8e823c8b7168af4f1d328359e2edf8d9031933c41", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated feature additions in different model families, so they should not be clustered as duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the cluster contains two unrelated PRs, one for TimesFM covariates and one for DINOv3 image classification support.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good cluster representative for duplicate triage because they address different code paths and different models.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different model subsystems and different changes: TimesFM forecasting covariates vs DINOv3 image classification support. They do not look like the same underlying bug or mergeable 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": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "51d36f3504c2d0785e6f009c3dbc87d9f6338d06c098d70bc3873e8c6653f3eb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are workflow maintenance updates, but they target different dependency bumps and different failure modes: 42975 is centered on GitHub Actions Node 24 compatibility, while 42976 upgrades Docker build actions and other workflow action versions. They overlap in CI files, but not in the same concrete change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs are related only at a broad CI-maintenance level, but they update different actions for different reasons and are not the same underlying change.", + "best_issue_reason": null, + "best_pr_reason": "42975 is the better representative of the broader GitHub Actions compatibility effort because its title and diff clearly describe the Node 24 upgrade theme across many workflows.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Reject: both are CI/workflow dependency bumps, but 42975 upgrades actions for Node 24 compatibility while 42976 updates docker/build-push-action and other actions to newer versions. They touch overlapping files yet do not fix the same concrete code-path problem and would not naturally merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-7", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490", + "pull_request:46562", + "pull_request:46658", + "pull_request:46705" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:42978|pull_request:46705", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43448|pull_request:46705", + "pull_request:43451|pull_request:45490", + "pull_request:43451|pull_request:46705", + "pull_request:45490|pull_request:46562", + "pull_request:45490|pull_request:46705", + "pull_request:46562|pull_request:46658" + ], + "prepared_review_unit_hash": "61fbb237f91aca9a808b29df9aef39a6c7a9adbc2f6da432d0b7aa07f0e15b53", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 12, + "serialized_chars": 24369, + "estimated_input_tokens": 6093, + "estimated_eval_tokens": 12442 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 12, + "serialized_chars": 24369, + "estimated_input_tokens": 6093, + "estimated_eval_tokens": 12442 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are mostly unrelated model-addition branches that only overlap on standard docs/auto-registration files. The only likely duplicate-like pair is the two Step 3.7 Flash PRs (46562 and 46658); everything else appears to be a distinct model or feature addition.", + "confidence": 0.78, + "canonical_issue_reason": null, + "canonical_pr_reason": "46562 is the clearest representative of the one near-duplicate pair: it is the more polished Step 3.7 Flash model-addition PR, while 46658 looks like a WIP/companion implementation for the same feature.", + "best_issue_reason": null, + "best_pr_reason": "46562 is the best standalone PR to anchor the cluster because it corresponds to the same Step 3.7 Flash feature as 46658 and is the more complete user-facing addition; the rest of the PRs are separate model introductions.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model families (ViT NEPA vs Molmo). Shared docs/auto files are just boilerplate registration, not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model families (ViT NEPA vs Molmo2); same scaffolding files do not indicate the same code-path change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Unrelated model additions (ViT NEPA vs CTSM). Overlap is only in generic docs/auto wiring." + }, + { + "left": "pull_request:42978", + "right": "pull_request:46705", + "accept": false, + "reason": "A new model PR and a distributed-config PR are different kinds of changes with no shared concrete bug or feature." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are distinct model releases/families, not the same underlying implementation or fix." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are unrelated model additions; shared repo wiring does not make them duplicates." + }, + { + "left": "pull_request:43448", + "right": "pull_request:46705", + "accept": false, + "reason": "Model addition vs distributed configuration work are unrelated changes." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are different models in different domains, so they are not the same change." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46705", + "accept": false, + "reason": "Molmo2 model support and distributed config work do not touch the same underlying code path." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "CTSM time-series model addition and Step 3.7 Flash multimodal model addition are unrelated." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46705", + "accept": false, + "reason": "CTSM model support and distributed config changes are different features with no shared concrete fix." + }, + { + "left": "pull_request:46562", + "right": "pull_request:46658", + "accept": true, + "reason": "Both PRs appear to implement the same Step 3.7 Flash model addition; 46658 reads like a WIP/mini-model companion for the same underlying feature and 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": 7, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-9", + "nodes": [ + "pull_request:43291", + "pull_request:43498", + "pull_request:43665", + "pull_request:44815", + "pull_request:45114", + "pull_request:46380", + "pull_request:46539", + "pull_request:46579", + "pull_request:46719" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43498|pull_request:46579", + "pull_request:43498|pull_request:46719", + "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:43665|pull_request:46579", + "pull_request:44815|pull_request:46539", + "pull_request:44815|pull_request:46380", + "pull_request:44815|pull_request:46579", + "pull_request:45114|pull_request:46579", + "pull_request:46380|pull_request:46539", + "pull_request:46380|pull_request:46579", + "pull_request:46579|pull_request:46719", + "pull_request:43665|pull_request:46719" + ], + "prepared_review_unit_hash": "2231cae11425457db498de4bf865f651bdba33602cb9436e12ebb84b67394f21", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 17, + "serialized_chars": 26394, + "estimated_input_tokens": 6599, + "estimated_eval_tokens": 13454 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 17, + "serialized_chars": 26394, + "estimated_input_tokens": 6599, + "estimated_eval_tokens": 13454 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: a few weakly similar maintenance/test-fix PRs appear alongside unrelated generation, CI, FP8, sparse-attention, and typo-cleanup changes. I don\u2019t see any pair that clearly fixes the same underlying bug or could realistically be merged as one PR.", + "confidence": 0.34, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46579 is the cleanest single-theme representative in the set: a broad but coherent typo/cleanup sweep across code, with a clear purpose and no sign of mixed bug classes.", + "best_issue_reason": null, + "best_pr_reason": "PR 46579 is the best standalone representative because it is internally consistent and better described than the others, which are either ambiguous or clearly about different subsystems/problems.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Both are test-related, but one is Whisper decoding/test fixes and the other is CLIP/ViT test edits; they do not look like the same underlying bug or change." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenizer/test changes and a dequantization fix in FP8/model loading are unrelated code paths." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "These touch different areas: Whisper decoding vs generation/watermarking validation and many model tests." + }, + { + "left": "pull_request:43498", + "right": "pull_request:46579", + "accept": false, + "reason": "Backward-compatibility for tie_weights is unrelated to a broad typo cleanup sweep." + }, + { + "left": "pull_request:43498", + "right": "pull_request:46719", + "accept": false, + "reason": "Different problems: a compatibility alias for tie_weights versus sparse-attention/minimax model logic changes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "Test-only edits in CLIP/ViT do not match the CI memory/device-map integration fix in Phimoe tests and workflows." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "A set of test edits is not the same as a dequant/fp8/model-loading bug fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "These are not the same underlying change: random model test edits versus generation config/Whisper validation fixes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46579", + "accept": false, + "reason": "Model test edits are unrelated to repository-wide typo cleanup changes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46719", + "accept": false, + "reason": "No shared concrete bug: broken-looking test edits vs sparse attention/minimax model code." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8/dequantization and Phimoe CI test infrastructure are different subsystems and failure modes." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "These fix different code paths: dequant/model-loading versus generation validation and model tests." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46579", + "accept": false, + "reason": "A numeric/dequantization fix is not the same as typo cleanup across many files." + }, + { + "left": "pull_request:45114", + "right": "pull_request:46579", + "accept": false, + "reason": "Doctest documentation cleanup and code typo fixes are only broadly similar, not the same PR-worthy bug/change." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation validation/Whisper fixes do not match the Phimoe CI resource-limit workaround." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46579", + "accept": false, + "reason": "These are unrelated: generation config behavior versus typo corrections in assorted code files." + }, + { + "left": "pull_request:46579", + "right": "pull_request:46719", + "accept": false, + "reason": "Typo cleanup across code is not the same underlying change as sparse-attention and Minimax model fixes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46719", + "accept": false, + "reason": "Test-file edits and sparse-attention/model implementation changes are different bugs and should not be merged together." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "b7fed5af45d39e1e229019f3d0b9af9855f78f86bafa24dfc5d3f948ec2f79c3", + "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": "These two open PRs are both marked for test/bot experimentation, but they change different parts of the repo: one is a setup/dependency/version tweak, the other is a CI/workflow and testing-utility change. Their only overlap is a shared Vit test file, which is not enough to treat them as the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #43532 is the more representative item if one must be chosen because it spans the CI/testing harness surface across multiple files; #43488 is a narrower setup/dependency/test tweak and does not look like the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Reject: the overlap is only a shared Vit test file. #43488 changes setup.py/dependency metadata and a test fixture, while #43532 changes GitHub workflows, testing utilities, and related notification code. They are not the same underlying bug or fix and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "fd4b8ef3ceb337e9eb5d4cdd4c74b87e4855ee8156d18cae4b3380d3e5999710", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs add new segmentation pipelines and share scaffolding, but they target different features: promptable concept segmentation vs promptable visual segmentation. The overlap is mostly documentation and framework registration, not the same concrete change set.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: these are distinct pipeline additions with different underlying model/support integration and different task semantics.", + "best_issue_reason": null, + "best_pr_reason": "If a representative PR is needed, 43613 is the broader implementation with more model/test integration; however, it is still not a duplicate of 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registration scaffolding is not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "6a102a949fcab3c1aa7219fe4a021eb28daa3660f7acb6702ffdbc9c7d9972f0", + "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 two PRs are both in `trainer.py` but they address different extension points: one changes gradient-accumulation loss scaling, the other factors out signature-based column filtering. They are not the same change and should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 43651 is the stronger representative of the cluster only because it makes a concrete runtime behavior change in `training_step`; however, it is not a duplicate of 43663, which adds a separate customization hook for column filtering.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and behaviors: 43651 changes loss scaling during training, while 43663 refactors model signature inspection for dataset column filtering. They are not the same underlying fix and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "2f2fc1bf7bf5366a35e51061ce23bc3653e26c64dfba490238e5957b5d3304d1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch the same CLI file and both fix annotation/import-time NameError risks, but they target different runtime failures. They should stay separate; no duplicate merge is warranted.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43842 is the most direct match to the TypeAdapter/pydantic NameError: it moves the TypeAdapter import under TYPE_CHECKING and quotes the annotation in `_validate_request`.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the best representative because it is narrowly scoped to the concrete TypeAdapter NameError path, whereas #43656 is a broader annotation/postponed-evaluation cleanup affecting other symbols.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Both are annotation-related changes in `src/transformers/cli/serve.py`, but they fix different problems: #43656 addresses runtime NameErrors from unquoted model/processor type hints, while #43842 fixes `TypeAdapter` resolution when pydantic is absent. They are related, not the same concrete bug or mergeable duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44722", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request: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:44073|pull_request:44161", + "pull_request:44059|pull_request:44071", + "pull_request:44056|pull_request:44101", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44068", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:43997|pull_request:44018", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44010|pull_request:44073", + "pull_request:44018|pull_request:44129", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:43997|pull_request:44129", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "ed60f99e4255811baefba09b4014c66eca51e8803c436ae19412ebcad2532001", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84899, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75095, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a batch of model-specific output-tracing refactors tied to the same umbrella effort (#43979), but most artifacts are separate model implementations and should not be merged as duplicates. Only a few pairs cover the same concrete model code path.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best anchor: it explicitly targets #43979 and captures the core GPT-J/CodeGen output-tracing migration in a representative, non-trivial patch.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the strongest representative of the cluster because it directly implements the standardized output-tracing refactor for the core GPT-J/CodeGen path and links the work to the tracking issue.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths (CVT/FNet vs GPT-J), so this is not the same underlying change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "CVT/FNet and DeBERTa v2 are unrelated implementations; same umbrella theme, but not the same bug/change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same code path; 44085 looks like a narrower variant of the same change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet and EfficientNet are separate model implementations; these are distinct refactors, not duplicates." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet output tracing and GPT-J/CodeGen output tracing are different model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen versus EfficientNet; same tracing pattern, but not the same underlying code-path problem." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are unrelated model code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different models; this is only a thematic similarity." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are separate model implementations, so these are not duplicates." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both target GPT-Neo output tracing on the same model path; they appear to be the same refactor expressed in slightly different patches." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J and GPT-Neo are different model code paths, even though both are output-tracing refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations, so not the same underlying change." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated implementations; same theme, not same bug/change." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are separate model paths, so these are distinct changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo versus GPT-J/CodeGen: same family of refactor, but different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated models." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different model code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 and MGP-STR are unrelated model implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different code paths." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and XLM/Flaubert are separate model families; not the same change." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT2 and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are unrelated implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and SegFormer are distinct vision model implementations; same broad area, but not the same bug/change." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44161", + "accept": false, + "reason": "MPT and LongT5 are unrelated model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are separate model implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated models." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and XLM/Flaubert are different code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are unrelated model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are unrelated models." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are unrelated model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and SegFormer are different code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBERT and MPT are unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBERT and VisualBERT are different model code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBERT and GPT-Neo are unrelated implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44129", + "accept": false, + "reason": "TextNet and SpeechT5 are different model code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBERT and XLM/Flaubert are unrelated model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44071", + "accept": false, + "reason": "MPNet and MPT are separate model code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBERT and DeBERTa v2 are unrelated models." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa v2 and TextNet are different model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen standardized output-tracing refactor on the same core code path; these look like duplicate versions of the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same model path; 44722 is the broader version of the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet are distinct model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are different model implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and ViTs are unrelated model code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly same umbrella work with separate model-specific code paths, and the only accepted pairs are the ones that clearly share the same model path and appear mergeable as duplicate/refinement PRs. No pair verdict appears overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 35, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "28e62bb5b94574c104a3bfe1cd809f804c8bf8dcd74f8c59a6e421d2d43a4a9f", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84899, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17125, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All items are pull requests in the same broad Hugging Face refactor wave for standardized output tracing, but they touch different model code paths and should not be deduplicated as the same fix. No issue is present in the cluster. The best representative PR is SpeechT5 because it has the most concrete scoped change and an explicit issue link.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44129 is the strongest canonical representative: it has an explicit issue target, a clear refactor title, and a focused change to one model\u2019s output tracing implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR 44129 is the best overall representative of the cluster because it is concrete, well-scoped, and clearly tied to a tracked issue; the other PRs are similar but less anchored.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized-output refactors, but they apply to different model files and different execution paths (FocalNet vs ImageGPT), so they are not the same underlying change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "These touch VisualBert and TextNet respectively; despite similar decorator migration, they are separate model-specific refactors and not one mergeable fix." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert have distinct tracing implementations and model internals; this is a shared refactor theme, not the same bug or change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are unrelated code paths; the similarity is only the output-capture migration pattern." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are separate model implementations, so these are parallel refactors rather than duplicates of one concrete fix." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 each change their own output tracing logic; they are not the same underlying code-path problem and would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44028", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "0d24dfe595ab6f3ac4550a79accf0e97a4fc4d2615ded0e990cb9f818ede3f0e", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40172, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a cluster of similar-looking output-tracing refactors across many different model implementations, not a set of duplicate fixes for one shared bug or code path.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the strongest representative of the cluster: it is a broader two-file Mamba/FalconMamba refactor and includes the shared output-capture machinery, so it best captures the overall theme.", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the best single representative because it is the most substantial and central example of the output-tracing refactor pattern in this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and code paths; only the refactor pattern is shared." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "5939ab1f80e8a160ad19dcefa2bf471a31d686ec2c91dd75f3410b0192091d81", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two different model-addition PRs that only overlap on shared Transformers plumbing and docs/auto-registry files. They target different models and code paths, so they should not be merged as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear duplicate pair here; if forced to pick a representative, PR #45401 is the more focused and better-scoped change because it has an explicit issue target and is centered on one concrete TTS feature.", + "best_issue_reason": null, + "best_pr_reason": "PR #45401 is the better standalone representative for the cluster because it is narrowly tied to Voxtral TTS support and includes an explicit tracking issue, whereas PR #44662 is a broader PenguinVL model implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model implementations (PenguinVL vs Voxtral TTS) with different filenames, model classes, and functionality. Shared auto/config/docs files are expected for any new model integration and do not indicate 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": 16, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "3ab961925b038cad1e75aa3374c84841413c7e15ab0900818122a6f3fe41eac1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related thematically around module/kernel fusion, but they are not the same change: #44979 introduces a new `module_fusion.py` API scaffold, while #45363 wires `KernelConfig` into model loading and adds integration/example code for n-to-1 kernel fusion. They should not be deduplicated.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "#45363 is the better cluster representative because it is the more concrete end-user feature: it spans loader/integration code, `KernelConfig`, and a usage example, whereas #44979 is mainly the lower-level module fusion API groundwork.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related fusion work, but different concrete changes and code paths: one adds a module-fusion API module, the other adds kernel-config-based n-to-1 fusion integration. They are not mergeable as one PR without conflating separate features." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "81dcc477885e4f5405e1199838c536737f2a35baf3b13306b4e15186151ed7d9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests fix the same Mistral regex AttributeError in the same code path, with 45317 appearing to be the more complete version because it adds tests. They are strong duplicates and can be treated as one change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR because it implements the same fix as 45086 and also adds test coverage, making it the more complete patch.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best representative for the cluster since it contains the same bug fix plus tests, while 45086 is the narrower earlier version.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same file and code path, with 45317 extending 45086 by adding tests." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-4", + "nodes": [ + "pull_request:45115", + "pull_request:45569", + "pull_request:46332", + "pull_request:46565" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569", + "pull_request:45569|pull_request:46565", + "pull_request:46332|pull_request:46565" + ], + "prepared_review_unit_hash": "7192bd8d301fc5cd72a77703e7ae5546e6dd48d26587f30f68544af8318bbca8", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 13738, + "estimated_input_tokens": 3435, + "estimated_eval_tokens": 7126 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 13738, + "estimated_input_tokens": 3435, + "estimated_eval_tokens": 7126 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all Nemotron-related but split across different model families and feature tracks: NemotronH refactors versus Nemotron ASR streaming. The overlaps are broad thematic similarity, not duplicate code-path changes.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the most central NemotronH architecture change and most likely subsumes the narrower 45115 refactor, making it the best representative PR in this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 has the clearest, broadest scope within the NemotronH line and is the strongest anchor for this cluster; the ASR PRs are separate model additions rather than the same change.", + "soft_edge_verdicts": [ + { + "left": "Refactor/nemotron h inherit granitemoehybrid", + "right": "Proper nemotron H and 3 and 2", + "accept": false, + "reason": "Related model family, but 45115 is a narrow inheritance refactor while 45569 is a broader split/cleanup across NemotronH dense/sparse models; not the same concrete change." + }, + { + "left": "Proper nemotron H and 3 and 2", + "right": "Add Nemotron 3.5 ASR Streaming", + "accept": false, + "reason": "Both touch Nemotron naming and auto/docs plumbing, but one is NemotronH architecture work and the other is ASR streaming; different model families and code paths." + }, + { + "left": "Add Nemotron ASR Streaming", + "right": "Add Nemotron 3.5 ASR Streaming", + "accept": false, + "reason": "Same ASR area, but these are separate model-version additions, not one shared bug or a single change that could plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "3b1168a35d2cdffd5711712378df7332a21cbc4da2e396294ca7c254924c7ace", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related PRs in the same heterogeneous-config initiative, but they are not duplicates: one adds the per-layer config plumbing, the other consumes it to add broader modeling/runtime support.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45333 is the narrower foundational change: it introduces per-layer configuration support in `configuration_utils` and the heterogeneity helpers, which the larger modeling PR appears to build on.", + "best_issue_reason": null, + "best_pr_reason": "PR #45333 is the better representative of the cluster because it is the core abstraction change and is more self-contained than the downstream modeling integration in #45332.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Same feature area, but not the same concrete change. #45333 adds per-layer config infrastructure; #45332 adds model/cache/masking behavior that uses that infrastructure. They could be merged separately and are not duplicate fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "602bcd3606ceb09dbb5cb2e2b280822fff4853137cbe6131249cae39a9c89d50", + "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": "These are both typing-related PRs and share a few files, but they address different problems: #45415 is a broad repository-wide type-checking cleanup, while #45425 adds concrete typing protocols/aliases to support specific integrations and model utility code. They should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45425 is the better canonical PR for this cluster because it has a focused, concrete typing change set centered on `_typing.py`, `modeling_utils.py`, and PEFT-related integration code, rather than a broad sweep of type-checking adjustments.", + "best_issue_reason": null, + "best_pr_reason": "PR #45425 is the best representative because it targets a specific compatibility/typing gap with clear implementation scope, whereas #45415 is a much wider type-checking pass across many files and is not the same underlying change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Shared filenames and general typing theme are not enough: #45415 is a broad `src/transformers/*py` type-checking enablement pass, while #45425 introduces specific typing aliases/protocols for modeling_utils/PEFT integration. They fix different code-paths and are not plausibly one merged PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45513-2", + "nodes": [ + "pull_request:45513", + "pull_request:46084" + ], + "soft_pairs": [ + "pull_request:45513|pull_request:46084" + ], + "prepared_review_unit_hash": "437e784edbdaa3f47d75e1c4503c42fbeb34f701c30f420fd7fef323ab63e2d8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4948, + "estimated_input_tokens": 1237, + "estimated_eval_tokens": 2730 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated open PRs that both fix cached-forward bugs, but in different model families and with different code paths: Qwen3.5/OLMo hybrid linear-attention cached multi-token forward vs. Mamba2 use_cache with seq_len>1. They should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46084 is the cleaner canonical PR: it has an explicit issue target, a narrowly scoped fix for a concrete seq_len>1 cache bug, and a focused test update.", + "best_issue_reason": null, + "best_pr_reason": "PR #46084 is the best representative of the cluster because it is the most self-contained and explicitly tied to a tracked bug, whereas #45513 spans multiple model files and a broader model-family-specific fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Different underlying bugs and different model code paths. #45513 targets Qwen3.5/OLMo hybrid linear-attention cached multi-token forward; #46084 targets Mamba2 cache handling when seq_len > 1. Shared theme is too broad to treat as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "68bc090dd8ee2951d321d3065f29a70e5a1db7331c01abb6d201c54e6f5b77a4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests both touch EP-related loading code, but they fix different failure modes: one is about preserving distributed config / enabling EP-aware loading, the other is about preventing expert tensors from being overwritten during FSDP2/rank-0 broadcast. They are related, but not duplicates.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the better canonical candidate because it targets a concrete data-corruption bug in a specific loading path (experts being silently overwritten) and has the more self-contained, user-visible fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the strongest representative of the cluster topic: it fixes a specific EP/FSDP2 loading bug with a clear code-path and direct correctness impact, whereas #45548 is more about compatibility plumbing for accelerate/ZeRO-3.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Related EP loading changes, but they address different bugs and code paths: #45548 preserves distributed config / adds EP awareness, while #45662 prevents expert parameters from being overwritten during sharded loading. They would not plausibly merge as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "333577ab690e2f9b0f5dc24a7ebf48a4aa53109fe7a04d36172550526cc50503", + "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-related captured_info collision in `testing_utils.py`; PR #45645 is the more complete version because it also updates the CI workflow to read multiple per-worker log files.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#45645 is the better canonical PR because it contains the core xdist-safe logging fix plus the downstream workflow change needed to consume the new per-worker artifacts.", + "best_issue_reason": null, + "best_pr_reason": "#45645 is the best representative of the cluster since it fixes the full concrete problem end-to-end, while #45639 only covers the writer side.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: `captured_info.txt` collisions under pytest-xdist. #45645 appears to be a superset of #45639's fix and could 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": 24, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "93f25b81be90a6bb698fcdb5aa39b292cdc2940497bd1de9995c2b93559c47aa", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7248, + "estimated_input_tokens": 1812, + "estimated_eval_tokens": 3880 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7248, + "estimated_input_tokens": 1812, + "estimated_eval_tokens": 3880 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs in the same repository touch some of the same model files, but they implement different linting rules and fix different code patterns. The overlap looks incidental rather than duplicate work.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: #45720 is about TRF003 (return-tuple/forward signature cleanup across model forwards), while #46259 is about TRF018 (model weight initialization refactor). They target different underlying changes.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good duplicate target for the other. If forced to pick the more representative cluster member, #45720 is slightly broader and more central to the shared file overlap, but the pair is not a true duplicate set.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different linter rules and different code paths: TRF003 changes forward return-tuple handling, TRF018 changes initialization logic. Shared model filenames do not indicate the same underlying fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "c6d96aba025fd2fd8f43cfd5e55625391506506e68ae1aabb317289e7449bce6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5200, + "estimated_input_tokens": 1300, + "estimated_eval_tokens": 2856 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5200, + "estimated_input_tokens": 1300, + "estimated_eval_tokens": 2856 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are not duplicates. PR 45932 is a mixed CI/telemetry and model-config change, while PR 46100 is a test migration adding kernel tests and related test infra. They share only broad testing/CI context, not the same concrete bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR as a duplicate pair; the changes target different code paths and different goals.", + "best_issue_reason": null, + "best_pr_reason": "PR 46100 is the better cluster representative because it is a more coherent, self-contained test migration/change set. PR 45932 is mostly unrelated CI/workflow and model configuration churn.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different underlying changes: 45932 updates CI/workflow secrets and several model configs, while 46100 adds kernel tests and test-migration files. They do not appear mergeable into one PR for the same concrete 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": 27, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "5898c557ca3f5acf198a58842cd631e4cf16a9d595966c7679dbf537968c9d7e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5525, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: PR #46126 is a focused EP/Trainer integration change around tensor-parallel router handling, while PR #46269 is a broad distributed branch base touching many unrelated distributed/FSDP areas. They share some filenames, but not the same concrete fix or mergeable code path.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46126 is the more specific, self-contained change focused on EP router/tensor-parallel integration and trainer wiring; it has a clear single-purpose scope.", + "best_issue_reason": null, + "best_pr_reason": "PR #46126 is the best representative because it targets one concrete feature/fix, whereas PR #46269 is a large umbrella branch-base PR with mixed distributed changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Shared files alone are not enough here. PR #46126 is an EP/router-specific integration, while PR #46269 is a broad distributed branch base covering multiple unrelated subsystems (FSDP, CI, integrations, model loading). They do not appear to fix the same underlying bug or change, and would not plausibly merge as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-10", + "nodes": [ + "pull_request:46138", + "pull_request:46432", + "pull_request:46501", + "pull_request:46502", + "pull_request:46503", + "pull_request:46504", + "pull_request:46554", + "pull_request:46558", + "pull_request:46563", + "pull_request:46716" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46554", + "pull_request:46138|pull_request:46558", + "pull_request:46138|pull_request:46563", + "pull_request:46138|pull_request:46716", + "pull_request:46138|pull_request:46504", + "pull_request:46432|pull_request:46503", + "pull_request:46432|pull_request:46501", + "pull_request:46432|pull_request:46502", + "pull_request:46432|pull_request:46554", + "pull_request:46432|pull_request:46558", + "pull_request:46501|pull_request:46503", + "pull_request:46501|pull_request:46502", + "pull_request:46501|pull_request:46504", + "pull_request:46501|pull_request:46554", + "pull_request:46502|pull_request:46503", + "pull_request:46502|pull_request:46504", + "pull_request:46502|pull_request:46554", + "pull_request:46503|pull_request:46504", + "pull_request:46503|pull_request:46554", + "pull_request:46504|pull_request:46554", + "pull_request:46504|pull_request:46558", + "pull_request:46554|pull_request:46558", + "pull_request:46554|pull_request:46563", + "pull_request:46554|pull_request:46716", + "pull_request:46558|pull_request:46563", + "pull_request:46558|pull_request:46716", + "pull_request:46563|pull_request:46716", + "pull_request:46504|pull_request:46563", + "pull_request:46504|pull_request:46716" + ], + "prepared_review_unit_hash": "005b295efe4efd27f43c7a72ac004245220cb69c99a1b2cdf6067be4a00b79c3", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 29, + "serialized_chars": 25928, + "estimated_input_tokens": 6482, + "estimated_eval_tokens": 13220 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 29, + "serialized_chars": 25928, + "estimated_input_tokens": 6482, + "estimated_eval_tokens": 13220 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly unrelated workflow-maintenance PRs that were grouped by title similarity. The only true overlap is that several PRs touch .github/workflows/pr-ci-security-gate.yml, but they make different security-related changes and should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the items are distinct workflow edits, and even the repeated pr-ci-security-gate changes are separate follow-up changes rather than one duplicate fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 46554 is the most substantive representative of the pr-ci-security-gate sequence, but it still is not a duplicate of the later edits.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows: self-comment CI vs PR CI security gate, with unrelated behavior changes." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated changes; only title style is similar." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows and unrelated changes." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows: stdin handling in self-comment CI vs commit-message parsing in build-ci-docker-images." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs benchmark dispatch trigger." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs anti-slop permissions." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflows and unrelated concerns: SSH runner hardening vs reviewer-assignment checkout ref." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated concerns." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated concerns." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows with different purposes: anti-slop permissions vs benchmark dispatch support." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same file, but different concrete security-gate edits: one adds repo checkout and file checks, the other changes permissions/approval flow." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "Same file, but different changes: security checks vs reusable-workflow safety notes." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different changes: security checks vs secrets inheritance tightening." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Same file, but different concrete changes in the security-gate workflow." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different concrete changes in the security-gate workflow." + }, + { + "left": "pull_request:46563", + "right": "pull_request:46716", + "accept": false, + "reason": "Same file, but different concrete changes in the security-gate workflow." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46716", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46152-2", + "nodes": [ + "pull_request:46152", + "pull_request:46207" + ], + "soft_pairs": [ + "pull_request:46152|pull_request:46207" + ], + "prepared_review_unit_hash": "117b492514d80860027b0d8b0ace4c6e0ac2f66988160d944f217880e36fa847", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs share a theme of avoiding brittle source-file reads, but they fix different code paths in different modules: one in `import_utils.py` and one in `modeling_utils.py` with tests. They are not the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46207 is the better representative of the cluster because it is a concrete, self-contained fix with test coverage for a specific source-reading heuristic.", + "best_issue_reason": null, + "best_pr_reason": "PR #46207 is the strongest standalone candidate: it targets a precise bug in `_can_set_attn_implementation` and includes validation tests.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Same broad idea of avoiding `open()` for source inspection, but they touch different functions and modules with different failure modes; not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 30, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46154-2", + "nodes": [ + "pull_request:46154", + "pull_request:46196" + ], + "soft_pairs": [ + "pull_request:46154|pull_request:46196" + ], + "prepared_review_unit_hash": "2cfc50d71a1d48c5db64dd8806aa030e0713b8b8344b44695b6f9999c163f511", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5700, + "estimated_input_tokens": 1425, + "estimated_eval_tokens": 3106 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5700, + "estimated_input_tokens": 1425, + "estimated_eval_tokens": 3106 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs add SDPA attention dispatch support, but they target different model families and code paths (T5 vs wav2vec2/audio models). They are related in theme, not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "Same broad goal, but different architectures and implementation paths. These would not plausibly be merged into one PR as a duplicate 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": 31, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46173-2", + "nodes": [ + "pull_request:46173", + "pull_request:46474" + ], + "soft_pairs": [ + "pull_request:46173|pull_request:46474" + ], + "prepared_review_unit_hash": "8ac88de891156efc041ae6eac2ff2b6f957f37c7ee42f44803b224f91f9b44ec", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3631, + "estimated_input_tokens": 908, + "estimated_eval_tokens": 2072 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3631, + "estimated_input_tokens": 908, + "estimated_eval_tokens": 2072 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests make the same code change in `src/transformers/cache_utils.py` to widen `StaticLayer.get_seq_length` from `int` to `int | torch.Tensor`; they are near-duplicate implementations of the same fix for issue #45987.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46173 is the better canonical representative because it is the earlier, more explicitly titled version of the same fix and already points to the tracking issue #45987.", + "best_issue_reason": null, + "best_pr_reason": "PR #46173 is the strongest standalone representative of the change: it contains the same minimal patch, the same target issue, and a clearer description of the intended fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46173", + "right": "pull_request:46474", + "accept": true, + "reason": "Same underlying bug fix, same file, same diff hunk, and same explicit issue target (#45987); these are functionally duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs touch the same file, make the same type-annotation change for `StaticLayer.get_seq_length`, and both explicitly target issue #45987. The canonical-pr rationale is also conservative and supported by the titles/metadata. No overstatement beyond the visible diff evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 32, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46310-2", + "nodes": [ + "pull_request:46310", + "pull_request:46546" + ], + "soft_pairs": [ + "pull_request:46310|pull_request:46546" + ], + "prepared_review_unit_hash": "482a13eb32ea69b10273afaa02091c07a9fec2845d3d4e4c0b29a4deef73f98f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4850, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4850, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs fix the same sliding-window right-boundary off-by-one bug (`<` to `<=`) in Moonshine streaming; PR 46546 is the broader, more complete version and also patches Gemma4 plus tests.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46546 is the best canonical PR because it contains the shared Moonshine streaming fix, extends the same bug fix to Gemma4, and includes test coverage, making it the more complete representative of this change.", + "best_issue_reason": null, + "best_pr_reason": "PR 46546 is the best representative for the cluster since it addresses the same concrete boundary bug while covering more affected code paths and adding tests.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46310", + "right": "pull_request:46546", + "accept": true, + "reason": "Same underlying off-by-one boundary bug in the sliding-window mask; the Moonshine change is identical, and 46546 simply broadens the fix to another model plus tests, 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": 33, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "325f8cb6ab627d525a6b73e291606663699f044fef36ac74f94c9fb7c510404e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5939, + "estimated_input_tokens": 1485, + "estimated_eval_tokens": 3226 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5939, + "estimated_input_tokens": 1485, + "estimated_eval_tokens": 3226 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are unrelated: one adds multimodal support to continuous batching internals, while the other introduces the HunYuan VL model package, docs, and tests. They do not appear to be duplicates or mergeable into a single PR.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46417 is the more self-contained and clearly scoped change set, with a complete model addition across code, docs, and tests. PR 46376 is a separate draft infrastructure refactor for continuous batching.", + "best_issue_reason": null, + "best_pr_reason": "PR 46417 is the strongest standalone representative of the cluster because it is a complete feature addition. PR 46376 is a different subsystem change and not a duplicate of it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different underlying changes and code paths: continuous batching multimodal cache support versus adding a new HunYuan VL model. No shared filenames, targets, or concrete bug/fix overlap." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 34, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46574-2", + "nodes": [ + "pull_request:46574", + "pull_request:46580" + ], + "soft_pairs": [ + "pull_request:46574|pull_request:46580" + ], + "prepared_review_unit_hash": "f4d6feade69219ea43c36ad1c6db0333769718b98c3bebc8f2a5ddfe064885cf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4410, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated tokenizer fixes: one for CodeLlama decode stripping, one for DeepSeek checkpoint/tokenizer-class resolution. They share a broad theme but not the same underlying bug or code path.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46574 is the narrower, cleaner bug fix centered on a concrete decode regression in CodeLlama and includes a targeted tokenizer change plus tests.", + "best_issue_reason": null, + "best_pr_reason": "PR #46574 is the stronger representative because it fixes a single, well-scoped runtime bug with a direct code-path change; PR #46580 is a separate auto-tokenizer mapping issue.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46574", + "right": "pull_request:46580", + "accept": false, + "reason": "Different problems and code paths: CodeLlama decoder initialization/strip behavior vs DeepSeek checkpoint-prefix tokenizer-class override. They would not plausibly be merged into one PR as a single 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": 35, + "component_count": 35, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46603-2", + "nodes": [ + "pull_request:46603", + "pull_request:46604" + ], + "soft_pairs": [ + "pull_request:46603|pull_request:46604" + ], + "prepared_review_unit_hash": "35bad15ba8f2447d982fa3aeb685899268b0425ec47a56caa7e93e2e96186940", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5495, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5495, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are both Mistral-related PRs, but they address different changes: one is a package refactor/move of MistralConverter, the other adds tekken tokenizer detection/conversion/save support. They are related but not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46604 is the better canonical PR because it contains the more substantive user-facing tokenizer functionality; #46603 is mostly a code-organization move of MistralConverter.", + "best_issue_reason": null, + "best_pr_reason": "PR #46604 is the best representative of the cluster since it introduces the main behavioral change, while #46603 is a narrower refactor in the same area.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46603", + "right": "pull_request:46604", + "accept": false, + "reason": "Different concrete changes: #46603 moves MistralConverter into a new package location, while #46604 adds tekken tokenizer detection/conversion/save utilities. They touch similar files but not the same underlying fix 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 + } + ] +} diff --git a/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/manifest.json b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..84168d9e29db51b9d0e8d72c942bc7435c31e411 --- /dev/null +++ b/snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260618t000034z", + "artifacts": { + "hybrid": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-18T00:09:37Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260618T000034Z", + "variant": "hybrid" +} diff --git a/snapshots/20260618T000034Z/manifest.json b/snapshots/20260618T000034Z/manifest.json index d57f1fe89370f3ee581e62d75c5327b93bf3c3d1..7280aab81562e0fa9c05934629228a71cd611d4f 100644 --- a/snapshots/20260618T000034Z/manifest.json +++ b/snapshots/20260618T000034Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 619 }, "extracted_at": "2026-06-18T00:00:34Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260618t000034z", + "runs": { + "hybrid-model-20260618t000034z": { + "analysis_id": "hybrid-model-20260618t000034z", + "artifacts": { + "hybrid": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260618T000034Z/analysis-runs/hybrid-model-20260618t000034z/manifest.json", + "model": null, + "published_at": "2026-06-18T00:09:37Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260618T000034Z", "watermark": {