diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 3757254980fe0138f16e964414f45df6bb017422..f262cbad1c48a8995074e07308f6880f8463464d 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": "20260528T060028Z", - "generated_at": "2026-05-28T06:05:22Z", + "snapshot_id": "20260528T120029Z", + "generated_at": "2026-05-28T12:12:30Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1148,100 +1148,6 @@ } ] }, - { - "cluster_id": "cluster-46032-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #46032.", - "status": "open", - "confidence": 0.75, - "canonical_issue_number": 46032, - "canonical_pr_number": 46084, - "issue_numbers": [ - 46032 - ], - "pr_numbers": [ - 46084, - 46098 - ], - "evidence_types": [ - "closing_reference", - "shared_issue_target" - ], - "pr_comparisons": [ - { - "left_pr_number": 46084, - "right_pr_number": 46098, - "code_similarity": 0.43, - "size_similarity": 0.5, - "file_overlap": 0.667, - "area_overlap": 0.226, - "patch_similarity": 0.339, - "shared_filenames": [ - "src/transformers/models/mamba2/modeling_mamba2.py", - "tests/models/mamba2/test_modeling_mamba2.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/models/mamba2/modeling_mamba2.py", - "left_ranges": [ - [ - 256, - 262 - ], - [ - 303, - 353 - ], - [ - 461, - 467 - ], - [ - 471, - 491 - ], - [ - 496, - 502 - ], - [ - 599, - 609 - ] - ], - "right_ranges": [ - [ - 255, - 373 - ], - [ - 481, - 495 - ] - ] - }, - { - "filename": "tests/models/mamba2/test_modeling_mamba2.py", - "left_ranges": [ - [ - 206, - 236 - ], - [ - 285, - 322 - ] - ], - "right_ranges": [ - [ - 476, - 542 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -2122,15 +2028,6 @@ "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-46032-3", - "canonical_pr_number": 46084, - "duplicate_pr_numbers": [ - 46098 - ], - "target_issue_number": 46032, - "reason": "PRs in cluster-46032-3 are treated as duplicates because they converge on issue #46032 with closing_reference, shared_issue_target evidence." - }, { "cluster_id": "cluster-46100-2", "canonical_pr_number": 46100, diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 56f3cae39a990d282262fc37efd3023aa6552393..4dd09f72f8cc54330f81320023ba7a06b65f3465 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": "20260528T060028Z", - "generated_at": "2026-05-28T06:05:22Z", + "snapshot_id": "20260528T120029Z", + "generated_at": "2026-05-28T12:12:30Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -13,8 +13,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -138,370 +138,550 @@ "issue:43866|issue:44863", "issue:44292|issue:46015" ], - "prepared_review_unit_hash": "2af61c8f01717da63b8f1b78a3ca1aaa3fba9ad63163c190d1fa8ecd6a2db575", + "prepared_review_unit_hash": "f270fcaf8003e02597617bfd45fb5a51b4e0d409170c3bafa404b2b6be4686b6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 69, - "serialized_chars": 48405, - "estimated_input_tokens": 12102, - "estimated_eval_tokens": 24460 + "serialized_chars": 48409, + "estimated_input_tokens": 12103, + "estimated_eval_tokens": 24462 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated transformers issues. The only clear duplicate families are the Bert/ZeRO-3 index-error reports, the v5 AutoTokenizer token-ID regression reports, and a likely shared Qwen3_5Moe rotary-embedding offload/materialization bug.", - "confidence": 0.69, - "canonical_issue_reason": "issue:45920 is the broadest and most representative report of the v5 AutoTokenizer token-ID regression, subsuming the narrower Granite-only report in the accepted pair.", + "summary": "The cluster is mostly heterogeneous and not a single duplicate set. The only clear duplicate families are the AutoTokenizer wrong-token-ID v5 regression and the near-identical DeepSpeed ZeRO-3 BERT index-out-of-bounds reports.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 45920 is the broadest and most representative active report of the AutoTokenizer token-ID regression affecting multiple model families.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 best captures the broader tokenizer-regression family and is the strongest single issue to anchor that duplicate set.", + "best_issue_reason": "Issue 45920 is the best global issue anchor because it captures the widest version of the recurring tokenizer regression and is still open for follow-up.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different models and failure modes: Qwen3.5 config attribute error vs Qwen3.5 streaming/chat completions failure." + "reason": "Different Qwen3.5 streaming/API failure vs config attribute error; not the same bug." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Unrelated hardware/backend problems: AMD GPU support for GPT-OSS vs an MPS histogram op limitation." + "reason": "Different model and hardware paths: AMD GPU failure vs MPS histogram/int error." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different bug class and subsystem: image-preprocessor/model dtype mismatch vs GPT-OSS on AMD GPUs." + "reason": "Different multimodal dtype mismatch vs GPT-OSS AMD GPU failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both mention runtime issues, but these are different models and different dtype/device paths." + "reason": "Both are dtype-related at a high level, but different models and code paths." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model families and symptoms: GPT-OSS on AMD GPUs vs GraniteMoEHybrid invalid method call." + "reason": "Different model failures; not the same concrete method/code-path bug." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/backend issues: Qwen-3-8B-NVFP4 runtime failure vs GPT-OSS on AMD GPUs." + "reason": "Different Qwen NVFP4 runtime issue vs GPT-OSS AMD GPU failure." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "Different concrete failures: DeepSpeed Zero3 collective mismatch vs ModernBERT ONNX export error." + "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Unrelated bugs: ONNX export for ModernBERT vs FineGrainedFP8 failure for Qwen3Moe." + "reason": "Export-to-ONNX issue vs FP8 config failure; unrelated bugs." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different subsystems and error signatures: ONNX export vs auto device mapping cudaErrorAssert." + "reason": "ONNX export and CUDA assert during auto device map are unrelated." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "A task crash report is not the same as a docs example cleanup issue." + "reason": "Table QA crash vs outdated RAG docs; different problem types." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions from v5 removals, but they target different removed APIs/pages and are not the same change." + "reason": "Both are docs cleanup for v5 removals, but they refer to different removed items and not one concrete duplicate change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "Different code paths entirely: ONNX export error vs import path breakage for GenerationMixin." + "reason": "ONNX export bug vs GenerationMixin import breakage; unrelated." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Both are shape/device-related errors, but they involve different models and different failure points." + "reason": "Different failures: mask shape mismatch vs Gemma4 device mismatch." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT embedding unpacking and resume-with-trust-remote-code are distinct training issues." + "reason": "TRL SFT embedding unpacking vs resume/trust-remote-code training issue; unrelated." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are IndexError/shape issues, but the reported tensors, triggers, and code paths differ." + "reason": "Both are runtime IndexErrors, but different models and failure mechanisms." }, { - "left": "issue:43638", - "right": "issue:46165", + "left": "issue:43596", + "right": "issue:44805", "accept": false, - "reason": "Different DeepSpeed/Bert loading bug vs training resume/trust_remote_code problem." + "reason": "Different zero3/BERT init failure vs mask shape mismatch." }, { "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "BERT zero3 init bug vs training-resume/trust-remote-code issue; unrelated." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration in Qwen3-vl embedding vs training resume issue; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "StableLmConfig padding regression vs training resume issue; unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor preset bug vs ModernBERT ONNX export bug; unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but one is collective mismatch and the other is zero3 index error; different bugs." + }, + { + "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Similar surface-level IndexError, but the zero3/Bert load path is different from the mask-shape mismatch report." + "reason": "FSDP checkpointing recompute mismatch vs mask shape mismatch; unrelated." }, { - "left": "issue:44805", - "right": "issue:45161", + "left": "issue:45859", + "right": "issue:45902", "accept": false, - "reason": "Different problem domains: mask/tensor shape mismatch vs GPT-OSS tensor-parallel failure." + "reason": "Same model family, but CPU-offload incompatibility and uninitialized-memory materialization are different code-path problems." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy call issue is unrelated to GPT-OSS tensor-parallel behavior." + "reason": "Liger cross_entropy call vs GPT-OSS tensor-parallel issue; unrelated." }, { - "left": "issue:44805", + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Different model/config errors: Gemma3TextConfig attribute vs Mistral tokenizer patch crash." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Both Gemma4-related, but device mismatch and expensive reverse-embedding lookup are different issues." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility vs Qwen NVFP4 runtime error; unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Different dtype/autocast issue vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL embedding unpacking vs GPT-OSS TP bug; unrelated." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Both are runtime failures, but not the same model or code path." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different multimodal dtype issue vs MPS histogram int issue." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different models and failure modes; not duplicates." + }, + { + "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Different failures in different training stacks; no clear shared underlying bug." + "reason": "Different model families and errors." }, { - "left": "issue:43572", - "right": "issue:46165", + "left": "issue:45507", + "right": "issue:45685", "accept": false, - "reason": "StableLmConfig pad token regression is unrelated to training resume and trust_remote_code." + "reason": "GraniteMoEHybrid invalid method vs MPS histogram int error; unrelated." }, { - "left": "issue:34634", - "right": "issue:35545", + "left": "issue:43039", + "right": "issue:44918", "accept": false, - "reason": "BarkProcessor voice_preset bug and ModernBERT ONNX export error are unrelated." + "reason": "Liger kernel cross_entropy issue vs TRL embedding unpacking; unrelated." }, { - "left": "issue:30277", + "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is a zero3 BERT index error." + "reason": "Both are DeepSpeed ZeRO-3 related, but the models and concrete failures differ." }, { - "left": "issue:43596", - "right": "issue:43638", - "accept": true, - "reason": "Same underlying DeepSpeed ZeRO-3/Bert loading bug with the same zero-sized tensor IndexError; one is just the non-pretrained variant." + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Different docs gaps about different removed v5 features." }, { - "left": "issue:43975", - "right": "issue:44457", + "left": "issue:43716", + "right": "issue:44560", "accept": false, - "reason": "Different failures: DeepSeek detokenization regression vs LoRA merge/save/load output mismatch." + "reason": "Different model-specific issues; not the same bug." }, { - "left": "issue:44292", - "right": "issue:44457", + "left": "issue:43828", + "right": "issue:45685", "accept": false, - "reason": "Qwen NVFP4 runtime error and LoRA merge inconsistency are unrelated." + "reason": "Dtype/autocast issue vs MPS histogram int issue; unrelated." }, { - "left": "issue:24643", - "right": "issue:30277", + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL embedding unpacking vs MPS histogram int issue; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different subsystems and failures; not a duplicate." + }, + { + "left": "issue:30277", + "right": "issue:36331", "accept": false, - "reason": "Both mention DeepSpeed, but the concrete errors and affected paths are different." + "reason": "DeepSpeed collective mismatch vs trainer signature change; unrelated." }, { "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP mismatch vs GenerationMixin import error; unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU model failure vs docs example drift; unrelated." + }, + { + "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP recompute mismatch is a different bug from the mask/index shape mismatch." + "reason": "Liger kernel cross_entropy call vs tensor mask shape mismatch; unrelated." }, { - "left": "issue:45859", - "right": "issue:45902", + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different model/runtime errors; not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 loading issues, but Gemma3 and Bert failures are different concrete bugs." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "GPT-OSS AMD issue vs table QA crash; unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs problems, but they cover different pipeline/API removals." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral vs Gemma4 device mismatch; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs Qwen3-vl video StopIteration; unrelated." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization regression vs docs example drift; unrelated." + }, + { + "left": "issue:45812", + "right": "issue:45920", "accept": true, - "reason": "Same Qwen3_5MoeTextRotaryEmbedding offload/materialization path; both point to the same rotary-embedding state/CPU-offload bug." + "reason": "Same AutoTokenizer wrong-token-ID v5 regression across model families; 45920 is the broader report." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading bug vs outdated examples docs issue; unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint report vs model implementation load bug; not the same issue." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen NVFP4 runtime error vs docs issue; unrelated." }, { "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed/BERT zero3 failure vs training resume/trust-remote-code issue; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 + Bert index-0 out-of-bounds failure, just described with slightly different model wording." + }, + { + "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "DeepSpeed/Bert zero3 loading issue and GPT-OSS tensor-parallel issue are unrelated." + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." }, { - "left": "issue:34928", - "right": "issue:36331", + "left": "issue:43716", + "right": "issue:45507", "accept": false, - "reason": "Activation-checkpointing tensor recomputation mismatch is unrelated to CustomTrainer.compute_loss signature breakage." + "reason": "Different model-specific failures and different code paths." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs GPT-OSS TP issue; unrelated." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different models and failure modes: DeepSpeed/Bert zero3 IndexError vs GPT-OSS tensor-parallel issue." + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Mistral dtype mismatch and GraniteMoEHybrid invalid method call do not share the same underlying bug." + "reason": "Different model-specific failures and different code paths." }, { - "left": "issue:39290", - "right": "issue:45081", + "left": "issue:43039", + "right": "issue:44918", "accept": false, - "reason": "Different components and symptoms: Gemma3 config attribute issue vs Mistral tokenizer regex patch crash." + "reason": "Different subsystems: Liger kernel loss call vs TRL embedding unpacking." }, { - "left": "issue:45823", - "right": "issue:45874", + "left": "issue:41762", + "right": "issue:43638", "accept": false, - "reason": "Both are Gemma4-related, but one is device-mapping forward mismatch and the other is an expensive reverse embedding lookup." + "reason": "Different ZeRO-3 failure modes and different models." }, { - "left": "issue:43425", - "right": "issue:44292", + "left": "issue:43827", + "right": "issue:44512", "accept": false, - "reason": "Torch version incompatibility and Qwen NVFP4 runtime failure are different issues." + "reason": "Different doc gaps; not a single duplicate change." }, { - "left": "issue:43828", - "right": "issue:45161", + "left": "issue:43716", + "right": "issue:44560", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS tensor-parallel failures." + "reason": "Different multimodal issues; not duplicates." }, { - "left": "issue:44918", - "right": "issue:45161", + "left": "issue:43828", + "right": "issue:45685", "accept": false, - "reason": "TRL SFT embedding unpacking and GPT-OSS tensor-parallel problems are not the same bug." + "reason": "Different runtime errors and different models." }, { - "left": "issue:43828", - "right": "issue:44292", + "left": "issue:44918", + "right": "issue:45685", "accept": false, - "reason": "Different models and different runtime errors; no common code path indicated." + "reason": "Different failures: TRL SFT embedding unpacking vs MPS histogram issue." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Liger Kernel cross-entropy call issue is unrelated to Phi-tiny-MoE autocast dtype mismatch." + "reason": "Different subsystems and symptoms; not a duplicate." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and CustomTrainer keyword-argument breakage are unrelated." + "reason": "DeepSpeed collective mismatch vs trainer signature break; unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Activation-checkpointing/FSDP mismatch is unrelated to GenerationMixin import failure." + "reason": "FSDP checkpointing mismatch vs GenerationMixin import error; unrelated." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "GPT-OSS on AMD GPUs is not the same as an outdated RAG example docs issue." + "reason": "Model/runtime failure vs docs example drift; unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Cross-entropy/Liger Kernel issue and mask-shape mismatch are unrelated." + "reason": "Liger kernel cross_entropy issue vs shape mismatch error; unrelated." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 vLLM config regression and Qwen3 auto-device-mapping cuda assert are different bugs." + "reason": "Different model/runtime issues; not the same underlying bug." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Gemma3 zero3 loading IndexError and BertModel zero3 loading IndexError are similar in shape but not clearly the same bug." + "reason": "Gemma3 ZeRO-3 loading issue vs BERT ZeRO-3 loading issue; similar area, different bug." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "GPT-OSS AMD GPU support and table-question-answering crashes are unrelated." + "reason": "GPT-OSS AMD failure vs table QA crash; unrelated." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "TextClassificationPipeline docs mismatch is unrelated to the removed transformers run command docs issue." + "reason": "Different documentation updates for different removed APIs." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral and Gemma4 PLE device mismatch are different model-specific problems." + "reason": "Processor failure vs Gemma4 device mismatch; unrelated." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross-entropy invocation and Qwen3-vl video StopIteration are unrelated." + "reason": "Liger kernel issue vs Qwen3-vl video StopIteration; unrelated." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "DeepSeek detokenization regression and outdated RAG examples are not the same underlying change." - }, - { - "left": "issue:45812", - "right": "issue:45920", - "accept": true, - "reason": "Same v5 AutoTokenizer token-ID regression; 45920 is the broader report and 45812 is the narrower Granite-only instance." + "reason": "Detokenization regression vs docs drift; unrelated." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading failure is unrelated to RAG documentation examples." + "reason": "Checkpoint loading bug vs docs issue; unrelated." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Ovis2 checkpoint corruption and NemotronH checkpoint loading are distinct model issues." + "reason": "Corrupted checkpoint report vs model load bug; not the same issue." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen NVFP4 runtime failure and outdated RAG docs are unrelated." + "reason": "Model runtime error vs docs issue; unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The accepted pairs are supported by the titles and the soft-pair evidence, and the rejected pairs avoid overclaiming shared root cause where the issues only look superficially similar. No soft-edge verdict appears overstated." + "feedback": "The summary is grounded in the packet: most items are clearly heterogeneous, and the two accepted duplicate families are well supported by the titles. The soft-edge rejections stay conservative, and the accepted pairs (45812/45920 and 43596/43638) look justified from the available evidence." }, "evaluator_used": true, "retried": false, @@ -517,8 +697,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -612,268 +792,268 @@ "issue:43404|issue:45005", "issue:40990|issue:44960" ], - "prepared_review_unit_hash": "49a695780a3e08cbeb0e7aeaaaff8d95ed415dd1e881b6bb73351685fb0f6997", + "prepared_review_unit_hash": "d83af649a88d0d52de6bb2587079d09172d244da442a6ae3b6ec33f4cbb98687", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 39, - "serialized_chars": 39132, - "estimated_input_tokens": 9783, - "estimated_eval_tokens": 19822 + "serialized_chars": 39127, + "estimated_input_tokens": 9782, + "estimated_eval_tokens": 19820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue reports across different models, loaders, and runtime features; there is no reliable duplicate cluster. The only broad theme is recent Transformers compatibility regressions, but the soft pairs are still too symptom- and model-specific to merge.", - "confidence": 0.91, - "canonical_issue_reason": null, + "summary": "This cluster is mostly a grab bag of unrelated closed issues. Several pairs are superficially similar around model loading, zero-3, tied weights, and custom-module handling, but they do not look like the same concrete bug/change, so I would not merge any of the soft edges.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45020 is the broadest and most umbrella-like report in the set: it captures a general regression class (\u2018recent transformers versions break models using remote_code\u2019), which makes it the best candidate to anchor this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "#45020 is the broadest regression report in the set (recent versions breaking `remote_code` models), so it is the closest representative issue, but it does not truly subsume the others.", + "best_issue_reason": "Issue 45020 is the most reusable triage target because it is broad, regression-oriented, and likely to subsume multiple adjacent compatibility complaints better than the model-specific tickets.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Different models and different config attributes; both are sliding-window related but not the same bug." + "reason": "Both are model-specific config issues, but they involve different models and different attributes; too broad to be the same bug." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper runtime failure vs table-question-answering crash; unrelated code paths." + "reason": "Whisper runtime failure vs table-question-answering crash; different feature areas and code paths." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Different model families with different loading failures; no shared concrete defect." + "reason": "One is a GLM-4.7-Flash load/test issue, the other a corrupted Ovis2 checkpoint; unrelated failures." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are tensor index/shape errors, but the reproductions and affected paths differ." + "reason": "Both mention deepspeed zero-3 and out-of-bounds errors, but one is a mask-shape mismatch while the other is a Bert init/indexing failure." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "One is a Torch version compatibility report; the other is a PretrainedConfig/pydantic regression." + "reason": "Torch-version incompatibility is not the same bug as the pydantic PretrainedConfig regression." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "DeepSpeed ZeRO-3/Bert loading issue vs a Torch/Transformers version incompatibility complaint." + "reason": "Both are compatibility/runtime complaints, but one is a zero-3 BertModel init problem and the other is a torch/transformers version mismatch." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Detokenization bug and table QA crash are different failures." + "reason": "Incorrect detokenization and table QA crashing are different symptoms and code paths." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper execution failure is not the same as the PretrainedConfig regression." + "reason": "Whisper runtime failure does not match the PretrainedConfig field regression." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5 chat, but the symptoms and API paths differ enough to treat as separate bugs." + "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming API failure; not the same concrete bug." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "`init_empty_weights` argument handling is unrelated to table-question-answering crashes." + "reason": "Different models and unrelated failure modes." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Different models and different failure modes; not the same sliding-window/device bug." + "reason": "Different model families and different underlying issues (config attribute vs device mismatch)." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Chat streaming failure is unrelated to an MPS MoE histogram kernel issue." + "reason": "Streaming chat API failure and an MPS histogram kernel issue are unrelated." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Tied-weight bug vs corrupted checkpoint report; not the same underlying problem." + "reason": "Weight-tying bug vs corrupted checkpoint; not the same problem." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Different MoE/runtime failures on different backends; no common concrete code-path." + "reason": "Grouped-mm tracing failure and an MPS int kernel limitation are different code paths." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both are dtype/runtime issues, but on different models and execution contexts." + "reason": "Both mention dtype/precision, but they are different models and different execution contexts." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor/model-loading failure is not the same as the template-node compilation error." + "reason": "Processor loading failure and template-node compilation error are unrelated." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for `GenerationMixin` and a Gemma sliding-window attribute issue are unrelated." + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding-window config errors." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/Transformers incompatibility vs Gemma4 device-map mismatch; different bugs." + "reason": "A package-version incompatibility is not the same as a Gemma4 device-map mismatch." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are indexing errors, but one is a mask-length mismatch and the other is a zero-sized tensor on load." + "reason": "Both are zero-3 related index errors, but they affect different models and appear to stem from different failure conditions." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing `pad_token_idx` in a config class is different from the pydantic `PretrainedConfig` regression." + "reason": "Missing pad_token_idx after v5 update is a different regression than the PretrainedConfig pydantic issue." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are separate model/support requests." + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are distinct requests/issues." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Different MoE issues; one is a grouped_mm tracing error, the other is tensor-parallel support." + "reason": "Torch-dynamo grouped_mm tracing failure is not the same as GPT-OSS tensor-parallel support." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Quantization memory/OOM regression is unrelated to tied-weights translation-model behavior." + "reason": "Int4 OOM regression and tied-weights translation-model bug are different code paths." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch and chat streaming failure are different runtime problems." + "reason": "Autocast dtype mismatch and streaming chat API failure are unrelated." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "`init_empty_weights` argument mismatch is not the same as the `PretrainedConfig` field breakage." + "reason": "init_empty_weights argument handling is separate from the PretrainedConfig field regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF model support and Qwen chat streaming are unrelated features." + "reason": "GPT-OSS GGUF support and Qwen3.5 streaming chat failure are unrelated." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code model breakage and Torch 2.4.1 incompatibility are different regression classes." + "reason": "Broad remote_code breakage and a specific torch/transformers version mismatch are different issue classes." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Sliding-window model config bug vs pydantic config field breakage; no shared code path." + "reason": "Qwen3-MoE sliding_window behavior is not the same as the pydantic PretrainedConfig issue." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Grouped_mm tracing error and Float8 storage lookup failure are distinct backend issues." + "reason": "Torch-dynamo grouped_mm tracing and missing Float8 storage are unrelated runtime issues." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index errors, but the affected model/load sequence differs." + "reason": "Different models and different zero-3 error shapes; not one underlying bug." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Different SigLIP2 reports: degraded embedding quality vs nonsensical pipeline output." + "reason": "Both concern siglip2 quality/behavior, but one is an embedding-quality complaint and the other is an AutoModel/pipeline result bug." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Composite-config `from_config` behavior and rotary embedding uninitialized memory are unrelated." + "reason": "Composite-config unwrapping and uninitialized inv_freq memory are distinct bugs." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorBoard/TensorFlow backend detection issue is unrelated to Torch compatibility." + "reason": "TensorBoard-without-TF backend error is unrelated to torch/transformers incompatibility." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Gemma3 image preprocessor dtype mismatch is not the same as Qwen3.5 embedding unpacking failure." + "reason": "Gemma/Mistral image-text dtype mismatch and Qwen3.5 TRL embedding unpacking are different failures." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape error and Qwen3.5 embedding unpacking are different problems." + "reason": "FSDP weight-shape error and Qwen3.5 embedding unpacking are unrelated." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to flash-attn fallback logic." + "reason": "Notebook custom-model initialization and flash-attn fallback gating are separate problems." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both involve custom modules, but one is a cache-path collision and the other is wrong-module loading after save/load." + "reason": "Both touch custom-module loading, but one is a leaf-directory cache collision and the other is wrong-module selection after save_pretrained; not the same concrete bug." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but the models and failure conditions are different." + "reason": "Both mention tied weights, but they concern different model classes and likely different fixes; too broad to merge." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Perplexity regression on gpt-oss and a generic GLM5 issue are not the same bug." + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are unrelated." } ] }, @@ -892,8 +1072,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -982,247 +1162,247 @@ "issue:43881|issue:43883", "issue:45685|issue:45823" ], - "prepared_review_unit_hash": "ded2abc075621133f6c2e428a9e724c2d27fb5542b688dcd88504343d4adb34f", + "prepared_review_unit_hash": "27b865c65caf95e645b0382ae9b8d7e29d5a98bc5acc177f067003a59d29a26b", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37572, - "estimated_input_tokens": 9393, - "estimated_eval_tokens": 19042 + "serialized_chars": 37581, + "estimated_input_tokens": 9396, + "estimated_eval_tokens": 19048 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated model/loading/quantization issues. The clearest duplicate pair in this packet is the Transformers 5.x non-persistent-buffer regression: `issue:43950` and `issue:44534`.", - "confidence": 0.82, - "canonical_issue_reason": "`issue:43950` is the broadest and most explicit report of the non-persistent-buffer corruption regression, and it matches the strongest duplicate candidates in this packet.", + "summary": "This cluster is mostly heterogeneous. The only clear duplicate pair is the Transformers v5 non-persistent buffer regression; the rest are related-at-a-distance model/loading/quantization reports but not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest, most detailed report of the non-persistent-buffer regression, and issue:44534 is essentially the same bug described more generally.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:43950` is the best representative because it states the concrete regression clearly and at the right abstraction level.", + "best_issue_reason": "issue:43950 best represents the one true duplicate pair in this cluster and has the strongest problem description and reproduction context.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Different failure modes: FP8 config error vs fallback gating when flash-attn2 is absent." + "reason": "Different problems: FP8 config failure in Qwen3MoE vs flash-attn2 fallback being blocked by checks." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "Different problems: Qwen3.5 cache crash vs tied-weights handling on translation models." + "reason": "PagedAttentionCache crash on Qwen3.5 is unrelated to tied-weights handling in translation models." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to the tied-weights regression." + "reason": "Quantization/offloading behavior is a different code path from v5 tied-weights issues." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "DSA ReLU discussion is unrelated to a docs typo about the removed run command." + "reason": "A DSA indexer ReLU discussion is unrelated to docs mentioning a removed CLI command." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is not the same as the v5 tied-weights issue." + "reason": "NemotronH checkpoint loading failure is not the same as translation-model tied weights." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and code paths: TRL SFT unpacking vs GraniteMoE invalid method." + "reason": "Qwen3.5 embedding unpacking in TRL is a different failure than GraniteMoEHybrid calling an invalid method." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different models and symptoms; one is detokenization, the other is an NVFP4 runtime error." + "reason": "Incorrect detokenization and a Qwen-3 NVFP4 runtime error are distinct bugs." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A cache-strategy typo is unrelated to docs still mentioning the removed run command." + "reason": "A cache-strategy typo is unrelated to a removed docs command." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Separate model/runtime errors with no shared underlying bug shown." + "reason": "NVFP4 execution failure and GraniteMoEHybrid invalid method are separate issues." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying Transformers 5.x non-persistent-buffer corruption/junk-initialization bug." + "reason": "Both report the same v5 regression where non-persistent buffers registered with persistent=False get corrupted/junk values on load." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "SigLIP2 docs example errors are not the same as the reported embedding-quality discrepancy." + "reason": "Docs example errors for SigLIP2 are not the same as a broader text-embedding quality complaint." }, { - "left": "issue:43644", + "left": "issue:43643", "right": "issue:43950", - "accept": true, - "reason": "Both reports describe the same non-persistent-buffer corruption/junk behavior in Transformers 5.x." + "accept": false, + "reason": "Remote-code config field loss is a different bug from non-persistent buffer corruption." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different issues: autocast dtype mismatch vs RLHF gradient explosion from 3D position_ids/backend fallback." + "reason": "Both mention Qwen3.5, but one is a dtype mismatch under autocast and the other is an RLHF NaN/SDPA fallback issue." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Trainer API kwarg mismatch is unrelated to Gemma3 loading under ZeRO-3." + "reason": "CustomTrainer compute_loss API mismatch is unrelated to Gemma3 loading with ZeRO-3." }, { "left": "issue:41628", "right": "issue:45800", "accept": false, - "reason": "Import/export surface issue is unrelated to torch/transformers version incompatibility." + "reason": "Missing AutoImageProcessor import and torch/transformers version incompatibility are not the same bug." }, { "left": "issue:43950", "right": "issue:45902", "accept": false, - "reason": "Both involve buffers, but one is a broad from_pretrained regression and the other is a model-specific materialization bug." + "reason": "Both involve buffers/materialization, but one is a from_pretrained regression and the other is a Qwen3_5Moe rotary embedding uninitialized-memory bug." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different bitsandbytes/storage compatibility failures." + "reason": "bitsandbytes Int8Params argument incompatibility is different from a Float8 storage lookup failure." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption is unrelated to generic loading noise." + "reason": "A corrupted Ovis2 checkpoint is not the same as a generic loading-noise complaint." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different loading failures: bad AutoTokenizer error message vs missing tied-weights attribute." + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to missing all_tied_weights_keys on Molmo." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Zero probabilities in SigLIP2 is not the same as lower text-embedding quality." + "reason": "Zero probabilities in SigLIP2 and worse text embeddings are related model-quality reports but not the same concrete defect." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 key error is unrelated to Qwen3_5Moe rotary embedding initialization." + "reason": "Gemma4 FSDP2 shared_kv_states rebuild and Qwen3_5Moe inv_freq uninitialized memory are separate model-specific bugs." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen3.5-related, but one is flash-attn position_ids handling and the other is CPU-offload compatibility." + "reason": "Flash-Attention packed-sequence misinterpretation and CPU-offload incompatibility in rotary embedding are different failures." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "Model quality complaint is unrelated to a tokenization-change regression." + "reason": "SigLIP2 embedding quality regression is not the same as a general tokenization-version change report." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Bad AutoTokenizer error message and glm-4v loading failure are distinct bugs." + "reason": "Voxtral tokenizer error handling and glm-4v-9b loading failure are unrelated." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Gradient-checkpointing ineffectiveness is not the same as int4 CUDA reserved-memory OOM." + "reason": "LoRA gradient checkpointing ineffectiveness is a different issue from int4 quantization memory growth/OOM." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise is not the same bug as a specific model failing to load." + "reason": "Generic loading noise is not the same as inability to load a specific model repo." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Detokenization regression and AMD GPU support failure are unrelated." + "reason": "DeepSeek detokenization regression and GPT-OSS-20B AMD GPU failure are unrelated." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Different SigLIP2 problems: embedding quality vs HF/JAX implementation discrepancy." + "reason": "SigLIP2 text embedding quality and HF-vs-JAX implementation discrepancy are not the same report." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Two model-loading complaints, but no evidence they stem from the same concrete defect." + "reason": "Generic loading noise is not the same as a specific sleng-bert load failure." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "`trust_remote_code` missing fields is more specific than the broad remote-code compatibility report." + "reason": "Missing fields in AutoConfig with trust_remote_code is too specific to merge with a broader remote_code breakage report." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Distinct training/memory issues with different symptoms and likely different fixes." + "reason": "Gradient checkpointing with PEFT LoRA and inefficient MoE memory usage are different problems." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 loading and Qwen3Moe FP8 failure are unrelated." + "reason": "Gemma3 ZeRO-3 loading and Qwen3MoE FP8 config failure are unrelated." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model-loading failures with different root causes." + "reason": "glm-4v-9b loading failure and Molmo tied-weights attribute error are different model-specific bugs." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram-op failure is unrelated to Gemma4 PLE device placement mismatch." + "reason": "MPS histogram support in MoE is unrelated to Gemma4 PLE device mismatch under device_map=auto." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The overall clustering is mostly grounded, but at least one soft-edge rationale is not precise enough: the `issue:43644` vs `issue:43950` explanation mentions `trust_remote_code` missing-fields behavior, which is a different issue (`issue:43643`) and not the pair under review. That kind of mismatch makes the verdicts less reliable. I\u2019d also soften the summary from \"the only clear duplicate pair\" to \"the clearest duplicate pair in this packet\"." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43950/issue:44534, is well supported by the matching v5 non-persistent-buffer regression titles, and the rejected soft pairs are framed as distinct bugs without overstating certainty. No PR mergeability concerns apply here." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1235,8 +1415,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -1322,232 +1502,226 @@ "issue:44530|issue:44863", "issue:43366|issue:43828" ], - "prepared_review_unit_hash": "aa832cdd279b487f0768b97b8eb7c042fb23cc2602a6e642749f1300f4543028", + "prepared_review_unit_hash": "331e76db7a687030b5bd96fbc24fcfd54c12fb283e2eb0d7318c55b889d87155", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 32, - "serialized_chars": 37048, - "estimated_input_tokens": 9262, - "estimated_eval_tokens": 18780 + "serialized_chars": 37049, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose mix of Transformers regression/loading issues rather than one duplicate set. Clear duplicate-like pairs are the docs pipeline removal reports, the mask-shape mismatch reports, the bitsandbytes/init_empty_weights `_is_hf_initialized` reports, and likely the Qwen3 ZeRO-3 weight-loading reports.", - "confidence": 0.79, - "canonical_issue_reason": "No single issue cleanly represents the whole cluster; if one anchor is needed, 45020 is the broadest umbrella for recent model-loading/remote_code compatibility regressions.", + "summary": "Mostly unrelated issues, but there are a few true duplicates: the v5 pipeline-docs regression, the `_is_hf_initialized` TypeError during model loading, and the mask/tensor shape-mismatch IndexError. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 41093 has the clearest concrete bug report for the shape-mismatch IndexError, and 44805 is just the generalized/placeholder version of the same failure.", "canonical_pr_reason": null, - "best_issue_reason": "45020 is the most central and general issue in the set, and it overlaps with the largest number of other loading-regression reports.", + "best_issue_reason": "Issue 41093 is the strongest representative runtime bug in this set: specific, reproducible, and duplicated by the generalized 44805 report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Different failures: tokenizer loading for EMBEDDIA/est-roberta vs flash-attn fallback gating." + "reason": "Different subsystems and failures: tokenizer loading vs flash-attn fallback gating." }, { "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 config warning about `tie_word_embeddings`." + "reason": "Related Qwen/LoRA wording, but one is a model-compatibility failure and the other is a warning about `tie_word_embeddings`; not the same bug." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Importing a missing Qwen2.5-VL class is unrelated to the bitsandbytes `_is_hf_initialized` constructor error." + "reason": "Import error for a missing class vs bitsandbytes kwarg incompatibility; different failure modes." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Broad remote_code regressions and a Gemma4 PLE device-mapping mismatch are different bugs." + "reason": "Broad remote-code regression vs a Gemma4 device-mismatch bug; not the same underlying issue." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "`add-new-model-like` tokenizer-mapping failure is a distinct metadata/tooling bug from remote_code loading regressions." + "reason": "Tokenizer-mapping-name edge case is unrelated to the remote_code regression." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to Gemma4 PLE device placement." + "reason": "Different code paths: autocast dtype mismatch vs device placement mismatch." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral regex patch crash and flash-attn fallback blocking are separate tokenizer vs attention-path failures." + "reason": "Mistral regex patch crash is unrelated to flash-attn2 fallback checks." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight_only load failure and GLM5 support are different model-loading problems." + "reason": "Qwen3VL `weight_only` load error and GLM5 loading are different issues." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM key error is unrelated to flash-attn installation fallback logic." + "reason": "MobileLLM key error is unrelated to flash-attn fallback logic." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote_code breakage is broader but not the same as Kimi-K2.5 tokenizer codec/regression handling." + "reason": "Both mention recent-version regressions, but one is broad remote-code breakage and the other is a specific Kimi-K2.5 tokenizer regression." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "lm_head tying in Mistral3 is a model-specific weight-sharing bug, not a non-persistent buffer regression." + "reason": "Tied `lm_head` weights and non-persistent buffer corruption are different bugs." }, { "left": "issue:43638", "right": "issue:45823", "accept": false, - "reason": "ZeRO-3 empty-Bert index error and Gemma4 device mismatch are different code paths." + "reason": "ZeRO-3 index error with a non-pretrained Bert model is unrelated to the Gemma4 device mismatch." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "LoRA checkpoint compatibility across hf4.x/hf5.x is not the same bug as generic remote_code model breakage." + "reason": "Qwen3 MoE FP8 failure is unrelated to the remote_code tokenizer/model-loading regression." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both report the same docs problem: v5 docs still reference removed pipeline tasks like summarization/translation/text2text-generation." + "reason": "Same docs regression: v5 removed pipeline tasks, but the docs still reference them." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Whisper runtime failure is too broad to confidently equate with the specific `_torch_extract_fbank_features()` issue." + "reason": "Whisper runtime failure and `_torch_extract_fbank_features()` are too vague to confirm the same underlying bug." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same underlying bitsandbytes / `init_empty_weights` regression around unexpected `_is_hf_initialized` being passed to `Int8Params`." + "reason": "Same unexpected `_is_hf_initialized` TypeError during loading; one reports the bitsandbytes `Int8Params` path and the other the `init_empty_weights` path." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Both involve loading issues, but one is LoRA model compatibility and the other is a bitsandbytes constructor regression." + "reason": "Different model/loading problems: Qwen3 MoE FP8 vs bitsandbytes init kwarg incompatibility." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed is unrelated to an Ovis2 checkpoint corruption report." + "reason": "Qwen3 MoE weight-loading issue is unrelated to the corrupted Ovis2 checkpoint report." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Bad AutoTokenizer error messaging for Voxtral is not the same as est-roberta tokenizer loading failure." + "reason": "Tokenizer error-message quality issue vs a tokenizer loading failure for a different model." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing not working under PEFT LoRA is distinct from a missing `supports_gradient_checkpointing` flag on a trust_remote_code model." + "reason": "Gradient-checkpointing effectiveness and missing support flag are related in theme but not the same concrete bug." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Unexpected `num_items_in_batch` kwarg in `compute_loss` is unrelated to the mask/tensor shape mismatch." + "reason": "Unexpected `compute_loss` kwarg and mask-shape IndexError are unrelated training failures." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "RLHF NaN/SDPA fallback due to 3D position_ids is unrelated to a sink-less flash-attention `s_aux=None` crash." + "reason": "Catastrophic gradient explosion and flash-attn `s_aux=None` crash are different problems." }, { "left": "issue:43940", "right": "issue:45313", - "accept": true, - "reason": "Both describe the same ZeRO-3 weight-loading failure class on Qwen3-family models, with params ending up missing during sharded load." - }, - { - "left": "issue:43257", - "right": "issue:45305", "accept": false, - "reason": "Missing weight conversion for Qwen3 MOE is not the same as gradient averaging with DeepSpeed GAS." + "reason": "Both are DeepSpeed weight-loading issues, but they target different Qwen variants and symptoms; not enough to treat as one bug." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "GLM-4.7-Flash loading in tests and NemotronH checkpoint loading are different model-specific failures." + "reason": "Model-loading failure for GLM-4.7-Flash vs NemotronH checkpoint incompatibility are unrelated." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "bitsandbytes `_is_hf_initialized` incompatibility is unrelated to CLIPTokenizer's huge `model_max_length`." + "reason": "Bitsandbytes kwarg error is unrelated to CLIPTokenizer's `model_max_length` default." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Voxtral tokenizer error messaging and failure to load `cjvt/sleng-bert` are different tokenizer/model issues." + "reason": "Bad error message for Voxtral tokenizer loading is unrelated to `cjvt/sleng-bert` not loading." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "These are both video-related Qwen issues, but one is a broad v5.3.0 regression and the other is a specific StopIteration on Qwen3-vl-embedding." + "reason": "Both are video-input regressions, but the affected models and failure signatures differ; not the same concrete bug." }, { "left": "issue:41093", "right": "issue:44805", "accept": true, - "reason": "Same underlying mask shape mismatch: the indexed tensor and mask lengths differ by one in both reports." + "reason": "Same mask/tensor shape-mismatch IndexError; 44805 is the genericized duplicate of the exact error in 41093." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training RuntimeError about 2-D weights is unrelated to void segmentation map processing." + "reason": "DeepSpeed training weight-shape runtime error and segmentation-map processing are unrelated." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Qwen3.5 PagedAttentionCache linear_attention failure and NemotronH checkpoint loading are unrelated." + "reason": "PagedAttentionCache group-type crash and NemotronH checkpoint loading are different failures." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF gpt-oss support and Phi-tiny-MoE autocast dtype mismatch are different features/failures." + "reason": "GGUF gpt-oss support and autocast dtype mismatch are unrelated issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays conservative and is grounded in the packet. It correctly treats the cluster as mostly a loose set of loading/regression issues, and the cited duplicate-like pairs are supported by the issue titles and soft similarity edges (docs pipeline removals, mask-shape mismatch, bitsandbytes/init_empty_weights, and Qwen3 ZeRO-3 loading). The anchor choice of 45020 as the broadest umbrella is reasonable given its connections, without overclaiming a single exact duplicate set." + "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted duplicates are well supported by the titles/body patterns in the listed issues, and the rejected pairs avoid overstating weak similarity. No PR pairs are present, so there is no mergeability concern." }, "evaluator_used": true, "retried": false, @@ -1563,8 +1737,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -1650,14 +1824,14 @@ "issue:42491|issue:45276", "issue:42915|issue:44756" ], - "prepared_review_unit_hash": "105fb231299874807e331343cb4dad7d934e49d104554ac150956223bd3b3dab", + "prepared_review_unit_hash": "e1f931eabae8f43bc620cf9a1b1104d5ee597f14d6408426a53da1d74c5a1082", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, @@ -1672,198 +1846,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of model-specific issues, with only a few true duplicate pairs: the Transformers v5 non-persistent-buffer regression, the custom-model initialization regression, and the SigLIP2 output-corruption reports. No PRs are present.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43644 is the clearest representative of a concrete regression, and it has an almost verbatim duplicate (44534) with the same failure mode.", + "summary": "This cluster is mostly a mix of unrelated Transformers regressions across loading, config, and model-specific inference bugs. Only a few pairs look like true duplicates: custom model initialization, non-persistent buffer corruption, SigLIP2 bad outputs, and the BERT loading regression affecting multiple models.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43646 is the broadest, most representative issue in the cluster: it frames the shared Transformers v5 custom-model initialization regression in a generic way that can absorb the more specific notebook repro.", "canonical_pr_reason": null, - "best_issue_reason": "43644 is the best issue-level canonical candidate because it states the bug directly, is earlier, and cleanly anchors the strongest duplicate pair.", + "best_issue_reason": "Issue 43646 is the best overall anchor because it is the clearest umbrella for the custom-init regression and is less model- or environment-specific than the nearby duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Same model family, but different bugs/code paths: Qwen2.5-VL rope index scaling vs Qwen3.5-MoE rotary embedding materialization." + "reason": "Different bugs: Qwen2.5-VL temporal position scaling vs Qwen3_5Moe rotary embedding reading uninitialized memory after meta materialization." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4, but one is an FSDP2 training KeyError and the other is a device-map forward mismatch; not the same concrete bug." + "reason": "Same model family, but different failures: FSDP2 shared_kv_states KeyError vs device_map='auto' PLE device mismatch." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different topics: GGUF/gpt-oss architecture support vs Qwen3-MoE sliding_window behavior." + "reason": "Unrelated issues: GGUF support for gpt-oss vs a Qwen3-MoE sliding_window problem." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption and Kosmos2.5 long-OCR index errors are unrelated failures." + "reason": "Different code paths and symptoms: non-persistent buffer corruption vs Kosmos2.5 long-OCR index error." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Different models and failures: Qwen3-vl embedding video StopIteration vs an MPS histogram op missing for Int." + "reason": "Different model/runtime bugs: Qwen3-VL video StopIteration vs MPS histogram not implemented for Int." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Both are attention/compile regressions, but on different models and different mechanisms; not mergeable as one bug." + "reason": "Different concrete failures: Bamba torch.compile+SDPA crash vs Qwen3.5 flash-attention illegal memory access from 3D position_ids." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying Transformers v5 custom-model initialization regression; 43645 is just the Jupyter-specific manifestation of 43646." + "reason": "Same underlying Transformers 5.0 custom-model initialization regression; the notebook wording is just a more specific reproduction." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Unrelated: GGUF support for gpt-oss vs a Float8 storage lookup error." + "reason": "Unrelated: gpt-oss GGUF support request vs Float8 storage lookup error." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Different Qwen3.5-related symptoms and code paths; one is video embedding generation, the other SFT embedding unpacking." + "reason": "Different bugs: Qwen3-VL video StopIteration vs Qwen3.5 embedding unpacking failure in TRL SFT trainer." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "MoE FP8 failure and meta-device loading breakage are different regressions." + "reason": "Different regressions: Qwen3-MoE FineGrainedFP8Config failure vs generic meta-device loading issues." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "Unrelated issues: 3d_parallel.py execution vs GGUF model support." + "reason": "No shared underlying bug: 3d_parallel.py failure vs GGUF gpt-oss support." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Different failures: missing Qwen2.5-VL import vs meta-device loading breakage." + "reason": "Different problems: missing export for Qwen2.5-VL class vs model loading with torch.device('meta')." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin is unrelated to Qwen3Moe FP8 behavior." + "reason": "Completely different: GenerationMixin import error vs Qwen3-MoE FP8 config failure." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer/model mapping failure and Gemma4 PLE device mismatch are different bugs." + "reason": "Different areas: tokenizer mapping/add-new-model-like bug vs Gemma4 PLE forward device mismatch." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "_is_hf_initialized flag regression is unrelated to Kosmos2.5 OCR index errors." + "reason": "Different regressions: _is_hf_initialized flag breakage vs Kosmos2.5 OCR index error." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Different symptoms and models; not the same underlying bug." + "reason": "Different symptoms and likely different causes: video StopIteration vs mask/tensor shape mismatch." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 bad-output regression: unexpected zero probabilities and nonsensical AutoModel/pipeline results on the same checkpoint." + "reason": "Same SigLIP2 model regression: zero probabilities and nonsensical outputs point to the same bad inference/loading behavior." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batched-track leakage and Whisper empty-transcription failure are unrelated." + "reason": "Different model-specific bugs: AudioFlamingo3 batched leakage vs Whisper empty-transcription failure." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Zero3 BertModel index error is unrelated to add-new-model-like / tokenizer mapping handling." + "reason": "Different issues: ZeRO-3 BertModel index error vs add-new-model-like/tokenizer-mapping failure." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Both mention loading, but the concrete bugs differ: Deepspeed Zero3 BertModel indexing vs tokenizer mapping logic." + "reason": "Different bugs: zero3 BertModel initialization index error vs tokenizer mapping validation issue." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Different bug classes: model-registration/mapping vs attention-mask shape mismatch." + "reason": "Unrelated failures: tokenizer mapping logic vs mask shape mismatch at indexing time." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both concern config validation, but for different models and different failures." + "reason": "Both mention config validation, but the affected models and failures are different; not enough to be the same bug." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Dequantized model offload loading and Strix Halo mmap OOM are not the same bug." - }, - { - "left": "issue:38175", - "right": "issue:39692", - "accept": false, - "reason": "Same SigLIP2 area, but one is an inference/output regression and the other is a docs example with model/processor mismatch and quantization errors." + "reason": "Different concerns: gpu+cpu offload dequantized load failure vs disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different mechanisms: Qwen3_MoE LoRA incompatibility across versions vs BitNet packed-weight loading." + "reason": "Different code paths: qwen3_moe LoRA incompatibility across versions vs BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both involve Qwen3, but auto device mapping on A800 and FineGrainedFP8Config failures are distinct code paths." + "reason": "Different runtime failures: auto device mapping cudaErrorAssert vs Qwen3-MoE FineGrainedFP8Config." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "GLM5 and Qwen3.5 flash-attention failures are separate model-specific regressions." + "reason": "Unrelated model bugs: GLM5 issue vs Qwen3.5 flash-attention generation failure." }, { "left": "issue:44451", "right": "issue:44488", - "accept": false, - "reason": "Both are model-load failures, but the titles only show different checkpoints; not enough evidence they share one concrete root cause." + "accept": true, + "reason": "Same loading regression class: both report current/latest Transformers failing to load specific BERT models." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact same bug: both report Transformers v5 filling non-persistent buffers with junk." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "One is ZeRO-3 loading of language_model weights; the other is save_pretrained writing incorrect visual encoder keys." + "reason": "Different Qwen3.5 issues: ZeRO-3 load failure for language_model vs incorrect visual encoder keys on save_pretrained." }, { - "left": "issue:43644", - "right": "issue:44534", - "accept": true, - "reason": "Near-verbatim duplicate reporting the same non-persistent-buffer junk regression in Transformers v5." + "left": "issue:44910", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen3.5 problems: flash-attention misreading 3D position_ids vs TRL SFT embedding unpacking failure." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the provided titles and metadata, and the accepted soft-edge pairs are conservative enough: 43644/44534 is a near-verbatim duplicate, 43645/43646 is a plausible same regression, and 38175/43994 stays within the same SigLIP2 output-corruption symptom class. The canonical issue choice is also well supported by the earlier, clearer duplicate pair. No PRs are present." + "feedback": "The summary is well grounded in the packet, and the accepted soft-edge pairs are conservative. The rejected pairs are also phrased cautiously and do not overclaim duplicate status." }, "evaluator_used": true, "retried": false, @@ -1879,8 +2053,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -1968,14 +2142,14 @@ "issue:45276|issue:45292", "issue:43854|issue:45216" ], - "prepared_review_unit_hash": "c26910eb47e3fad5439ebb228f5156e2f0b2520ad19822b374fafc5140a2ac79", + "prepared_review_unit_hash": "813553c9573c56a14ea9c10c98698b7aa6e8c4b1168ce41eb9cdc8167fd1961f", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, @@ -1990,217 +2164,220 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose set of Transformers issues spanning quantization/offloading, tokenizer/loading regressions, and model-specific bugs. The soft pairs look related by broad subsystem, but none appear to be the same concrete bug/change.", - "confidence": 0.84, - "canonical_issue_reason": "issue:43873 is the broadest and most central open report around quantization/offloading memory behavior, which is the closest shared theme in this otherwise heterogeneous cluster.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues. The only pair that looks like a true duplicate is the resize_token_embeddings regression affecting output embeddings in Gemma4/generic form (45276 and 45292).", + "confidence": 0.71, + "canonical_issue_reason": "issue:45276 is the most representative issue because it describes the broad resize_token_embeddings failure and includes the concrete affected model plus the shared output-embedding symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best single representative because it is open, fairly well-discussed, and broad enough to cover the cluster\u2019s dominant memory/offloading theme, even though the items are not true duplicates.", + "best_issue_reason": "issue:45276 is the best single issue to keep: it is the clearest, most actionable representative of the resize_token_embeddings regression and has the strongest duplicate overlap with another item.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume step math and unused SGD args are unrelated training bugs." + "reason": "Different training bugs: checkpoint resumption step \u0a97\u0aa3culation vs optimizer arguments being ignored." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both involve SigLIP2, but one is a broken documentation example and the other is a model/pipeline inference regression." + "reason": "Both involve SigLIP2, but one is a broken documentation example while the other is a model/pipeline output regression." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Tokenizer regressions in v5, but different code paths and model families (DeepSeek vs Granite)." + "reason": "Different tokenizer regressions affecting different model families and different failure modes." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different model loading failures for unrelated models." + "reason": "Different model-loading failures for different models; no clear shared underlying bug." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Qwen3 MoE loading/acceleration conversion vs GLM unit-test loading failure are unrelated." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to GLM-4.7-Flash loading." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory, but one is Qwen3 MoE training memory usage and the other is quantization/offloading correctness." + "reason": "Both mention memory/quantization, but one is training-memory inefficiency and the other is offloading not working under quantization." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "LoRA checkpoint compatibility across HF versions is not the same bug as a Qwen3-VL shape mismatch on load." + "reason": "LoRA compatibility across hf versions is not the same issue as a Qwen3-VL weight-shape mismatch." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Two separate Transformers 5 regressions: non-persistent buffers vs custom model initialization in notebooks." + "reason": "Both are Transformers 5 regressions, but one is non-persistent buffers and the other is notebook custom-model initialization." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is training-memory inefficiency and the other is int4 reserved-memory growth/OOM." + "reason": "Related to memory, but the concrete problems differ: Qwen3 MoE training memory vs int4 quantization CUDA reservation/OOM." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA compatibility and resize_token_embeddings propagation are different problems." + "reason": "Different problems: LoRA model compatibility versus token-embedding resize propagation." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Different model/tokenizer loading failures for different checkpoints." + "reason": "Both are tokenizer/load failures, but the affected models and likely root causes differ." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "FlashAttention crash from 3D position_ids misinterpretation is distinct from the s_aux=None AttributeError." + "reason": "Different flash-attention crashes with different root causes and code paths." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Different inference entrypoints and different model classes; not the same failure mode." + "reason": "Different runtime failures: chat invocation crash vs serve processor attribute error." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "save/from_pretrained extra_state handling and init_empty_weights argument handling are separate save/load issues." + "reason": "Serialization extra_state handling is unrelated to init_empty_weights/_is_hf_initialized initialization." }, { "left": "issue:45276", "right": "issue:45335", "accept": false, - "reason": "Both are resize_token_embeddings regressions, but on different model architectures and submodules." + "reason": "Same API area, but the affected models/fields differ enough that this looks model-specific rather than one concrete duplicate bug." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "An import error for GenerationMixin is unrelated to Whisper pipeline return_language behavior." + "reason": "ImportError for GenerationMixin and Whisper return_language pipeline regression are unrelated." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch version compatibility and meta-device loading are different compatibility bugs." + "reason": "General Torch 2.10 incompatibility is not the same as meta-device model-loading breakage." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both concern lm_head tying, but they are model-specific implementations rather than one concrete shared bug." + "reason": "Both concern untied lm_head weights, but they appear to be separate model-specific bugs rather than one concrete duplicate." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "A SigLIP2 documentation/example issue is not the same as a GLM5 model-loading report." + "reason": "Documentation/example errors are unrelated to the GLM5 issue." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Tokenizer vocab-size mismatch and loading failure for a different model are not the same issue." + "reason": "Both mention tokenizer/vocab loading, but they are different models and likely different causes." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing/PEFT and an apex RMSNorm memory leak are unrelated." + "reason": "Gradient checkpointing/LoRA and an apex RMSNorm memory leak are unrelated." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states and Qwen multimodal config unwrapping are distinct code paths." + "reason": "FSDP2 KeyError in Gemma4 attention is unrelated to Qwen multimodal text_config unwrapping." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Different models and different failure modes; not a shared underlying bug." + "reason": "Different model families and different offload/path failures." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Related memory/quantization theme, but one is offloading correctness and the other is reserved-memory growth under int4." + "reason": "Quantization/offloading issues, but not the same bug or change." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo track leakage and Gemma4 device mismatch are unrelated model-specific bugs." + "reason": "AudioFlamingo3 batched inference leak is unrelated to Gemma4 device-map mismatch." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights type handling and Strix Halo mmap OOM are separate problems." + "reason": "Empty-weights initialization and Strix Halo mmap OOM are unrelated." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and Gemma4 FSDP2 key handling are unrelated." + "reason": "Sparse-layer creation in SwitchTransformers is unrelated to Gemma4 FSDP2 attention state rebuilding." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different quantization formats and different model loading failures; too far apart to merge." + "reason": "Different quantization formats and different load failures; not the same underlying bug." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "A generic GLM5 issue and a save/load regression for sinq-quantized models are not the same bug." + "reason": "GLM5 is unrelated to sinq save/load breakage." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Both involve OOM, but the causes are different: int4 reserved memory vs mmap behavior on Strix Halo." + "reason": "Different OOM causes and different environments." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Gradient checkpointing with PEFT and checkpointing-in-eval-mode are different API/behavior bugs." + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode restriction." }, { "left": "issue:45276", "right": "issue:45292", - "accept": false, - "reason": "The titles are closely related, but one is Gemma4-specific and the other is a broader output_embeddings regression; not clearly the same concrete bug." + "accept": true, + "reason": "These describe the same resize_token_embeddings regression: output embeddings are not updated, with 45276 being the model-specific variant and 45292 the generic one." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Different model families and different checkpoint/save_pretrained symptoms." + "reason": "Different model-loading/save-pretrained issues affecting different models and versions." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The 45276/45292 duplicate call is well supported by the packet, and the other soft-edge rejections stay conservative without overstating evidence. The summary is mostly accurate; at most, 45276/45335 is a borderline similarity, but not enough in the packet to overturn the analyst\u2019s conservative split." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -2212,8 +2389,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -2299,235 +2476,235 @@ "issue:43596|issue:44683", "issue:43994|issue:44079" ], - "prepared_review_unit_hash": "2ff4a0daa0d66f1e51a214a14c24863c204d4845edc5b5aa5f65901304dd5b5a", + "prepared_review_unit_hash": "dae3557c8203ff772db92fee53f59acae40748a6b95242ee6697011b169175a6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 32, - "serialized_chars": 36879, - "estimated_input_tokens": 9220, - "estimated_eval_tokens": 18696 + "serialized_chars": 36883, + "estimated_input_tokens": 9221, + "estimated_eval_tokens": 18698 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports across different subsystems. The only clear duplicate pair is the resize_token_embeddings regression: a general report plus a t5gemma-specific instance. No pull requests are present.", + "summary": "This cluster is mostly heterogeneous Transformers regression reports with one clear duplicate pair: `resize_token_embeddings` not updating output embeddings (`issue:45292` / `issue:45335`). The other soft-similarity pairs are related at a high level but describe different failures or code paths, so they should stay separate.", "confidence": 0.76, - "canonical_issue_reason": "issue:45292 is the broader, cleaner formulation of the resize_token_embeddings bug; issue:45335 is a model-specific instance of the same underlying failure to update downstream embeddings.", + "canonical_issue_reason": "`issue:45292` is the clearest canonical for the only true duplicate pair: it states the generic `resize_token_embeddings` output-embedding mismatch, while `issue:45335` is a model-specific instance (`t5gemma`/decoder.embed_tokens).", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 best represents the cluster because it states the general API breakage, while the paired t5gemma report is just a narrower reproduction.", + "best_issue_reason": "`issue:43611` is the strongest single representative issue overall: it is a broad, well-scoped Transformers 5.0.0 loading regression affecting model initialization/loading, which matches the cluster\u2019s main theme better than the narrower model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Different failures: eval-mode gradient checkpointing restriction vs Qwen3.5 RLHF NaN/SDPA fallback behavior." + "reason": "Both mention training/runtime instability, but one is about gradient checkpointing in eval mode and the other about Qwen3.5 RLHF NaNs from 3D position ids / SDPA fallback; different bug paths." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Unrelated bugs: extra_state save/load handling vs mmap-related OOM on Strix Halo." + "reason": "One is about `extra_state` save/load handling, the other about disabling mmap on Strix Halo to avoid OOM; unrelated mechanisms." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Different model families and symptoms: video input regression vs strict config loading failure." + "reason": "Video-input regression for Qwen VL models vs strict config blocking `granite_speech`; different model families and failure modes." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Different code paths: flash_attention sink-less crash vs Gemma4 FSDP2 shared_kv_states KeyError." + "reason": "Flash-attention crash on `s_aux=None` for sink-less models is a different attention-path bug than Gemma4 FSDP2 `shared_kv_states` rebuilding." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "No shared underlying bug; one is device mapping/cuda assert on Qwen3, the other is batched inference token leakage in AudioFlamingo3." + "reason": "Auto device mapping CUDA assert on Qwen3 vs AudioFlamingo3 batched inference token/embedding leakage; not the same underlying defect." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both concern custom-model save/load, but one is file permission propagation and the other is loading the wrong module; not the same concrete bug." + "reason": "Read-only permission propagation in `save_pretrained` is a packaging/filesystem issue; wrong custom module loaded after save is an auto-class/module resolution issue." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage is unrelated to disabling mmap for OOM on Strix Halo." + "reason": "Meta-device loading breakage on some models vs mmap OOM on Strix Halo; shared theme is loading, but the root causes differ." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to Trainer checkpoint-resume max_steps calculation." + "reason": "Custom model initialization regression vs Trainer checkpoint-resume `max_steps` miscalculation; separate trainer/loading code paths." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook-defined custom model init is not the same as extra_state persistence in save/from_pretrained." + "reason": "Notebook-specific custom model init breakage vs `save/from_pretrained` failing to handle `extra_state`; different subsystems." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same underlying bug: resize_token_embeddings fails to update the model's effective output/decoder embeddings; one is the generic report and the other a t5gemma-specific instance." + "reason": "Same underlying regression: `resize_token_embeddings` updates one embedding path but not the output/decoder embedding; `t5gemma` is a model-specific manifestation of the same bug." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Different issues: autocast dtype mismatch vs meta-device model loading problems." + "reason": "Both involve dtype/device behavior during loading or inference, but one is an autocast mismatch in Phi-tiny-MoE and the other is meta-device loading failures across models." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is a broad environment issue, not the same as Phi-tiny-MoE autocast dtype mismatch." + "reason": "Torch version incompatibility is a broad environment issue; the other is a specific autocast dtype mismatch on a particular model." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Both relate to loading/tokenizer setup, but the specific failure modes are different and not clearly the same bug." + "reason": "Tokenizer vocab-size mismatch for one model vs inability to load a different model at latest version; too little evidence they share one code path." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to mmap OOM behavior." + "reason": "Float8 storage deserialization error is unrelated to mmap/OOM on Strix Halo." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is unrelated to Gemma4 device-map forward mismatch." + "reason": "Compiled flex_attention on torch>=2.9 vs Gemma4 PLE device mismatch under `device_map=auto`; different layers and bugs." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Different model-loading failures for different architectures; too broad to treat as the same bug." + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed is distinct from NemotronH checkpoint loading failure." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper return_language pipeline regression is unrelated to a general tokenization-change report." + "reason": "Whisper pipeline `return_language` regression and tokenization changing across transformers versions are different features with different symptoms." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch 2.10 compatibility and Gemma4 device mismatch are unrelated." + "reason": "Torch 2.10 compatibility is a general dependency issue, while Gemma4 PLE device mismatch is a model-forward placement bug." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both involve Transformers 5 custom-model behavior, but one is base_model_prefix loading and the other is notebook-defined model initialization; not the same concrete defect." + "reason": "Both concern Transformers 5 custom model handling, but one is about `base_model_prefix` loading and the other about notebook redefinition/init behavior; not the same concrete bug." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch compatibility issue vs multimodal dtype mismatch; no shared root cause." + "reason": "Torch 2.10 incompatibility vs Mistral-3 image preprocessor/model dtype mismatch; different scope and root cause." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to a compiler error about non template nodes." + "reason": "PagedAttentionCache crash on Qwen3.5 linear_attention vs compile-time template-node error; unrelated failures." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction issue does not match the specific Qwen3.5 PagedAttentionCache crash." + "reason": "An isolated reproduction of an older issue vs a specific Qwen3.5 cache crash; no evidence they are the same code-path bug." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization loading fallback." + "reason": "Whisper pipeline `return_language` and MXFP4 quantization fallback are unrelated model/pipeline behaviors." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export with torch_compilable_check is unrelated to Perceiver non-default resolution/interpolate_pos_encoding failure." + "reason": "`torch.export`/`torch_compilable_check` failure vs Perceiver non-default resolution interpolation failure; different APIs." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression for Qwen models is unrelated to a mask/tensor shape mismatch IndexError." + "reason": "Video-input regression in several Qwen VL models vs tensor mask shape mismatch; not the same failure mode." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch version incompatibility is unrelated to Qwen3.5 SFT trainer embedding unpacking." + "reason": "Torch version incompatibility vs Qwen3.5 embedding unpacking with TRL SFT trainer; different problem domains." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs mismatches, but they cover different removed/changed pipeline APIs and are not the same underlying bug." + "reason": "Docs drift about `return_all_scores` vs docs mentioning removed pipeline tasks; documentation-only issues, but not the same missing/changed content." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Different failures: missing Float8 storage object vs compiler rejecting non-template nodes." + "reason": "Float8 storage lookup failure vs compile template node error; unrelated symptoms." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "extra_state save/load handling is unrelated to config unwrapping for composite Qwen multimodal models." + "reason": "`extra_state` save/load handling vs `AutoModelForCausalLM.from_config` not unwrapping `text_config`; separate model-construction paths." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer method-name conflict in T5 is not the same as gguf_file breaking AutoTokenizer/AutoModel loading." + "reason": "T5Tokenizer special-token method conflict vs `gguf_file` breaking `from_pretrained`; no shared bug mechanism." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed ZeRO-3/BertModel index error is unrelated to compiled flex_attention on newer torch." + "reason": "Deepspeed ZeRO-3 init index error in BertModel vs compiled flex_attention breaking on torch>=2.9; unrelated." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Siglip2 nonsensical outputs are a model/pipeline behavior issue, not the same as ModelOutput key assignment when previous value is None." + "reason": "Siglip2 nonsensical outputs with AutoModel/pipeline vs `ModelOutput` key assignment bug; different layers of the stack." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, and the 45292/45335 duplicate call looks plausible. However, at least one soft-edge verdict is too strong for the evidence available: issue:43906 is only titled as an isolated reproduction of another issue, so saying it has 'no shared underlying bug' with issue:44530 overstates what the packet supports. Please soften such cases to 'insufficient evidence to merge' unless the bodies clearly establish unrelated failures." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge calls. The lone duplicate pair (`issue:45292` / `issue:45335`) is a reasonable canonical merge, and the other pairs are rejected without overclaiming shared root cause. No evidence of overstated similarity." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2540,8 +2717,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -2551,7 +2728,6 @@ "issue:39404", "issue:40444", "issue:40990", - "issue:42175", "issue:42915", "issue:43299", "issue:43317", @@ -2584,14 +2760,15 @@ "issue:44737", "issue:44756", "issue:44811", + "issue:44849", "issue:44987", "issue:44991", "issue:45081", "issue:45084", + "issue:45325", "issue:45584", "issue:45859", - "issue:45987", - "issue:46178" + "issue:45987" ], "soft_pairs": [ "issue:43299|issue:43931", @@ -2623,216 +2800,216 @@ "issue:43452|issue:43482", "issue:43317|issue:43482", "issue:36032|issue:43645", - "issue:42175|issue:46178" + "issue:44849|issue:45325" ], - "prepared_review_unit_hash": "991bd1aff0f37ff73ef51cf321c2e0bad5abf19a83542496dbfca9f10081f9d7", + "prepared_review_unit_hash": "ab75c416dd885b5860c0d286f9cecb61bb033a71a62b9ebddfe4363de662ad4c", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 35996, - "estimated_input_tokens": 8999, - "estimated_eval_tokens": 18254 + "serialized_chars": 36003, + "estimated_input_tokens": 9001, + "estimated_eval_tokens": 18258 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of distinct bug reports, many around Transformers v5 regressions and model/tokenizer loading, but the soft pairs do not look like true duplicates. The best representative issue is the broad custom model initialization regression.", + "summary": "This cluster is a loose semantic grouping of Transformers v5 regressions across many unrelated models/components. I don\u2019t see any soft pair that clearly describes the same underlying bug, so I\u2019d keep the issues separate and use one broad model-init regression as the representative issue.", "confidence": 0.69, - "canonical_issue_reason": "issue:43646 is the broadest, least environment-specific regression here: it targets generic custom model initialization, making it a better cluster representative than narrower model-, tokenizer-, or hardware-specific reports.", + "canonical_issue_reason": "Issue 43646 is the broadest, most central representative of the shared v5 regression theme around model initialization/loading; the other items are mostly model-specific or unrelated subsystems.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the most general standalone issue in the set and best captures the cluster\u2019s shared v5-regression theme without overfitting to one model or platform.", + "best_issue_reason": "43646 is the best single issue to anchor this cluster because it is general enough to represent the common Transformers v5 breakage pattern without being tied to one model or backend.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen3-VL loading problems, but one is a dev-version breakage and the other is a checkpoint shape mismatch." + "reason": "Both are Qwen loading failures, but they involve different model families and different failure modes (FineGrainedFP8Config vs weight-shape mismatch)." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading failure vs. Mistral regex-patch crash; different code paths and symptoms." + "reason": "Both affect tokenizer loading, but one is an Mistral regex patch crash and the other is a specific tokenizer regression for EMBEDDIA/est-roberta; not the same bug." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Training memory inefficiency and Strix Halo mmap OOM are different issues." + "reason": "Both mention memory/OOM, but one is Qwen3 MoE training memory usage and the other is Strix Halo mmap OOM; different code paths and fixes." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "Masked-LM prediction error and tokenizer loading failure are unrelated." + "reason": "CamemBERT masked-LM prediction corruption is not the same issue as tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline language return and CamemBERT masked-LM predictions are unrelated bugs." + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM prediction bug are unrelated." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "Both touch device/offload behavior, but XLNet positional encoding and Qwen3 MoE rotary embedding are different code paths." + "reason": "Both are device/offload related, but one is XLNet CPU allocation and the other is Qwen3.5 MoE rotary embedding CPU-offload compatibility; different bugs." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to a video-input regression." + "reason": "Save/from_pretrained extra_state handling and video-input regression are distinct problems." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Apple Silicon warmup TypeError and `_is_hf_initialized` regression are distinct failures." + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized flag regression." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Loading a tokenizer and crashing in Mistral regex patching are different bugs." + "reason": "Both are loading failures, but the models and root causes differ; one is a specific repository loading issue and the other is a Mistral tokenizer patch crash." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "A model-mapping failure and compiled flex_attention on torch>=2.9 are unrelated." + "reason": "Tokenizers mapping/model registration failure is unrelated to compiled flex_attention failing on torch >= 2.9." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 loading regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization." + "reason": "Base_model_prefix loading regression and non-persistent buffer junk initialization are separate v5 regressions." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer GPU broadcast NoneType error and Whisper return_language failure are unrelated." + "reason": "Trainer _gpu_broadcast_one NoneType error is unrelated to Whisper return_language behavior." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "High perplexity on gpt-oss and tokenizer post-processor updates are not the same bug." + "reason": "High perplexity on gpt-oss-20b and runtime post-processor updating for special tokens are not the same bug." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor dtype mismatch and template compilation TypeError are unrelated." + "reason": "Image preprocessor dtype mismatch and template compilation error are unrelated." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Different checkpoints and different loading failures; not the same underlying bug." + "reason": "Both are model-loading regressions, but they target different repositories and likely different code paths." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "T5 tokenizer method conflict and Qwen3MoE FP8 config failure are unrelated." + "reason": "T5 tokenizer special-tokens conflict is unrelated to Qwen3Moe FineGrainedFP8Config failure." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "Missing AutoConfig fields and PagedAttentionCache linear_attention crash are different issues." + "reason": "AutoConfig trust_remote_code field loss and PagedAttentionCache invalid group type on Qwen3.5 models are different issues." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Trainer token-count tracking and multi-image IterableDataset fine-tuning failures are unrelated." + "reason": "TrainerState token counter update and Qwen2.5-VL iterable dataset multi-image failure are unrelated." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing tied-weights attribute and tokenizer loading failure are different bugs." + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is not the same as a tokenizer/model loading failure for cjvt/sleng-bert." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both are Whisper-related, but batch_decode skip_special_tokens and empty-transcription generation failures are distinct." + "reason": "Both are Whisper-related, but one is batch_decode skip_special_tokens handling and the other is empty-transcription generation after align_special_tokens." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Different models and different loading regressions; not a duplicate pair." + "reason": "Two separate model-loading regressions for different repositories; no evidence they share the same underlying fix." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch dynamo grouped_mm tracing error and template-node compilation failure are unrelated." + "reason": "Torch dynamo grouped_mm tracing failure for Mixtral is unrelated to the template node compilation TypeError." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3MoE FP8 config failure and Qwen3.5 PagedAttentionCache crash are different code paths." + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache linear_attention crash are different model/backend problems." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are v5 initialization regressions, but one is buffer filling and the other is custom model initialization." + "reason": "Non-persistent buffer junk and custom model initialization regressions are both v5 issues, but not the same defect." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation with compiled forward and StaticCache length typing are unrelated." + "reason": "Chunked generation inconsistency with compiled forward and StaticCache seq-length return type mismatch are distinct cache/compilation issues." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "The titles point to different audio feature-extraction failures; not enough evidence of one concrete bug." + "reason": "The generic _torch_extract_fbank_features issue and the LasrFeatureExtractor center-arg regression are related in area but not clearly the same bug." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Generic gguf_file breakage and a specific Qwen2.5-GGUF loading failure are not clearly the same underlying bug." + "reason": "Both mention GGUF loading regressions, but one is a broad from_pretrained breakage and the other is a specific Qwen2.5-GGUF failure; not enough to merge." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "device_map auto dequantized offload and GGUF model loading are different mechanisms." + "reason": "device_map/offload loading failure and GGUF loading failure are different code paths." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer special-token conflict and custom model initialization in notebooks are unrelated." + "reason": "Tokenizer method-name conflict and Jupyter custom-model initialization regression are unrelated." }, { - "left": "issue:42175", - "right": "issue:46178", + "left": "issue:44849", + "right": "issue:45325", "accept": false, - "reason": "Missing tensorflow in the torch extra and tensorflow_text backend mapping errors are different packaging/backend issues." + "reason": "Qwen3.5 hidden-states bug and Qwen2.5-VL temporal position-id scaling bug are separate model-specific regressions." } ] }, @@ -2851,14 +3028,15 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ "issue:36032", "issue:40990", "issue:41720", + "issue:42175", "issue:42915", "issue:43089", "issue:43262", @@ -2874,7 +3052,6 @@ "issue:43645", "issue:43646", "issue:43742", - "issue:43792", "issue:43819", "issue:43883", "issue:43994", @@ -2892,7 +3069,6 @@ "issue:44683", "issue:44756", "issue:44811", - "issue:44849", "issue:44877", "issue:44987", "issue:44991", @@ -2903,10 +3079,11 @@ "issue:45325", "issue:45357", "issue:45507", - "issue:45987" + "issue:45987", + "issue:46178" ], "soft_pairs": [ - "issue:44849|issue:45325", + "issue:42175|issue:46178", "issue:45084|issue:45507", "issue:43883|issue:44991", "issue:43262|issue:43377", @@ -2937,244 +3114,213 @@ "issue:36032|issue:41720", "issue:43425|issue:44403", "issue:44479|issue:45325", - "issue:40990|issue:44811", - "issue:43792|issue:44589" + "issue:40990|issue:44811" ], - "prepared_review_unit_hash": "45dce90c40204dd8ac656563e71b0f80a1cfb9aba504ce5ab31ef774ae44485f", + "prepared_review_unit_hash": "14e890e2ced348dd75dc27e37439f4088c9734087c9913acabc798a0ddb26b2f", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37527, - "estimated_input_tokens": 9382, - "estimated_eval_tokens": 19020 + "soft_pair_count": 32, + "serialized_chars": 37300, + "estimated_input_tokens": 9325, + "estimated_eval_tokens": 18906 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous Transformers bug reports. The only strong duplicate-like overlap is the Qwen3.5 `save_pretrained` regression pair (45216/45357); the rest are different models or different code paths and should stay separate.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45216 is the best canonical anchor: it states the broader Qwen3.5 `save_pretrained` regression, while issue:45357 looks like a narrower follow-up about the same saved-checkpoint/visual-key bug.", + "summary": "Most pairs are unrelated and should stay separate. The only solid duplicate-like match is the Qwen3.5 `save_pretrained` regression pair (`issue:45216` / `issue:45357`). The Qwen-VL video/rope-index pair (`issue:44479` / `issue:45325`) is too weakly supported from titles alone and should not be accepted as the same bug.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 45216 is the best canonical issue anchor because it describes the broader Qwen3.5 `save_pretrained` regression and has the clearest duplicate-style follow-up in 45357.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45216 is the clearest representative of the only true duplicate pair and captures the underlying serialization regression more broadly than issue:45357.", + "best_issue_reason": "Issue 45216 is the strongest overall issue representative because it has the cleanest confirmed overlap and a broader regression framing than its duplicate follow-up.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:42175", + "right": "issue:46178", "accept": false, - "reason": "Different Qwen VL bugs: hidden-state output vs temporal rope/index scaling." + "reason": "Different backend-mapping failures: missing TensorFlow backend in a torch install vs an offending tensorflow_text backend entry." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Different failures and code paths; template-node compilation error is unrelated to GraniteMoEHybrid invalid method calls." + "reason": "Unrelated failures: template-node compilation error vs an invalid method call in GraniteMoEHybrid." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Unrelated model-loading problems: Molmo tied-weights attribute vs tokenizer load failure for est-roberta." + "reason": "Different model-loading bugs with different symptoms and causes; not the same concrete regression." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Different audio bugs: chat-template sample-rate default vs MIMI batch padding-mask mismatch." + "reason": "Both are audio-related, but sampling-rate defaults and MIMI batching/padding-mask support are separate code paths." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are load failures, but for different models and different root causes." + "reason": "Different tokenizer/model loading regressions affecting different models and mechanisms." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead is unrelated to StaticCache seq-length return type." + "reason": "Generation performance overhead is unrelated to the StaticCache return-type contract issue." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Both are v5 regressions, but one is custom model init and the other is buffer serialization/state corruption." + "reason": "Custom model initialization in notebooks and non-persistent buffer junk are separate v5 regressions." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Different multimodal load/processing bugs: Llava Next segfault vs Qwen3OmniMoe video ValueError." + "reason": "Both are multimodal loading issues, but Llava Next segfault and Qwen3OmniMoe video ValueError are different bugs." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Noise when loading a transformer is unrelated to the Mistral regex patch crash." + "reason": "Noise while loading a transformer is not the same concrete failure as the Mistral regex patch crash." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both involve tied weights, but one is model weight tying and the other is device-dependent serialization; not the same bug." + "reason": "Both involve tied weights, but one is missing tying in a model class and the other is serialization inconsistency." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Different symptoms and code paths: bad SigLIP2 outputs vs OmDet-Turbo processor image-size mismatch." + "reason": "Different subsystems: SigLIP2 inference quality vs OmDet-Turbo image-size mismatch." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3Moe FP8 config failure is unrelated to template-node compilation errors." + "reason": "FP8 config loading failure and non-template-node compile failure are unrelated." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Notebook custom-model init breakage is not the same as non-persistent buffer junk values." + "reason": "Notebook custom-model initialization breakage is not the same as non-persistent buffer corruption." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Both are audio-related, but padding-mask batching and DAC STE mismatch are different underlying bugs." + "reason": "Different audio/model numeric bugs affecting different code paths." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export failures and processor image-size mismatch are separate issues." + "reason": "torch.export compatibility with torch_compilable_check is unrelated to the OmDet processor resolution mismatch." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT loading failure is unrelated to the Mistral tokenizer regex crash." + "reason": "Different tokenizer-loading failures for different models and different causes." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Both are load-time failures, but they affect different models and code paths." + "reason": "Loading physical-intelligence/fast is unrelated to the Mistral regex patch crash." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Feature-extraction fbank issue is unrelated to compiled flex_attention compatibility." + "reason": "fbank feature extraction and compiled flex_attention on torch>=2.9 are unrelated." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Same underlying Qwen3.5 `save_pretrained` regression; 45357 is a narrower manifestation about incorrect visual encoder keys." + "reason": "Same underlying Qwen3.5 `save_pretrained` regression; 45357 is a specific manifestation of the broader issue in 45216." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to the MobileLLM-125M load key error." + "reason": "Tokenizer add_special_tokens conflict and MobileLLM key error are different loading failures." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Different v5 regressions: FP8 config failure vs unexpected `_is_hf_initialized` init_empty_weights argument." + "reason": "FineGrainedFP8Config failure and unexpected _is_hf_initialized argument are separate v5 issues." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "`_is_hf_initialized` flag breakage is unrelated to disabling mmap on Strix Halo." + "reason": "_is_hf_initialized flag breakage and mmap OOM avoidance are unrelated." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "Different failures: FineGrainedFP8Config vs missing Float8 storage type." + "reason": "FP8 config failure is not the same as missing Float8 storage object resolution." }, - { - "left": "issue:44062", - "right": "issue:44987", - "accept": false, - "reason": "AddedToken keyword collision is a tokenizer-construction bug, not the fast-model load failure." - }, - { - "left": "issue:43296", - "right": "issue:43531", - "accept": false, - "reason": "Different models and failure modes: PaddleOCR-VL load failure vs Qwen3-MoE sliding-window behavior." - }, - { - "left": "issue:43404", - "right": "issue:45127", - "accept": false, - "reason": "Both concern tied weights, but one is missing weight tying and the other is LoRA merge collapse after vocab extension." - }, - { - "left": "issue:43531", - "right": "issue:44877", - "accept": false, - "reason": "Qwen3-MoE sliding-window bug is unrelated to granite_speech strict-config loading." - }, - { - "left": "issue:43611", - "right": "issue:44756", + { + "left": "issue:44062", + "right": "issue:44987", "accept": false, - "reason": "Base-model-prefix loading regression is unrelated to mmap OOM on Strix Halo." + "reason": "AddedToken special-argument conflict and model loading failure for physical-intelligence/fast are unrelated." }, { - "left": "issue:36032", - "right": "issue:41720", + "left": "issue:43296", + "right": "issue:43531", "accept": false, - "reason": "Tokenizer load conflict is unrelated to Qwen3 auto device mapping cuda asserts." + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window behavior are different bugs." }, { - "left": "issue:43425", - "right": "issue:44403", + "left": "issue:43404", + "right": "issue:45127", "accept": false, - "reason": "Torch 2.10 incompatibility is unrelated to transformer load noise." + "reason": "Both mention embeddings/tied weights, but missing tie in Mistral3ForConditionalGeneration is not the same as LoRA merge collapse with extended vocab." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen VL regressions, but one is broad video-input breakage and the other is a specific still-image rope-index scaling bug." + "reason": "Qwen-VL-related titles alone are not enough to show the same underlying bug: one is a broad video-input regression, the other a still-image temporal position-id scaling issue." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity on GPT-OSS and Whisper batch_decode skip_special_tokens are unrelated." - }, - { - "left": "issue:43792", - "right": "issue:44589", - "accept": false, - "reason": "Whisper runtime failure is unrelated to the Float8 storage lookup error." + "reason": "High perplexity on GPT-OSS is unrelated to Whisper batch_decode ignoring skip_special_tokens." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative on the soft edges. The only accepted overlap, 45216/45357, is plausibly the same Qwen3.5 save_pretrained serialization regression, and the rest of the pairwise rejections are appropriately model/code-path specific." + "accept": false, + "feedback": "The review is mostly conservative, but one accepted soft-edge verdict overstates the evidence. `issue:44479` and `issue:45325` are both Qwen-VL-related, yet the packet only shows titles: one is a broad video-input regression across several Qwen VL variants, while the other is a still-image temporal-position-id scaling bug in Qwen2.5-VL. That is enough to say they may be related, but not enough to confidently call them the same underlying regression or a duplicate symptom. The Qwen3.5 `save_pretrained` pair (`issue:45216` / `issue:45357`) looks better grounded. Because the summary and canonical-anchor reasoning lean on the stronger claim for `44479` / `45325`, I would not accept as-is." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -3187,8 +3333,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -3241,6 +3387,7 @@ "issue:45923" ], "soft_pairs": [ + "issue:43792|issue:44589", "issue:33666|issue:40444", "issue:41553|issue:43582", "issue:43632|issue:44534", @@ -3254,8 +3401,8 @@ "issue:44462|issue:45812", "issue:43482|issue:45538", "issue:44829|issue:45479", - "issue:42175|issue:45542", "issue:43317|issue:43856", + "issue:42175|issue:45542", "issue:43482|issue:43525", "issue:43818|issue:45357", "issue:41720|issue:45081", @@ -3270,202 +3417,208 @@ "issue:43994|issue:44898", "issue:44079|issue:44534" ], - "prepared_review_unit_hash": "e48c88955d03923d0f6ef9406feb64ec1c3b0163857dbbb0dabdad59512da4f8", + "prepared_review_unit_hash": "b4274158f2e1b3a16eee3976b2ca3ce96bbc07e76aa1c4d2cf55160d5d4d0368", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35744, - "estimated_input_tokens": 8936, - "estimated_eval_tokens": 18128 + "soft_pair_count": 29, + "serialized_chars": 36040, + "estimated_input_tokens": 9010, + "estimated_eval_tokens": 18276 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated, with a few loose thematic overlaps around Transformers v5 regressions, tokenizer/loading behavior, and model-specific multimodal bugs. None of the soft pairs look like the same underlying issue closely enough to merge.", - "confidence": 0.78, - "canonical_issue_reason": "Best canonical issue candidate is issue:45812 because it is a clear, high-signal regression report with broad impact and strong duplicate-attracting value, while still being concrete about the failing code path.", + "summary": "This cluster is mostly a set of separate model- or backend-specific regressions that only share superficial themes like \u201cv5 breakage,\u201d \u201cloading fails,\u201d or \u201ctokenizer/model init issues.\u201d I did not find any soft pair that looks like the same underlying bug/change.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45020 is the best umbrella issue here: it\u2019s the broadest, most central report about recent Transformers versions breaking remote-code models, which matches the cluster\u2019s overall pattern of v5 regressions better than the narrower model-specific tickets.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest stand-alone issue in this set: specific, impactful, and framed as a silent regression affecting all Granite models, which makes it a good representative bug report.", + "best_issue_reason": "issue:45020 is the best representative issue for this mixed set because it captures the widest class of affected failures and is the most plausible anchor for related downstream regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Both are model-loading failures, but one is Whisper-specific and the other is a float8 storage/type error; they are different concrete bugs and code paths." + }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen-VL training failures, but they describe different model versions and different failure modes (multi-GPU training vs iterable dataset with multiple images)." + "reason": "Both involve Qwen VL training, but one is generic multi-GPU training and the other is an IterableDataset/multi-image prompt failure; not the same bug." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Unrelated problems: tokenizer loading error messaging for Voxtral vs a TypeError in an Apple Silicon allocator warmup function." + "reason": "These are unrelated: Voxtral tokenizer error messaging versus an Apple Silicon TypeError in caching_allocator_warmup." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Both mention Transformers v5 regressions, but one is about `_is_hf_initialized` and model init flow; the other is about non-persistent buffers being filled with junk." + "reason": "Both are Transformers v5 regressions, but one is about _is_hf_initialized and the other about non-persistent buffers; different mechanisms." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Different bugs: compiled flex attention on torch>=2.9 vs recent Transformers versions breaking remote-code models." + "reason": "Torch 2.9 compiled flex_attention failure is unrelated to remote_code model breakage." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise vs a tokenizer-loading regression for a specific model; no shared underlying code-path is evident." + "reason": "Both are loading-related, but one is generic loading noise and the other is a specific tokenizer-loading regression for one model family." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both involve tokenizer special tokens, but one is runtime post-processor updates and the other is a specific tokenizer not adding BOS/EOS in v5." + "reason": "Both touch special-token behavior, but one is runtime post-processor updates and the other is add_special_tokens not adding BOS/EOS; not the same concrete issue." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment and Gemma-4 FSDP2 attention state rebuilding are unrelated failure modes." + "reason": "Unrelated bugs: ModelOutput key assignment versus Gemma-4 FSDP2 shared_kv_states rebuilding." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing in eval mode is a runtime-mode issue; missing `supports_gradient_checkpointing` on a trust_remote_code variant is a model-definition issue." + "reason": "Gradient checkpointing in eval mode is a different issue from a missing supports_gradient_checkpointing flag on a trust_remote_code model." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Same model family, but different bugs: finetuning with multiple images vs incorrect temporal position scaling in rope index computation." + "reason": "Both are Qwen2.5-VL issues, but one is multi-image IterableDataset training and the other is rope-index scaling for still images; different code paths." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Both are loading regressions in newer Transformers, but one is an unexpected `_is_hf_initialized` argument and the other is a missing float8 storage type." + "reason": "Both involve loading/initialization regressions, but one is an unexpected _is_hf_initialized argument and the other is a float8 storage lookup failure." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Both concern tokenizer handling, but one ignores `tokenizer.json` while the other produces wrong token IDs for Granite models; not the same bug." + "reason": "Both concern AutoTokenizer behavior, but one ignores tokenizer.json while the other produces wrong token IDs for Granite models; not the same bug." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "Qwen2.5-GGUF loading failure and CLIPTokenizer model_max_length handling are unrelated." + "reason": "GGUF loading failure and CLIPTokenizer max-length behavior are unrelated." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Different training bugs: flash_attention_3 degeneracy vs zero loss for single-label classification with one label." + "reason": "Both cause bad training dynamics, but one is flash_attention_3 degeneration and the other is single-label classification with num_labels=1; different root causes." }, { - "left": "issue:42175", - "right": "issue:45542", + "left": "issue:43317", + "right": "issue:43856", "accept": false, - "reason": "Both touch TensorFlow-related packaging/backend setup, but one is missing tf in the torch extra and the other is an undefined tf backend error with tensorboard-only installs." + "reason": "Device-map/offload loading failure is unrelated to Qwen3 MoE memory inefficiency during training." }, { - "left": "issue:43317", - "right": "issue:43856", + "left": "issue:42175", + "right": "issue:45542", "accept": false, - "reason": "Device-map/offload loading failure and Qwen3 MoE memory inefficiency are not the same code-path problem." + "reason": "Both are packaging/backend issues around TensorFlow, but one is missing TF in a torch extra install and the other is tensorboard installed without TF causing backend errors; not clearly the same bug." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Loading Qwen2.5-GGUF in v5 and missing `pad_token_id` on Llama4Config are different compatibility issues." + "reason": "Different model-loading problems: Qwen2.5-GGUF failure versus Llama4Config missing pad_token_id." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA architecture/state issues and Qwen3.5 save_pretrained key regression are unrelated." + "reason": "Completely different model families and failures: Video-LLaVA tower issues versus Qwen3.5 save_pretrained key regression." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "CUDA assert on auto device mapping and a Mistral tokenizer regex patch crash are different subsystems and failure modes." + "reason": "An A800 cudaErrorAssert in Qwen3 device mapping is unrelated to a Mistral tokenizer regex patch crash." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both involve batching/padding behavior, but MIMI encoder output mismatch and Gemma3 NaN embeddings stem from different model/attention bugs." + "reason": "Both mention padding/masking-related batching issues, but they affect different models and different failure modes; not the same bug." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and `_is_hf_initialized` init_empty_weights loading errors are unrelated." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the _is_hf_initialized init_empty_weights regression." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export compatibility and bfloat16 dtype mismatches are separate issues." + "reason": "Torch export failures with torch_compilable_check and dtype mismatches in bfloat16 inference are different code paths and bugs." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Packaging extras missing TensorFlow backend support vs vLLM failing to load PaddleOCR-VL on Transformers 4.53.0 are unrelated." + "reason": "Packaging/backend dependency issues versus a vLLM model-loading failure for PaddleOCR-VL are not the same problem." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are broad Transformers v5 regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; too different to merge." + "reason": "Both are Transformers v5 regressions in model initialization/loading, but the reported failures are different enough that they do not look like one concrete bug." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "`_is_hf_initialized` flag regression and dtype mismatches in inference do not share the same concrete fix path." + "reason": "_is_hf_initialized regressions and bfloat16 dtype mismatches are unrelated." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper model runtime failure and meta-device loading breakage are different problems." + "reason": "Whisper loading failure and torch.device(\"meta\") loading breakage are separate model-init issues with different symptoms." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Both are model-specific vision/ocr issues, but one is Qwen2.5-VL rope indexing and the other is Kosmos2.5 long OCR input index error." + "reason": "Qwen2.5-VL temporal position scaling and Kosmos2.5 long OCR index errors are different multimodal bugs." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "SigLIP2 wrong outputs with AutoModel/pipeline and Perceiver non-default resolution failure are distinct model bugs." + "reason": "SigLIP2 incorrect outputs with AutoModel/pipeline and Perceiver interpolation failures are unrelated model-specific issues." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer initialization are unrelated transformer internals." + "reason": "ModelOutput key assignment and non-persistent buffer filling are different internal regressions with no clear shared code path." } ] }, @@ -3484,8 +3637,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -3569,214 +3722,214 @@ "issue:43317|issue:43873", "issue:43873|issue:45701" ], - "prepared_review_unit_hash": "408c7af9d65317c72ff2d89342cb0132a20a8dadfa1624f7462b777a0606cf60", + "prepared_review_unit_hash": "1ac53a031010810160be5b850d2ee3fb59724744ccfdfe89b9866a65a448a4e5", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36293, - "estimated_input_tokens": 9074, - "estimated_eval_tokens": 18404 + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is heterogeneous; most soft pairs are only loosely related by subsystem/theme, not the same underlying bug. I would not merge any of the proposed soft edges, and I\u2019d use one broad tokenizer-regression report as the closest umbrella issue.", - "confidence": 0.58, - "canonical_issue_reason": "Issue 45920 is the best umbrella issue here: it is the broadest open, user-facing regression report and explicitly spans multiple models/tokenizers, making it the strongest representative if one item must be picked.", + "summary": "These items are mostly unrelated model- or subsystem-specific regressions that only share broad vocabulary (tokenization, quantization, device_map, Qwen/VL, etc.). I found no pair that looks like the same concrete bug/change, so the safe triage result is to keep them separate.", + "confidence": 0.96, + "canonical_issue_reason": "issue:43873 is the broadest concrete regression in the set: it is open, has inbound references, and clearly describes an offloading/quantization failure path, making it the most plausible representative issue for this heterogeneous cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the strongest global representative because it captures a broad v5 tokenizer regression affecting several models, whereas the rest of the items are mostly narrow, unrelated failure reports.", + "best_issue_reason": "issue:43873 is the best global representative because it is reproducible, reasonably specific, still open, and has the strongest signal of being a user-facing regression among the listed issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both involve tied `lm_head` behavior, but one is a missing weight-tying bug and the other is device-dependent serialization inconsistency; not the same concrete code-path problem." + "reason": "Related theme (tied lm_head weights), but one is a config/tie-weights generation bug and the other is a serialization inconsistency; not the same concrete failure." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Different failures: multi-image IterableDataset finetuning vs Qwen2.5-VL video position-id regression." + "reason": "Different modalities and failure modes: iterable-dataset multi-image finetuning vs video position-id regression." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Unrelated issues: FineGrainedFP8Config on Qwen3MoE vs GGUF loading failure on transformers v5." + "reason": "Completely different models and bugs: FP8 config failure vs GGUF loading failure." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Both are in Qwen2.5-VL position-id logic, but one is still-image temporal scaling and the other is video vision_position_ids; too different to treat as the same bug." + "reason": "Both touch Qwen2.5-VL position handling, but one is still-image temporal scaling and the other is video vision_position_ids; too different to merge as one bug." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer initialization is unrelated to ModelOutput key assignment when a prior value was None." + "reason": "Unrelated: non-persistent buffers filled with junk vs ModelOutput key assignment behavior." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "One is a specific PaddleOCR-VL/vLLM loading failure; the other is a broad torch 2.4.1 / transformers 5.8 incompatibility report." + "reason": "Different layers and symptoms: model loading in vLLM vs torch/transformers version incompatibility." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "Different areas: CLIPVisionModel hidden_states regression vs ModelOutput key bookkeeping." + "reason": "Hidden_states being dropped in CLIPVisionModel is unrelated to ModelOutput key assignment semantics." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Both are model-loading/forward issues, but the concrete failures differ: video-input regression vs device mismatch under auto device mapping." + "reason": "Different regressions: video-input handling in Qwen VL models vs Gemma4 PLE device mismatch." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to Qwen2.5-VL multi-image finetuning." + "reason": "Both are generation/training-related, but one is progressive generation with inputs_embeds/past_key_values and the other is multimodal finetuning with multiple images." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU-offload device mismatch and `torch.export` failures with `torch_compilable_check` are different problems." + "reason": "CPU offload device mismatch is not the same bug as torch.export failures around torch_compilable_check." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument during init-empty-weights loading is not the same as strict granite_speech config loading." + "reason": "Different bugs: unexpected _is_hf_initialized argument vs strict config blocking granite_speech loading." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Both mention memory, but one is an int4 quantization regression and the other is a Sam3Video OOM report; not the same bug." + "reason": "Both mention OOM, but one is reserved-memory growth under int4 quantization and the other is a model-specific CUDA OOM." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Different model-loading failures with no shared concrete code path." + "reason": "Both are model loading failures, but they involve different models and likely different root causes." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope-index / position-id computation." + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL rope/position-id handling." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Tokenizer special-token insertion regression and wrong token IDs are related only at a high level, but not the same failure." + "reason": "Both are tokenizer regressions, but they affect different tokenizer/model families and appear to be separate defects." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "CPU offload device mismatch is unrelated to DAC latent reconstruction mismatch." + "reason": "Device offload mismatch is unrelated to DAC.from_latents not matching forward pass." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both are device-mapping problems, but they affect different models and failure modes, so they should stay separate." + "reason": "Both involve device_map, but the concrete failure modes and models differ too much to treat as one bug." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crashes on NER/padding are unrelated to `torch.export` failures from `torch_compilable_check`." + "reason": "Tokenizer crashes on NER/batching and torch.export with torch_compilable_check are unrelated issues." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Tied-weight generation corruption and Bamba torch.compile SDPA failure are different code paths." + "reason": "Different problems: tied weights / garbage generation vs torch.compile+SDPA failure." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting has nothing to do with PEFT adapter checkpoint loading." + "reason": "Audio chat-template sample-rate defaulting is unrelated to PEFT checkpoint loading for GraniteSpeech." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "Documentation example errors are unrelated to finetuning failures with multiple images per prompt." + "reason": "Documentation/example errors for SigLIP2 are unrelated to Qwen2.5-VL multi-image finetuning failures." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Perplexity evaluation regression and Whisper empty-transcription generation are unrelated." + "reason": "Different domains and root causes: perplexity regression on GPT-OSS vs Whisper empty-transcription generation failure." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder is unrelated to OmDet-Turbo\u2019s processor/model resolution mismatch." + "reason": "Missing padding-mask support in MIMI encoder is unrelated to OmDet-Turbo processor/model input-size mismatch." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "AyaVision tied-weight generation bug is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Both involve model behavior under certain setups, but one is tied weights and the other is dtype mismatch in inference." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Missing `pad_token_idx` in StableLmConfig and `from_config` failing to unwrap `text_config` are different config-loading regressions." + "reason": "Unrelated: missing pad_token_idx in StableLmConfig vs composite multimodal config unwrapping." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Qwen3MoE weight conversion with accelerate+deepspeed is unrelated to a missing `supports_gradient_checkpointing` flag." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is a different failure than missing supports_gradient_checkpointing on a trust_remote_code model." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "High perplexity on gpt-oss and ineffective gradient checkpointing on PEFT LoRA do not share a concrete bug." + "reason": "Perplexity regression and gradient-checkpointing/LoRA ineffectiveness are unrelated." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "MIMI padding-mask behavior and Perceiver non-default-resolution failure are separate model-specific issues." + "reason": "Different models and bugs: MIMI padding-mask output mismatch vs Perceiver non-default-resolution failure." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both involve offloading/quantization, but one is a dequantized model load failure and the other is a broader offloading-not-working report without the same concrete path." + "reason": "Both concern offload/quantization, but one is loading a dequantized model with device_map=auto and the other is a broader offloading-not-working report; not the same concrete bug." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Generic offloading/quantization issues are unrelated to a tokenizer-version regression." + "reason": "Broad offloading/quantization issue vs tokenization changes; no shared concrete failure path." } ] }, @@ -3795,8 +3948,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -3885,238 +4038,238 @@ "issue:41720|issue:42617", "issue:41720|issue:45161" ], - "prepared_review_unit_hash": "fcdf464a44b5b6f6a9fe9dba1f521db6d7639ee5001fba0721d475d7a57a63d3", + "prepared_review_unit_hash": "0cf55e411c7c99347dce64e8fdf7ada79a028d09dbd6e9ab99b30e8975fdfa82", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37960, - "estimated_input_tokens": 9490, - "estimated_eval_tokens": 19236 + "serialized_chars": 37954, + "estimated_input_tokens": 9489, + "estimated_eval_tokens": 19234 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is heterogeneous: the soft-similarity pairs mostly link unrelated issues across different models, features, and failure modes. I found no true duplicate pair to merge.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "This cluster is a mixed bag of mostly unrelated Transformers issues. The closest shared theme is loading/offloading/quantization regressions, but none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43873 is the broadest representative of the loading/offloading + quantization subgroup and is already open with active discussion/inbound references.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the broadest representative of the loading/offloading/quantization subgroup, but it still does not subsume the rest of the cluster.", + "best_issue_reason": "issue:43873 best matches the cluster\u2019s dominant technical theme and is the most plausible umbrella representative.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "Different bugs: BeitImageProcessorFast label reduction vs Molmo missing all_tied_weights_keys." + "reason": "Different subsystems and failures: image processor label reduction vs missing tied-weights attribute on a model." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Different model/code paths: Mistral3 lm_head tying vs BitNet packed-weight unpacking during accelerate loading." + "reason": "Both involve loading, but they target different models and distinct code paths (weight tying vs packed-weight unpacking)." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "Unrelated failures: glm-4v loading vs tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Unrelated failures: one is a model load failure, the other is tokenizer loading for a different model family." }, { "left": "issue:43296", "right": "issue:44991", "accept": false, - "reason": "Different stack and symptom: PaddleOCR-VL load in vLLM vs tokenizer loading regression." + "reason": "Different stack and symptom: vLLM/PaddleOCR-VL load failure vs tokenizer regression for est-roberta." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Both are Qwen3-family, but one is auto device mapping cuda assert and the other is flash-attention generation; not the same concrete bug." + "reason": "Same model family, but different bugs: device-map CUDA assert vs flash-attention generation regression." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Different serialization issues: lm_head.weight inconsistency vs wrong visual encoder keys on save_pretrained." + "reason": "Both touch save/serialization, but one is lm_head tying/serialization by device and the other is incorrect visual encoder keys." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "Different problems: 3d_parallel.py execution failure vs Qwen3Moe FineGrainedFP8Config load failure." + "reason": "Different failures entirely: 3d_parallel.py execution vs Qwen3Moe FP8 config load." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device mapping cuda asserts." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device-map CUDA asserts." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Different models and root causes: Gemma3n variable collision vs Gemma4 PLE device mismatch." + "reason": "Both are Gemma-related, but the bugs are different: variable name collision vs device mismatch during forward." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Unrelated issues: BLIP2 dtype not propagating vs Trainer resume_from_checkpoint max_steps calculation." + "reason": "Different areas: dtype handling in Blip2 vs Trainer resume/max_steps calculation." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "BarkProcessor voice_preset bug is unrelated to progressive generation with inputs_embeds." + "reason": "Bark voice preset issue is unrelated to progressive generation with embeddings and cache state." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Different config/loader problems for different model types; no shared underlying defect." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batch-shape bug vs tokenization/version-change report; different subsystems." + "reason": "Video processor batching shape bug is unrelated to a general tokenization change report." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to dtype mismatches in inference." + "reason": "Tokenizer padding/truncation crash and bf16 dtype mismatches are different bugs." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "Different model bugs: torch.compile with SDPA vs packed-weights unpacking during loading." + "reason": "Different code paths: torch.compile/SDPA behavior vs accelerate loading of packed BitNet weights." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Different model families and failures: lm_head tying vs incorrect visual encoder keys on save_pretrained." + "reason": "Weight tying at model init is not the same as incorrect save_pretrained key serialization." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Trainer API signature mismatch is unrelated to Qwen3Moe FineGrainedFP8Config loading." + "reason": "Trainer API signature mismatch vs model loading failure; unrelated issues." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "compute_loss argument mismatch is unrelated to the 3d_parallel.py failure." + "reason": "Unexpected trainer keyword argument is unrelated to 3d_parallel execution failures." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both are multimodal serving crashes, but the concrete failures differ: str.to vs missing _tokenizer." + "reason": "Continuous-batching multimodal crash and missing processor attribute are different failures." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both involve memory, but int4 reserved-memory OOM and apex FusedRMSNorm leak are different root causes." + "reason": "Both are memory-related, but one is quantization OOM and the other is an apex RMSNorm memory leak." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Related to offloading/memory, but one is quantization offload behavior and the other is Strix Halo mmap OOM; not the same bug." + "reason": "Both involve memory/offload concerns, but one is quantization offloading and the other is mmap on Strix Halo causing OOM." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Different models and issues: Bamba torch.compile SDPA vs Bark-small CPU-offload device mismatch." + "reason": "Different models and mechanisms: SDPA/torch.compile vs CPU-offload device mismatch." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both concern image sizing, but OmDet-Turbo processor output mismatch and Perceiver interpolation failure are distinct bugs." + "reason": "Both concern image resolution, but one is a processor output-size bug and the other is a Perceiver interpolation failure." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading is unrelated to non-persistent buffers being filled with junk." + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk are distinct loading/runtime issues." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer NER/batching crash is unrelated to OmDet-Turbo processor/model image-size mismatch." + "reason": "Tokenizer crash vs vision processor size mismatch; unrelated." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision for trust_remote_code is unrelated to read-only permission propagation in save_pretrained." + "reason": "Both are filesystem/path related, but cache-path collisions and read-only permission propagation are different bugs." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Segfault loading Llava Next is unrelated to sinq quantized model save/load breakage." + "reason": "Llava Next load segfault and sinq save/load breakage are unrelated." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression is not the same as generic tokenization differences across versions." + "reason": "The former is a specific processor/tokenizer regression; the latter is too broad and not clearly the same bug." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and gpu+cpu offload dequantized-model loading are different quantization/loading paths." + "reason": "Qwen3Moe FP8 load failure and dequantized device_map offload failure are different quantization/load paths." }, { "left": "issue:44479", "right": "issue:44918", "accept": false, - "reason": "Video-input regression for Qwen VL models is unrelated to unpacking input embeddings in TRL SFT training." + "reason": "Video-input regression for several Qwen models is not the same as TRL unpacking of input embeddings." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Different models and failure modes: cudaErrorAssert on Qwen3 device mapping vs Llava Next segfault." + "reason": "Different model families and error modes: device-map assert vs segmentation fault on Llava Next load." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "High perplexity on GPT-OSS-20B is unrelated to tokenization changing across versions." + "reason": "Perplexity anomaly on GPT-OSS is not the same as a generic tokenization-change report." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device-map cuda assert is unrelated to 3d_parallel.py execution failure." + "reason": "Qwen3 auto device mapping and 3d_parallel execution are unrelated." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Different problems: cuda assert on A800 vs tensor-parallel only not working for GPT-OSS MoE." + "reason": "Both mention large-model parallelism, but the bugs are different models and different parallelism issues." } ] }, @@ -4135,8 +4288,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -4218,202 +4371,202 @@ "issue:43421|issue:43927", "issue:44898|issue:45072" ], - "prepared_review_unit_hash": "6f9381ce7a81fca1f8553acd2bb5ff2da2077e9aa885eac070188a2fb7127f64", + "prepared_review_unit_hash": "1a2a74180d357118031690d9d6db50b2d7f9ec9a7662ce21466dc6979164a313", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35919, - "estimated_input_tokens": 8980, - "estimated_eval_tokens": 18216 + "serialized_chars": 35910, + "estimated_input_tokens": 8978, + "estimated_eval_tokens": 18212 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a mix of unrelated transformer issues. Only one soft pair is a true duplicate: the two Qwen2_5_VLProcessor.apply_chat_template crash reports (44514/44545).", - "confidence": 0.96, - "canonical_issue_reason": "issue:44545 is the best canonical issue: it is the clearer duplicate of issue:44514, has the same concrete crash scenario (`Qwen2_5_VLProcessor.apply_chat_template` with `padding=False` on batched input), and has slightly more follow-up activity.", + "summary": "This cluster is mostly heterogeneous: the only clear duplicate is the Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False (issues 44514 and 44545). The other soft pairs look like adjacent-but-distinct bugs affecting different models, code paths, or failure modes.", + "confidence": 0.95, + "canonical_issue_reason": "Pick issue:44514 as the canonical issue: it is the earlier, clearer report of the same exact batched padding=False crash later repeated in issue:44545.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44545 is the strongest representative issue in this set because it is the most clearly scoped, reproducible bug report and matches a true duplicate.", + "best_issue_reason": "Issue:44514 is the strongest representative because it has the same concrete repro as the duplicate, with a precise API/method target and a clear failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "Different failures: a GenerationMixin import error vs. a script execution/runtime problem." + "reason": "ImportError for GenerationMixin and a 3d_parallel execution failure are different problems with no shared code path evidence." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "Unrelated domains: BeitImageProcessorFast label reduction vs. glm-4v-9b model loading failure." + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are unrelated model/component issues." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer-related, but one is a Mistral regex patch crash and the other is Granite token ID regression; different bugs and fixes." + "reason": "Mistral regex patch crash in tokenizer loading and Granite token-id regression are different tokenizer bugs." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Different model/config issues: Qwen3-MoE sliding_window vs. StableLmConfig missing pad_token_idx." + "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx regression are distinct config/model issues." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Both are config serialization discrepancies, but they affect different models and fields; not the same bug." + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload regression are separate config serialization problems." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Different Qwen multimodal loading/processing failures with different error paths and model variants." + "reason": "Video-input ValueError in Qwen3OmniMoe and Qwen3VL weight_only load error are different failure paths." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM regression and Qwen recurrent-state reset are unrelated code paths." + "reason": "CamemBERT masked LM regression and Qwen3.5 recurrent-state reset bug affect different models and mechanisms." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Different tokenizer/task crashes in different models; no clear shared root cause." + "reason": "Both are tokenizer/task failures, but one is LayoutLMv2 NER padding/truncation and the other is MLuke task AttributeError; not the same bug." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Both concern tokenizers, but one is runtime post-processor updates and the other is incorrect Granite token IDs; distinct issues." + "reason": "Runtime post-processor update for special tokens and Granite AutoTokenizer wrong IDs are related to token handling but are not the same concrete defect." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "General slow/fast tokenizer inconsistency vs. a specific T5 tokenizer load-time AttributeError." + "reason": "Slow/fast tokenizer inconsistency is not the same as T5Tokenizer's add_special_tokens method conflict on load." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Different API surfaces and failures: `transformers serve` processor init vs. chat/completions streaming inference." + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are unrelated processor/runtime bugs." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Image processor void segmentation maps and BarkProcessor voice preset are unrelated processor bugs." + "reason": "Void segmentation map processing and BarkProcessor voice_preset failure are different processor issues." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Different subsystems: torch.compile/SDPA failure vs. tokenizer special-token registration bug." + "reason": "Bamba torch.compile/SDPA failure and BigBird mask-token decode issue do not share the same code path." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both involve loading/offloading, but one is quantization offload mismatch and the other is packed-weight unpacking during accelerate loading." + "reason": "CPU offload device mismatch for bark-small and BitNet packed-weight unpacking during accelerate loading are separate loading/offload bugs." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Tokenization behavior changes, but one is runtime special-token/post-processor handling and the other is a broader version-related tokenization change." + "reason": "Special-token post-processor refresh and generic tokenization-version changes are too broad and do not indicate one bug." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Llava Next segmentation fault are unrelated." + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated model-loading problems." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Different root causes: offloading/quantization vs. tokenizer special token registration." + "reason": "bark-small CPU offload mismatch and BigBirdTokenizer mask-token registration are unrelated." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Different model families and different failure modes; not mergeable as one bug." + "reason": "Both involve Qwen3 variants, but one is sliding_window behavior and the other is from_config text_config unwrapping; different code paths." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A cache-strategy typo and removed pipeline-task docs are unrelated." + "reason": "A cache-strategies typo and docs mentioning removed pipeline tasks are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counter update bug vs. progressive generation with inputs_embeds/past_key_values are different issues." + "reason": "TrainerState token counting and progressive generation with inputs_embeds/past_key_values are distinct issues." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Metric gathering label-drop bug and SAM3 text/bbox behavior are unrelated." + "reason": "gather_for_metrics label dropping and SAM3 text/bounding-box behavior are different subsystems." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same concrete bug report: `Qwen2_5_VLProcessor.apply_chat_template` crashes on batched input when `padding=False`." + "reason": "Same exact crash: Qwen2_5_VLProcessor.apply_chat_template on batched input with padding=False, reported twice." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Import warning for trailing-slash local paths vs. undefined video-branch variables in multimodal token counting." + "reason": "Local-path trailing-slash import warning and multimodal token-count video branch bug are unrelated." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization offload behavior and a T5 apex RMSNorm memory leak are unrelated." + "reason": "Quantization/offloading behavior and T5 apex FusedRMSNorm memory leak are different problems." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Different models and failure modes: Qwen3 auto device mapping CUDA assert vs. GraniteSpeech PEFT checkpoint loading." + "reason": "Qwen3 auto device mapping cuda assert and GraniteSpeech PEFT local checkpoint loading are not the same bug." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Both are config/tokenizer persistence bugs, but one is runtime special-token updates and the other is DiaConfig losing custom token IDs on save/load." + "reason": "Both touch token IDs, but one is runtime post-processor updates while the other is save/load loss of custom token IDs." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different failure classes: Perceiver positional interpolation on non-default resolutions vs. dtype mismatch in bfloat16 inference." + "reason": "Perceiver interpolation failure and bf16 dtype mismatch in SwitchTransformers/TimmWrapperModel are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet: most pairs are clearly different bug reports, and the only accepted soft edge, 44514/44545, is well supported by nearly identical titles and the same Qwen2_5_VLProcessor crash scenario. The other verdicts stay conservative and do not overclaim duplication." + "feedback": "The summary is well grounded in the packet and appropriately conservative. The only clear duplicate is issue:44514 / issue:44545, which have identical titles and matching crash context. The other soft-edge rejections stay on the cautious side and do not overstate duplication evidence." }, "evaluator_used": true, "retried": false, @@ -4429,8 +4582,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -4513,20 +4666,20 @@ "issue:30064|issue:36010", "issue:43720|issue:43819" ], - "prepared_review_unit_hash": "23dca3be36d03dcb215ab26cf4f3fc7d1d4d5af3208743e1a5516fdabddf959c", + "prepared_review_unit_hash": "1d9e1b3e17328b089b98f3816f881e344fcc10e925421570d2bac92e8b234dee", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36090, + "serialized_chars": 36091, "estimated_input_tokens": 9023, "estimated_eval_tokens": 18302 }, @@ -4535,192 +4688,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the items span unrelated models, tokenizer bugs, training/runtime regressions, and documentation issues. I only see one plausible duplicate pair around HF_HUB_OFFLINE tokenizer loading failures (42886/44843); the rest do not look like the same underlying bug.", - "confidence": 0.83, - "canonical_issue_reason": "Issue 42886 is the best local canonical: it states the broad offline-cache tokenizer failure, while 44843 appears to be a more specific root cause involving _patch_mistral_regex and model_info().", + "summary": "The items are mostly unrelated issue reports across different models/features. Only one soft edge looks like a true duplicate/same-bug pair: the HF_HUB_OFFLINE tokenizer cache failure and the Mistral regex/offline model_info call. Everything else appears to be distinct root causes or different model-specific regressions.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44843 is the clearest root-cause report in the accepted duplicate pair: it identifies the specific offline break in `_patch_mistral_regex` via an unconditional `model_info()` call, which is more actionable than the broader symptom report.", "canonical_pr_reason": null, - "best_issue_reason": "42886 is the most representative issue because it captures the broader user-facing regression and cleanly subsumes the more specific 44843 symptom.", + "best_issue_reason": "issue:44843 is the strongest representative issue overall because it names the concrete failing code path and the exact offline mode regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen-related, but one is a multimodal config/unwrapping bug and the other is a CPU-offload rotary embedding issue; different code paths." + "reason": "Both are Qwen3.5-related, but one is a config unwrapping bug and the other is a CPU-offload rotary embedding bug; different code paths." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading failures, but they affect different model repos and likely different tokenizer regressions." + "reason": "Both are tokenizer-loading regressions, but they affect different models and likely different failure causes; too broad to be the same bug." }, { "left": "issue:33666", "right": "issue:35707", "accept": false, - "reason": "Multi-GPU Qwen2-VL training is unrelated to progressive generation with inputs_embeds and past_key_values." + "reason": "One is Qwen2-VL multi-GPU training and the other is progressive generation with inputs_embeds/past_key_values; unrelated problems." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and a Float8 storage import/storage error are unrelated failures." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the Float8 storage import/runtime error." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "A flash_attention crash on sink-less models is not the same as a Qwen multimodal config unwrapping bug." + "reason": "Flash-attention sink-less crash and Qwen multimodal config unwrapping are different failures." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "One is a model/tokenizer loading regression, the other is an image-processor torchvision dependency issue." + "reason": "Both are loading/processor regressions, but one is a model loading failure and the other is an image processor torchvision dependency issue." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "AddedToken special-argument duplication is a different tokenizer bug than failing to load a specific tokenizer repo." + "reason": "AddedToken special-argument error is not the same as a model tokenizer loading failure." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "A tie_word_embeddings warning during LoRA fine-tuning is unrelated to incorrect visual-encoder keys on save_pretrained." + "reason": "One is a tie_word_embeddings warning during LoRA fine-tuning; the other is a visual-encoder key save regression." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both mention processor/model mismatch, but they concern different models and different failures." + "reason": "Both mention processor/model mismatch, but they concern different models and different concrete failures." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Both touch special-token handling, but one is runtime post-processor updates and the other is BigBird mask-token registration/empty decode output." + "reason": "Both involve special tokens, but one is a runtime post-processor update feature gap and the other is a BigBirdTokenizer decode bug." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "A T5 RMSNorm memory leak and a missing supports_gradient_checkpointing flag are unrelated." + "reason": "RMSNorm memory leak in T5 is unrelated to a missing supports_gradient_checkpointing flag." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "Gemma-4 FSDP2 shared_kv_states KeyError is unrelated to the Nemotron trust_remote_code flag issue." + "reason": "Gemma-4 FSDP2 shared_kv_states bug is unrelated to Nemotron trust_remote_code metadata." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "These are both trainer-related, but one is unpacking Qwen3.5 inputs_embeds in SFT and the other is train/evaluate lifecycle behavior." + "reason": "One is an inputs_embeds/past_key_values unpacking bug; the other is a train/evaluate lifecycle issue." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Gradient checkpointing in eval mode and abnormal gradient growth under ZeRO-3 are different bugs." + "reason": "Gradient checkpointing in eval mode and abnormal gradient growth under ZeRO-3 are different training issues." }, { "left": "issue:42886", "right": "issue:44843", "accept": true, - "reason": "Both describe the same offline tokenizer-loading regression; 44843 gives a likely root cause in _patch_mistral_regex/model_info()." + "reason": "Both describe the same offline tokenizer-loading regression; 44843 gives the specific root cause (`model_info()` called unconditionally in `_patch_mistral_regex`)." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping cudaErrorAssert are unrelated." + "reason": "BarkProcessor voice preset failure and Qwen3 auto device-mapping cuda assert are unrelated." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "A fused RMSNorm memory leak is not the same as Qwen3.5 MoE CPU-offload incompatibility." + "reason": "RMSNorm memory leak in T5 is unrelated to Qwen3.5 CPU-offload rotary embedding behavior." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Special-token post-processor refresh and LayoutLMv2 NER/padding crashes are distinct tokenizer failures." + "reason": "Runtime special-token post-processing and LayoutLMv2 NER padding/truncation crashes are different tokenizer issues." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Missing pad_token_id on Llama4Config and Qwen3-MoE sliding_window behavior are distinct config/model bugs." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "Eval-mode gradient checkpointing and an apex RMSNorm memory leak are not the same issue." + "reason": "Gradient checkpointing eval-mode behavior is unrelated to T5's apex RMSNorm memory leak." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen VL issues, but one is from_pretrained weight_only loading and the other is video vision_position_ids." + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL video vision_position_ids are different multimodal regressions." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both mention fix_mistral_regex/_patch_mistral_regex, but one is a crash and the other is a codec-handling regression; not the same concrete bug." + "reason": "Both mention fix_mistral_regex/tokenizer loading, but one crashes on backend_tokenizer access while the other reports codec handling and warning issues; not the same bug." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 doc/example issues are unrelated to quantization/offloading behavior." + "reason": "SigLIP2 documentation errors and quantization/offloading behavior are not the same underlying issue." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Qwen3.5 num_labels propagation and a generic single_label_classification zero-loss bug are different configuration problems." + "reason": "num_labels propagation in Qwen3.5 and degenerate zero-loss classification across models are related only at a high level, not the same defect." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "A SigLIP2 inference quality issue and a CamemBERT masked-LM regression are unrelated." + "reason": "Incorrect masked-LM predictions in CamemBERT and nonsensical AutoModel/pipeline results for SigLIP2 are different model-specific regressions." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "A DeepSpeed ZeRO-3 BertModel initialization error is unrelated to _torch_extract_fbank_features()." + "reason": "Deepspeed ZeRO-3 init/BertModel index error is unrelated to `_torch_extract_fbank_features()`." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "A tokenizer regression in Kimi-K2.5 is unrelated to list_repo_files failing to forward revision in CI." + "reason": "Kimi tokenizer regression and `list_repo_files` missing revision forwarding are unrelated." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation-map processing and a GenerationMixin import error are unrelated." + "reason": "Void segmentation maps in image processing and GenerationMixin import errors are unrelated." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading and DAC.from_latents missing STE are different model-specific bugs." + "reason": "Packed-weight unpacking during accelerate loading and DAC.from_latents STE mismatch are different code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The 42886/44843 pairing is strongly supported by the titles, and the other soft-edge rejections avoid overclaiming duplicate status from weak similarity alone." + "feedback": "The summary is well-grounded in the packet, and the one accepted duplicate pair (42886/44843) is supported by the titles and conservative reasoning. The other soft-edge rejections are cautious and do not overclaim duplication from broad topical similarity alone." }, "evaluator_used": true, "retried": false, @@ -4736,8 +4889,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -4816,205 +4969,205 @@ "issue:30064|issue:36331", "issue:44448|issue:44671", "issue:43421|issue:44811", - "issue:43782|issue:44977", - "issue:39692|issue:45834" + "issue:39692|issue:45834", + "issue:43782|issue:44977" ], - "prepared_review_unit_hash": "9c0b46ef1288fd1c986ade471ae712a55a7ffbf66c756eba41793adfe8ccf69f", + "prepared_review_unit_hash": "0ea76ffc1dc20c6c18740151a57c7141b2bbb5450380c1aca0e3a6daf5a573b6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35909, - "estimated_input_tokens": 8978, - "estimated_eval_tokens": 18212 + "serialized_chars": 35917, + "estimated_input_tokens": 8980, + "estimated_eval_tokens": 18216 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a broad mix of unrelated Transformers bug reports across loading, tokenization, multimodal preprocessing, generation, and training/runtime behavior. No soft pair looks like the same underlying defect, including the superficially related Mistral regex cases which fail in different ways.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the set is heterogeneous across model families and subsystems, so there is no single underlying bug to anchor the cluster.", + "summary": "This cluster is very heterogeneous. I would reject all soft pairs as duplicates; they share at most a broad subsystem, not the same concrete bug/change. The most representative issue is the broad quantization/offloading report.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43873 is the broadest, most central report in the set, with the most discussion/inbound references and a cross-cutting loading/offloading/quantization scope.", "canonical_pr_reason": null, - "best_issue_reason": "No single best issue for deduplication: none of these reports is a reliable representative of the others.", + "best_issue_reason": "issue:43873 is the best umbrella issue here because it is active, broadly scoped, and closer to an infrastructure-level regression than the other model-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Both involve model loading/saving, but one is PEFT local-checkpoint loading and the other is sinq quantization round-trip; different failure modes and code paths." + "reason": "Both are loading-related, but one is PEFT adapter loading from local checkpoints and the other is sinq quantization save/load corruption; different models and different failure modes." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Unrelated issues: one is a config validation rule for patchtsmixer, the other is num_labels propagation in Qwen3.5." + "reason": "Different areas entirely: PatchTSMixer post_init policy vs Qwen3.5 num_labels propagation in text config." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Tokenizer special-token registration vs packed-weight unpacking during accelerate loading; different components and bugs." + "reason": "Tokenizer special-token registration bug vs BitNet packed-weight unpacking during accelerate loading; unrelated code paths." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction failure and tokenizer-mapping codegen failure are unrelated." + "reason": "Audio feature extraction helper failure vs tokenizer-mapping code generation failure; no shared underlying bug." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Both are multimodal, but the first is a Qwen3OmniMoe video-processing ValueError and the second is a Gemma3n variable-name collision in forward()." + "reason": "Different multimodal models and different root causes: Qwen3OmniMoe video processing error vs Gemma 3n variable-name collision." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Different regressions in different model outputs: CLIP attentions assignment vs RTDetrV2 output divergence." + "reason": "Both are v5 regressions, but one is CLIP output attentions assignment and the other is RTDetrV2 output drift; not the same concrete bug." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is not the same as bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Torch.compile/SDPA failure in Bamba vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel; different symptoms and paths." }, { "left": "issue:42915", "right": "issue:43746", "accept": false, - "reason": "Qwen3Moe FP8 config failure and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Qwen3Moe FineGrainedFP8Config failure and GraniteSpeech PEFT checkpoint loading are unrelated beyond both touching model loading." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Both touch loading/offloading, but one is PEFT local checkpoint restore and the other is quantization offloading behavior; not the same defect." + "reason": "Both involve loading/quantization, but one is PEFT checkpoint loading while the other is offloading behavior under quantization; not the same bug." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Different multimodal bugs: processor.tokenizer/image-tokenizer regression vs Kosmos2.5 long-OCR index error." + "reason": "Processor/tokenizer API regression vs long-OCR index error in Kosmos2.5; different subsystems and failures." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "torch.compile with SDPA in Bamba is unrelated to DAC.from_latents not matching forward() without STE." + "reason": "Bamba compile/SDPA bug is unrelated to DAC.from_latents missing-STE mismatch." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Generic input validation messaging bug vs single-label-classification zero-loss bug; different code paths." + "reason": "Both concern input/label handling, but one is XOR validation messaging and the other is zero-loss classification logic; distinct bugs." }, { "left": "issue:43388", "right": "issue:46032", "accept": false, - "reason": "Metrics gathering drops tuple labels in the last batch; Mamba2Mixer use_cache correctness is a separate model inference bug." + "reason": "Distributed gather_for_metrics label truncation is unrelated to Mamba2Mixer use_cache incorrect results." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Different model-specific multimodal errors: dtype mismatch in Mistral-3 preprocessing vs Gemma3n attribute collision." + "reason": "Different model families and different issues: image preprocessor/model dtype mismatch vs a forward-pass variable collision." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Sparse-layer creation bug in SwitchTransformersConfig and RoPE-layer truncation in Smollm3 are unrelated configuration/model-structure issues." + "reason": "Both are model-configuration mistakes, but they affect different architectures and different misconfigurations." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to Perceiver image-classification interpolation failures." + "reason": "Tokenizer NER/padding crash vs image-classification interpolation failure; unrelated." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration and DAC latent decoding mismatch are completely different bugs." + "reason": "BigBird tokenizer special-token registration bug has nothing to do with DAC latent decoding mismatch." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoConfig trust_remote_code field loss are separate issues." + "reason": "Sliding-window behavior in Qwen3-MoE vs AutoConfig trust_remote_code field loss; different components." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken constructor conflict vs model/tokenizer loading failure for cjvt/sleng-bert; not the same root cause." + "reason": "AddedToken special-argument conflict vs model loading failure for cjvt/sleng-bert; not the same root cause." }, { "left": "issue:43950", "right": "issue:45987", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained and StaticCache.get_seq_length() returning a tensor are unrelated." + "reason": "Both are core API regressions, but one corrupts non-persistent buffers on from_pretrained while the other returns the wrong type from StaticCache.get_seq_length()." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both hit _patch_mistral_regex, but one is an unconditional model_info() offline failure and the other is a backend_tokenizer AttributeError; different triggers and fixes." + "reason": "Same helper function, but different bugs: unconditional hub metadata lookup offline vs missing backend_tokenizer attribute crash." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "trust_remote_code missing-fields regression and strict granite_speech config loading are not the same bug." + "reason": "Both are config-loading issues, but one is missing fields with trust_remote_code and the other is strict config blocking granite_speech; different failures." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both are video-input issues, but they affect different Qwen model families and different internal logic." + "reason": "Both involve video/multimodal Qwen models, but the concrete bugs differ: video input ValueError vs incorrect vision_position_ids." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image segmentation-map processing and CustomTrainer.compute_loss argument handling are unrelated." + "reason": "Completely unrelated: image processor segmentation-map handling vs CustomTrainer compute_loss signature breakage." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Different v5 regressions in different models/tasks: Pegasus generation vs CamemBERT masked LM prediction." + "reason": "Both are model output regressions in v5, but on different models and different code paths." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime special-token/post-processor refresh and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer/processor issues." + "reason": "Tokenizer post-processor update bug vs Whisper batch_decode skip_special_tokens handling; related theme, different underlying issue." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:39692", + "right": "issue:45834", "accept": false, - "reason": "from_pretrained weight_only loading error and flash-attention generation failure are separate problems." + "reason": "SigLIP2 docs example errors and Kosmos2.5 OCR index error are separate model-specific problems." }, { - "left": "issue:39692", - "right": "issue:45834", + "left": "issue:43782", + "right": "issue:44977", "accept": false, - "reason": "Documentation/example bugs with quantization mismatch are unrelated to Kosmos2.5 long OCR indexing." + "reason": "Different Qwen variants and different failures: weight_only load error vs flash-attention generation bug." } ] }, @@ -5033,8 +5186,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -5117,202 +5270,202 @@ "issue:43526|issue:43697", "issue:43650|issue:43906" ], - "prepared_review_unit_hash": "b443ba68a3f60683470948054713eeb9d08938f45d35cc28460872c4f943ecfc", + "prepared_review_unit_hash": "5a846f64cfdcad9fc4ff2a566bbe40c6db1446198c4d5af9eef366e07c437590", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36115, - "estimated_input_tokens": 9029, - "estimated_eval_tokens": 18314 + "serialized_chars": 36112, + "estimated_input_tokens": 9028, + "estimated_eval_tokens": 18312 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a loose mix of transformer regression reports across tokenizers, multimodal models, loading paths, and docs. The soft pairs share broad topical similarity but do not look like the same underlying bug/change, so none should be merged as duplicates.", - "confidence": 0.83, - "canonical_issue_reason": "Issue 45020 is the broadest regression-style report here and works best as a triage umbrella for recent-version breakage reports, even though it is not a true duplicate target for most of the other items.", + "summary": "This cluster is highly heterogeneous; the soft-linked pairs mostly share only broad subsystem similarity and do not look like true duplicates of the same bug/change.", + "confidence": 0.9, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the most useful general anchor because it frames a concrete version-regression class rather than a narrow model-specific failure.", + "best_issue_reason": "issue:45701 is the broadest tokenizer-regression report and the closest thing to a representative anchor, but the cluster overall is too mixed to have a strong canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "Both are multimodal-model problems, but one is a dtype-loading issue for BLIP2 and the other is a Video-LLaVA architecture/weight-sharing defect; different code paths." + "reason": "Blip2 dtype remains float32 vs Video-LLaVA missing temporal attention / near-shared towers; different models and failure modes." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription behavior after special-token alignment is unrelated to Kosmos2.5 long-OCR index errors." + "reason": "Whisper empty transcription after align_special_tokens vs Kosmos2.5 OCR long-input index error; unrelated modalities and code paths." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and Jinja/template compilation errors are unrelated subsystems." + "reason": "AutoImageProcessor URL loading vs Jinja/template compilation error; unrelated loading vs template parsing issues." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next load-time segfault and an MPS int histogram op error are not the same bug." + "reason": "Llava Next segfault on model load vs MPS histogram not implemented for Int in MOE; different subsystems and symptoms." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "CPU-offload device mismatch in Bark-small is different from bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "CPU offload device mismatch in Bark-small vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel; both are device/dtype issues but not the same bug." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Qwen VL video-input regression and Kosmos2.5 long-OCR index error affect different model paths and inputs." + "reason": "Qwen VL video-input regression vs Kosmos2.5 OCR index error; different models and input pipelines." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/padding are unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + "reason": "LayoutLMv2Tokenizer crashes on NER/batched padding vs AutoTokenizer failing to load a FastSpeech2ConformerTokenizer; both tokenizer-related but not the same failure." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "A BarkProcessor voice_preset bug and a GenerationMixin import breakage are unrelated failures." + "reason": "BarkProcessor voice_preset failure vs GenerationMixin import error; completely different APIs." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "One is an incorrect pipeline error message about removed tasks; the other is a bad-output report for SigLIP2. Different issues." + "reason": "pipeline() translation-support error message vs Siglip2 producing nonsensical outputs; docs/error-message issue vs model behavior bug." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both mention recent-version breakage, but one is about removing is_torch_fx_available and the other is broad remote_code failures; not the same concrete change." + "reason": "Both involve remote_code breakage, but one is a specific removed helper (is_torch_fx_available) and the other is a broad recent-version regression; not the same concrete bug." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "StaticCache sequence-length contract bug and return_tensors being ignored when text_kwargs is passed are different API issues." + "reason": "StaticCache seq_length return-type violation vs return_tensors being ignored with explicit text_kwargs; unrelated API behaviors." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "A float16/bfloat16 model-loading bug and an Apple Silicon warmup TypeError are distinct." + "reason": "Blip2 dtype stays float32 vs Apple Silicon caching_allocator_warmup TypeError; no shared code path." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to docs still mentioning removed pipeline tasks." + "reason": "GLM-5 RoPE implementation discussion vs docs still mentioning removed pipeline tasks; unrelated implementation vs documentation issues." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "Both involve tokenizer behavior, but one is specifically extra spacing after added tokens while the other is a broader slow-vs-fast inconsistency." + "reason": "Additional space after added token vs slow/fast tokenizer inconsistency; related area, but not clearly the same underlying bug." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "OmDet input-size mismatch and bfloat16 dtype mismatches are different failures." + "reason": "OmDet-Turbo image-size mismatch vs dtype mismatch in a different model family; unrelated regressions." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask-shape IndexError during tensor indexing is not the same as trainer.evaluate() failing after training." + "reason": "Mask shape IndexError vs trainer.evaluate() failing after train(); different failure sites and APIs." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight loading in BitNet and dtype mismatch inference problems are different code paths." + "reason": "Packed weights not unpacked during accelerate loading vs bfloat16 dtype mismatch in inference; different loading/inference bugs." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "gguf_file loading failures and tokenizer loading failures for a specific model are separate bugs." + "reason": "gguf_file breaking tokenizer/model loading vs a specific tokenizer load failure for EMBEDDIA/est-roberta; both loading regressions, but not the same concrete issue." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Runtime error messaging and stale docs references around removed pipeline tasks are related, but they are not the same artifact or fix." + "reason": "Both relate to v5 pipeline removals, but one is an error message and the other is stale docs; not the same bug/change." }, { "left": "issue:43404", "right": "issue:43782", "accept": false, - "reason": "Mistral3 weight tying and Qwen3VL weight_only loading are model-specific loading bugs, not duplicates." + "reason": "Mistral3 lm_head weight tying bug vs Qwen3VL weight_only=True from_pretrained error; different models and root causes." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "AppleSilicon warmup TypeError and CUDA AMP crash in LwDetrImageLoss are unrelated." + "reason": "AppleSilicon allocator warmup TypeError vs CUDA AMP crash in LwDetrImageLoss; unrelated platforms and subsystems." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Package-install backend availability and Mistral regex patch crashes are unrelated." + "reason": "Missing TensorFlow backend in pip extra vs Mistral regex patch AttributeError; unrelated packaging vs tokenizer internals." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "An isolated reproduction of an older issue and an FP8 B200 inference failure are not the same bug." + "reason": "Isolated reproduction of an old issue vs FP8 inference failure on B200; no shared underlying bug." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral tokenizer regex patch failures and Gemma4Processor missing _tokenizer are different loading issues." + "reason": "Mistral regex patch failing on tokenizer internals vs Gemma4Processor missing _tokenizer in transformers serve; different objects and code paths." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT masked-LM prediction regressions and generic tokenization changes are too broad to treat as the same bug." + "reason": "CamemBERT masked-LM regression vs a broad report that tokenization changes across versions; one is model prediction behavior, the other a tokenizer-level complaint." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate defaults and DAC forward/from_latents mismatch are different audio bugs." + "reason": "Audio processor sampling-rate default mismatch vs DAC.from_latents / forward-pass mismatch; unrelated audio bugs." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences are unrelated vision bugs." + "reason": "BeitImageProcessorFast reduce_labels bug vs RTDetrV2 output changes in v5; both vision-related, but not the same concrete defect." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "The placeholder issue 'ADD THE DATA' is not a duplicate of the isolated reproduction report." + "reason": "Placeholder 'ADD THE DATA' issue vs an isolated reproduction report; no substantive overlap." } ] }, @@ -5331,8 +5484,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -5415,208 +5568,208 @@ "issue:43606|issue:45859", "issue:42490|issue:43257" ], - "prepared_review_unit_hash": "3ae8a06133aff652f8b4ff983e3960a0abdd783f876a42b67b1257d33a830e2a", + "prepared_review_unit_hash": "63d87f296b65d8c95569a02ffa41628308654ad2041bd163889cfda91399d8c6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36168, - "estimated_input_tokens": 9042, - "estimated_eval_tokens": 18340 + "serialized_chars": 36175, + "estimated_input_tokens": 9044, + "estimated_eval_tokens": 18344 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a loose collection of unrelated Transformers issues across many subsystems (audio, vision, loading, quantization, training, and config handling). None of the soft pairs look like the same underlying bug or change, so there is no clear canonical duplicate here.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "The cluster is loosely related by Transformers loading/processor regressions, but the soft pairs mostly describe different concrete bugs. I did not find any safe duplicate merges, and there are no PRs in the set.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45020 is the broadest and most central loading/regression report here, covering recent-version breakage around remote/custom model handling that overlaps the most with the rest of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a safe canonical for the cluster because the items span distinct code paths and model families; picking one would be arbitrary rather than duplicate-driven.", + "best_issue_reason": "issue:45020 is the best general exemplar because it is high-level, user-facing, and closest to the shared theme of recent Transformers compatibility regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Different bugs: tokenizer special-token decoding vs MXFP4 quantization fallback." + "reason": "Different subsystems: tokenizer special-token registration vs MXFP4 quantization loading." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Different tied-weights warnings in different Qwen variants; not the same bug." + "reason": "Both mention tied-embedding warnings, but they describe different Qwen warning paths and symptoms." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "MobileLLM key error vs broad remote_code regression; unrelated failure modes." + "reason": "MobileLLM key error is a model-specific load failure, not the same as broad remote_code regressions." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation error vs CLIPTokenizer max_length behavior; different components." + "reason": "Unrelated issues: template compilation error vs tokenizer max length constant." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Undefined video multimodal-token helpers vs stale auto_mappings CI breakage." + "reason": "Video token counting bug is unrelated to removed auto-mapping config references." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup failure vs PIL image processors incorrectly requiring torchvision." + "reason": "Storage-object lookup error is unrelated to PIL backend torchvision requirements." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT/local checkpoint loading issue vs Whisper batch_decode skip_special_tokens bug." + "reason": "PEFT local checkpoint loading vs Whisper decode flag handling are different code paths." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer junk values vs FSDP CPU RAM efficient loading breakage." + "reason": "Both are loading-related, but one is buffer initialization corruption and the other is FSDP RAM-efficient loading." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Zero3/non-pretrained BERT index error vs trainer.evaluate after train; different paths." + "reason": "Zero-3/BERT index error is unrelated to trainer evaluate-after-train behavior." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile+SDPA Bamba issue vs MXFP4 quantization fallback; not the same problem." + "reason": "torch.compile+SDPA failure is not the same as quantization fallback during load." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent reconstruction mismatch vs DiaConfig token-ID persistence bug." + "reason": "DAC latent reconstruction mismatch and DiaConfig token-id persistence are unrelated bugs." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "GenerationMixin import error vs Jupyter custom model initialization regression." + "reason": "Importing GenerationMixin and Jupyter custom-model initialization are different failures." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "Bark voice_preset issue vs CustomTrainer compute_loss signature change; unrelated." + "reason": "Bark voice_preset and CustomTrainer loss signature are unrelated issues." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Model loading/unit-test failure vs flash-attention generation bug; different models and code paths." + "reason": "GLM-4.7-Flash loading test failure is not the same as Qwen3.5 flash-attention generation failure." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation example errors/quantization failure vs MXFP4 load fallback; different scope." + "reason": "SigLIP2 doc-example mismatch/quantization errors are different from MXFP4 load fallback." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight unpacking during accelerate load vs OmDet-Turbo processor image-size mismatch." + "reason": "BitNet packed-weight unpacking and OmDet processor image-size mismatch are unrelated." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero probabilities vs processor output shape mismatch; unrelated behavior." + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to OmDet-Turbo input-shape mismatch." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Tokenizer error masking in protobuf fallback vs bare except swallowing errors in Fuyu conversion." + "reason": "Both concern exception handling, but they are different functions and different failure modes." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "AyaVision lm_head tying bug vs BLIP2 dtype loading bug; different model internals." + "reason": "AyaVision lm_head tying and Blip2 dtype propagation are separate model bugs." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Import path regression vs trainer API signature regression; not the same issue." + "reason": "ImportError for GenerationMixin and trainer compute_loss signature mismatch are unrelated." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Beit reduce_labels truncation bug vs MPS attention correctness issue; unrelated." + "reason": "Beit reduce_labels bug and MPS attention correctness issue are unrelated." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding-window bug vs PIL torchvision dependency regression; different subsystems." + "reason": "Qwen3-MoE sliding-window behavior and PIL torchvision dependency are different problems." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sample-rate default vs Whisper batch decoding special-token handling." + "reason": "Audio chat-template sample-rate defaulting is unrelated to Whisper batch_decode special-token handling." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM loading breakage vs init_empty_weights unexpected argument TypeError." + "reason": "FSDP CPU RAM-efficient loading and init_empty_weights argument handling are distinct loading regressions." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE accelerate/deepspeed weight conversion vs multimodal token counting undefined names." + "reason": "Accelerate/deepspeed weight conversion and multimodal token counting are different code paths." }, { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "AutoConfig trust_remote_code field loss vs AutoModelForCausalLM.from_config text_config unwrap bug." + "reason": "Both involve config/model loading, but one is missing fields from trust_remote_code and the other is composite-config unwrapping." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Undefined video token helpers vs batched video processor shape bug; related area but different defects." + "reason": "Undefined video-token helper calls and batched video processor shapes are not the same bug." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Bark CPU offload device mismatch vs Qwen3_5MoE rotary embedding CPU offload compatibility." + "reason": "CPU offload device mismatch in Bark is different from Qwen3_5MoE rotary embedding offload incompatibility." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local model import warning vs Qwen3 MOE accelerate/deepspeed conversion bug." + "reason": "Trailing-slash path warning and Qwen3 MOE weight conversion are unrelated." } ] }, @@ -5635,8 +5788,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -5724,20 +5877,20 @@ "issue:43257|issue:43696", "issue:34567|issue:36331" ], - "prepared_review_unit_hash": "33350b9d48a7df8d3b5864b14461db8cd8611ffbce1f14d2993da0a9e97b67dc", + "prepared_review_unit_hash": "92b018eae36bb2194fab2962c31dd7ca487a4886b306774885e0e6c40d8f9be5", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37765, + "serialized_chars": 37766, "estimated_input_tokens": 9442, "estimated_eval_tokens": 19140 }, @@ -5746,217 +5899,214 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft links are mostly false positives: the cluster spans unrelated tokenizer, video, model-loading, quantization, DeepSpeed, and dtype bugs. None of the paired issues look like the same underlying defect with a plausible one-PR merge path.", - "confidence": 0.95, - "canonical_issue_reason": "issue:45488 is the most concrete, actionable bug report in the set; it states a specific tokenizer regression with a clear causal change and downstream breakage, whereas most other issues are either orthogonal or much more generic.", + "summary": "Most candidates are clearly unrelated despite superficial lexical overlap. The only strong duplicate-style match is the Qwen video-input regression pair, where the specific qwen2.5-vl vision_position_ids bug is plausibly the same underlying v5.3.0 video regression affecting multiple Qwen VL variants.", + "confidence": 0.73, + "canonical_issue_reason": "issue:44479 is the broadest, most central report: a v5.3.0 video-input regression spanning multiple Qwen VL variants, and it cleanly subsumes the more specific qwen2.5-vl symptom in issue:45381.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is the strongest standalone issue here because it has a precise failure mode, a named affected family, and a clearly scoped fix target.", + "best_issue_reason": "issue:44479 is the best representative issue because it describes the shared concrete regression at the right abstraction level and is broad enough to anchor the only convincing duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Unrelated problems: one is TrainerState bookkeeping, the other is perplexity for gpt-oss-20b. No shared code path." + "reason": "Unrelated symptoms: TrainerState token counters vs GPT-OSS perplexity on WikiText." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different failure modes (special-token addition vs codec/warning behavior). Not the same bug." + "reason": "Both are tokenizer bugs, but they affect different models and different failure modes; no clear same code path." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "Both involve video paths, but one is an undefined helper/variable bug in multimodal token counting and the other is a broader video-input regression across several models. Too broad to merge as one fix." + "reason": "Both involve video processing, but 43329 is an undefined-name bug in multimodal token counting while 44479 is a broader Qwen video-input regression." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is unrelated to TrainerState num_input_tokens_seen updates." + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState's token accounting." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different failures: Qwen3Moe with FineGrainedFP8Config versus a key error loading MobileLLM-125M. No concrete shared defect." + "reason": "Different loading failures: Qwen3 MoE FP8 config vs MobileLLM key error." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "One is init_empty_weights/_is_hf_initialized handling, the other is PIL image processors incorrectly requiring torchvision. Different code paths." + "reason": "init_empty_weights/_is_hf_initialized is unrelated to PIL image processor torchvision requirements." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug is model-specific; the other is a generic report that tokenization changes across versions. Not enough evidence of the same underlying bug." + "reason": "Qwen3.5 hidden-state output bug is unrelated to a tokenizer version-change report." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Accelerate+DeepSpeed MOE weight conversion issue is unrelated to a Qwen3-VL weight-shape mismatch on load." + "reason": "Both are Qwen3 loading issues, but one is MoE weight conversion under accelerate+deepspeed and the other is a weight-shape mismatch for Qwen3-VL." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf_file loader breakage and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "GGUF/AutoTokenizer loading and Apple Silicon allocator warmup are unrelated." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "CUDA assert with auto device mapping and a video-input ValueError in Qwen3OmniMoe are different failures." + "reason": "Different Qwen3 failures: auto device mapping cuda assert vs video-input ValueError." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig save/load token IDs loss is unrelated to MXFP4 quantization falling back to bf16." + "reason": "DiaConfig token-id persistence is unrelated to MXFP4 quantization fallback." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Qwen3Moe FP8 loading failure is unrelated to AutoImageProcessor URL loading." + "reason": "Qwen3 MoE FP8 loading and AutoImageProcessor-from-URL are unrelated." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output drift and Qwen3.5 visual-encoder save_pretrained keys are different model-specific bugs." + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key regression are different problems." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer issues, but one is add_special_tokens/BOS-EOS behavior and the other is LlamaTokenizer overriding tokenizer.json pre-tokenizer settings. Different root causes." + "reason": "Both are tokenizer regressions, but they concern different tokenizer classes and different underlying behavior." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error for CLIP model is unrelated to TrainerState token counting." + "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token counts." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "FineGrainedFP8Config loading failure and CLIPTokenizer model_max_length constant are unrelated." + "reason": "Qwen3 MoE FP8 loading is unrelated to CLIPTokenizer max_length handling." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Tokenizer loading failure for a specific model and CLIPTokenizer max-length handling are different bugs." + "reason": "Model loading failure for physical-intelligence/fast is unrelated to CLIPTokenizer model_max_length." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both are video-related, but one is a processing ValueError in Qwen3OmniMoe and the other is a v5 video-input regression across several Qwen families. Not the same code-path bug." + "reason": "Both mention video, but one is Qwen3OmniMoe video-input processing and the other is a broader Qwen VL regression." }, { "left": "issue:44479", "right": "issue:45381", - "accept": false, - "reason": "Both are video regressions, but they cite different affected models and symptoms; no clear one-to-one duplicate evidence." - }, - { - "left": "issue:43540", - "right": "issue:44560", - "accept": false, - "reason": "Video-input ValueError in Qwen3OmniMoe and StopIteration in Qwen3-vl-embedding are distinct failures." - }, - { - "left": "issue:43317", - "right": "issue:45538", - "accept": false, - "reason": "Dequantized model offload failure is unrelated to CLIPTokenizer max-length behavior." + "accept": true, + "reason": "45381 looks like the same video-input regression as 44479, just reported as a specific qwen2.5-vl vision_position_ids symptom." }, { - "left": "issue:24643", - "right": "issue:29159", + "left": "issue:44568", + "right": "issue:45488", "accept": false, - "reason": "DeepSpeed training RuntimeError and tokenizer inconsistency are unrelated." + "reason": "Different tokenizer issues: add_special_tokens/BOS-EOS behavior vs LlamaTokenizer overriding ByteLevel pre-tokenization." }, { - "left": "issue:45137", - "right": "issue:45405", + "left": "issue:44737", + "right": "issue:45588", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and a minimum PEFT version bump are unrelated." + "reason": "XLNet CPU tensor creation bug is unrelated to flash_attention s_aux=None crash." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Loading a specific model under transformers>=5.1.0 and forwarding revision in list_repo_files are unrelated." + "reason": "Loading a model called physical-intelligence/fast is unrelated to list_repo_files not forwarding revision." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "MobileLLM-125M load key error and init_empty_weights unexpected argument are different loading bugs." + "reason": "Key error loading MobileLLM and init_empty_weights argument handling are different code paths." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "BLIP2 dtype persistence and SwitchTransformers/TimmWrapperModel bfloat16 mismatches are separate dtype issues in different model families." + "reason": "Both are dtype-related, but they affect different model families and likely different bugs." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken constructor conflict and tokenizer loading failure for est-roberta are not the same defect." + "reason": "AddedToken special-arg duplication is not the same as tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Unexpected zero probabilities in a SigLIP2 model and Perceiver image-classification failure are unrelated vision issues." + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to Perceiver interpolation failure." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU-side arange placement bug and flash_attention s_aux=None crash are unrelated." + "reason": "Relative positional encoding on CPU is unrelated to sink-less flash-attention handling." }, { "left": "issue:43540", "right": "issue:45823", "accept": false, - "reason": "Qwen3OmniMoe video-input failure and Gemma4 PLE device mismatch are unrelated model/runtime bugs." + "reason": "Qwen3OmniMoe video-input failure is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:43404", "right": "issue:45987", "accept": false, - "reason": "lm_head weight tying in Mistral3 is unrelated to StaticCache.get_seq_length returning the wrong type." + "reason": "lm_head weight tying in Mistral3 image-text generation is unrelated to StaticCache sequence-length return type." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MOE conversion with accelerate+deepspeed is unrelated to GPT-oss-20b CUDA OOM." + "reason": "MoE weight conversion/loading and CUDA OOM are not the same bug." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState num_input_tokens_seen not updating is unrelated to compute_loss receiving an unexpected keyword argument." + "reason": "TrainerState token tracking and custom compute_loss kwargs are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed weight-dimension error is unrelated to tokenizer slow/fast inconsistency." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (issue:44479 / issue:45381) is a conservative fit: both describe the same Qwen video-input regression at the same version line, with 44479 explicitly spanning qwen2_5_vl and related Qwen VL variants. The rejections also stay conservative and do not overclaim shared root causes from superficial lexical overlap." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5968,8 +6118,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -6052,203 +6202,212 @@ "issue:44514|issue:45381", "issue:24643|issue:29127" ], - "prepared_review_unit_hash": "a6acdbd2483cf168caf6730fb4e1b01ebc580a448341a7218b7476514939ec92", + "prepared_review_unit_hash": "cdd7115b88461beb19037a12c196bba7a32e91e2c953e8810ad0060f26b469b6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36375, - "estimated_input_tokens": 9094, - "estimated_eval_tokens": 18444 + "serialized_chars": 36371, + "estimated_input_tokens": 9093, + "estimated_eval_tokens": 18442 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it spans many unrelated Transformers issues across tokenization, model loading, config handling, vision, and platform-specific failures. I do not see a single duplicate family to collapse; all soft edges should be rejected.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue fits the cluster well because the items cover distinct bugs and regressions across unrelated code paths.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues spanning loading, config propagation, tokenization, multimodal models, and backend/package regressions. I only see one plausible duplicate pair (SigLIP2 discrepancy vs nonsensical outputs); the rest are too different to merge safely.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43493 is the clearest root-cause style report and best matches the only plausible duplicate pair in the set; most other items are unrelated bug reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the best overall anchor if one must pick one: it is a broad, current regression report with multiple models affected.", + "best_issue_reason": "43493 is the strongest representative issue for the cluster because it describes the underlying SigLIP2 behavior rather than a downstream symptom. It is still not a good umbrella for the rest of the items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "Different config bugs: one is rotary_pct not persisting on reload, the other is a TypeError in rope validation on JSON-loaded lists." + "reason": "Both concern config/rope handling, but one is a reload defaulting bug and the other is a JSON-list type error in rope validation; different code paths and failure modes." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Both tokenizer-related, but one is an AddedToken argument collision and the other is a Mistral regex patch crash from a missing backend_tokenizer attribute." + "reason": "Both involve tokenizers, but one is an AddedToken constructor clash and the other is a Mistral regex patch/backend_tokenizer crash; not the same bug." }, { "left": "issue:43906", "right": "issue:45381", "accept": false, - "reason": "The first is an isolated reproduction of an older issue; the second is a Qwen2.5-VL video position-id regression. Different underlying failures." + "reason": "The first is an isolated reproduction of an older issue; the second is a Qwen2.5-VL video position-id regression. No concrete overlap." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization in v5 is a different code path from FSDP CPU RAM efficient loading breaking." + "reason": "One is custom model initialization breaking in Transformers 5.0; the other is FSDP CPU RAM efficient loading breaking. Similar area, but not the same problem." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to Smollm3 RoPE layer dropping." + "reason": "Different models and different bugs: Qwen3 MOE accelerate/deepspeed weight conversion vs Smollm3 RoPE layer dropping." }, { "left": "issue:43493", "right": "issue:43994", - "accept": false, - "reason": "Both involve SigLIP2, but one is an HF-vs-JAX implementation discrepancy while the other is a model loading/pipeline output issue." + "accept": true, + "reason": "Both are about SigLIP2 behaving incorrectly: an HF-vs-JAX discrepancy and a symptom of wrong/nonsensical outputs through AutoModel/pipeline. Same underlying model implementation issue." }, { "left": "issue:36010", "right": "issue:42175", "accept": false, - "reason": "Importing GenerationMixin and missing TensorFlow backend in a torch install are separate dependency/import problems." + "reason": "An import error for GenerationMixin is unrelated to a packaging/backend omission of TensorFlow in the torch extra." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Unexpected _is_hf_initialized during init_empty_weights is unrelated to loading AutoImageProcessor from a URL." + "reason": "The first is an init_empty_weights argument mismatch; the second is an AutoImageProcessor URL-loading issue. Different subsystems and failures." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Different multimodal model failures: one is a failing janus test, the other is loading Phi-4 multimodal instruct." + "reason": "Both are multimodal/model-loading complaints, but they involve different models and different breakages." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP RAM-efficient loading and disabling mmap for Strix Halo OOM are different loading/runtime issues." + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are different memory/loading issues with no shared concrete bug." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Missing TensorFlow backend in a torch install is not the same as Qwen2_5_VLForConditionalGeneration being unavailable to import." + "reason": "One is about backend packaging for pip extras; the other is a missing Qwen2_5_VLForConditionalGeneration import. Unrelated." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "PIL image processor torchvision dependency regression is unrelated to the Mistral tokenizer backend_tokenizer crash." + "reason": "PIL backend torchvision dependency regression is unrelated to Mistral tokenizer regex patch crashes." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Environment-level PyTorch missing error is distinct from a specific Llama 3.2 90B Vision model loading regression." + "reason": "A missing PyTorch ImportError is not the same as a specific Llama 3.2 Vision model-loading regression." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "One is incorrect visual encoder key saving for Qwen3.5; the other is wrong token IDs from AutoTokenizer across several models." + "reason": "Qwen3.5 save_pretrained key corruption and AutoTokenizer token-id regression are different bugs affecting different APIs." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "A key error when loading MobileLLM-125M is not the same as missing Float8 storage object handling." + "reason": "MobileLLM key error and Float8 storage lookup failure are distinct loading/runtime issues." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file handling in from_pretrained and sleng-bert load failure are not the same underlying defect." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "An unreleased PEFT version pin and auto_mappings referencing removed configs are separate maintenance issues." + "reason": "PEFT version pinning and auto_mappings stale config references are separate CI/package maintenance problems." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels truncation and zero-loss single_label_classification are different model families and code paths." + "reason": "BeitImageProcessorFast label reduction and sequence-classification zero-loss behavior are unrelated label-handling bugs." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Generic missing fields with trust_remote_code and Qwen3_5MoeVisionConfig strict-dropping a field are related in theme but not the same bug." + "reason": "Generic trust_remote_code field loss and a specific Qwen3_5MoeVisionConfig strict-field drop are different config-loading issues." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "A missing PyTorch dependency error is unrelated to a poor error message when LayoutLMv3 boxes lack enough information." + "reason": "Missing PyTorch and a LayoutLMv3 box-content clarity issue are unrelated." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated failures." + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading have no shared code-path." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Smollm3 dropping RoPE layers and Qwen2.5-VL still-image position scaling are both rope-related but clearly different bugs." + "reason": "Both mention RoPE-related behavior, but one drops layers in Smollm3 and the other mis-scales Qwen2.5-VL temporal positions; different bugs." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Qwen3.5 num_labels propagation and strict config blocking granite_speech loading are different config-handling issues." + "reason": "num_labels propagation and strict config loading are separate config-handling issues." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Apple Silicon allocator warmup TypeError is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "AppleSilicon caching_allocator_warmup TypeError is unrelated to Qwen3.5 flash-attention generation issues." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight unpacking during accelerate loading is a different path from Perceiver image classification failing with interpolate_pos_encoding." + "reason": "BitNet packed-weight loading and Perceiver interpolation failures are unrelated model-specific regressions." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "RTDetrV2 output drift in v5 and Qwen3.5 save_pretrained checkpoint corruption are different regression types." + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained corruption are different regression types." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL-related, but one crashes in apply_chat_template batching and the other has wrong vision_position_ids for video input." + "reason": "Both are Qwen2.5-VL related, but one is batched chat-template padding behavior and the other is video position-id logic; not the same bug." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed training weight dimensionality error is unrelated to LayoutLMv3 box-content error messaging." + "reason": "A DeepSpeed weight-shape runtime error is unrelated to LayoutLMv3 box-information error messaging." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The single accepted pair (43493/43994) is a plausible duplicate on the face of the titles, and the rest of the soft-edge rejections are appropriately cautious rather than overstated." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6260,8 +6419,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -6326,8 +6485,8 @@ "issue:43295|issue:45356", "issue:44484|issue:44704", "issue:43653|issue:44361", - "issue:45542|issue:46178", "issue:44361|issue:44442", + "issue:45542|issue:46178", "issue:44514|issue:45290", "issue:38175|issue:43295", "issue:43653|issue:44568", @@ -6346,20 +6505,20 @@ "issue:43643|issue:44561", "issue:43577|issue:45588" ], - "prepared_review_unit_hash": "ad0585f0ae0ccdc6e031811f515f3d885d3123aabded81fc0d474b71d3607d91", + "prepared_review_unit_hash": "48341bb8404662b4cdb453bb2ae411495b50b2dcfb7b87cd52753c238fade7bd", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36586, + "serialized_chars": 36587, "estimated_input_tokens": 9147, "estimated_eval_tokens": 18550 }, @@ -6368,198 +6527,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions (tokenizers, processors, generation, loading, backend/device issues). None of the soft pairs look like the same underlying bug closely enough to merge as duplicates.", - "confidence": 0.92, - "canonical_issue_reason": "No single canonical issue fits the cluster well: the items span different subsystems and failure modes, so any one issue would be a poor representative.", + "summary": "These items are a heterogeneous set of unrelated Transformers bugs and regressions. The proposed soft pairs mostly share only a broad subsystem label (tokenizers, processors, generation, loading, or trust_remote_code) but not the same concrete failure mode, so none should be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No global best issue for dedupe purposes. If forced, 43452 is broadly about loading regressions, but it does not represent the cluster as a whole.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Different bugs: ZeRO-3 Gemma3 loading index error vs trainer.evaluate() failing after train()." + "reason": "Different failures: Gemma3 ZeRO-3 load-time IndexError vs trainer.evaluate() after train." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "Different components and symptoms: AutoTokenizer/AutoModel loading with gguf_file vs BLIP2 dtype not propagating." + "reason": "Unrelated bugs: gguf_file loading regression vs Blip2 dtype staying float32." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "One is a reproduction of an earlier issue; the other is a Gemma pipeline roundtrip/tokenization problem." + "reason": "No same code path; one is a reproduction of an old issue, the other is a Gemma pipeline roundtrip/tokenization problem." }, { "left": "issue:44466", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but the concrete bugs differ: device-dependent lm_head serialization vs translation-model weight tying behavior." + "reason": "Both mention tied weights, but one is serialization-by-device and the other is translation-model behavior in v5; not the same bug." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Loading gguf_file regression is unrelated to model_info() being called offline in _patch_mistral_regex." + "reason": "Both concern tokenizer loading, but one is gguf_file support and the other is offline model_info() usage in a regex patch." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Unrelated failures: mistral_common ReasoningEffort import vs Windows ROCm torch.distributed dependency issue." + "reason": "Completely different issues: Gemma 4 processor import regression vs Windows ROCm torch.distributed import/type failure." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "Different code paths: DAC.from_latents parity bug vs Whisper batch_decode ignoring skip_special_tokens." + "reason": "Different features and symptoms: DAC latent decoding mismatch vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "Trainer lifecycle bug vs Gemma4 device_map=\"auto\" device mismatch during forward." + "reason": "Different subsystems and failures: trainer lifecycle vs Gemma4 device-mapping mismatch." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Segmentation-map processing failure is unrelated to Qwen2-VL multi-GPU training." + "reason": "No overlap beyond being model-related; segmentation-map processing vs Qwen2-VL multi-GPU training." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are regressions, but one is processor.tokenizer/images access and the other is Kimi tokenizer codec/warning handling." + "reason": "Different regressions: processor.tokenizer access/images-to-tokenizer breakage vs Kimi-K2.5 codec/fix_mistral_regex behavior." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "Different API concerns: max_shard_size default vs AutoProcessor failing to pass kwargs to cached_file." + "reason": "Save-pretrained shard-size question is unrelated to AutoProcessor kwargs propagation to cached_file." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "Different tokenizer bugs: BigBird mask token registration vs MLukeTokenizer task AttributeError." + "reason": "Both tokenizer-related, but one is special-token registration/decode output and the other is an AttributeError in MLukeTokenizer tasks." }, { - "left": "issue:45542", - "right": "issue:46178", + "left": "issue:44361", + "right": "issue:44442", "accept": false, - "reason": "Different backend problems: undefined tf backend with tensorboard vs tensorflow_text not in BACKENDS_MAPPING." + "reason": "Different tokenizer loading failures with different models and causes." }, { - "left": "issue:44361", - "right": "issue:44442", + "left": "issue:45542", + "right": "issue:46178", "accept": false, - "reason": "MLukeTokenizer task error is unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + "reason": "Both backend-related, but one is missing TensorFlow backend state and the other is an unsupported tensorflow_text backend mapping." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Same area (chat templating) but different edge cases and failures, so not the same bug." + "reason": "Different apply_chat_template crashes: batched padding=False vs assistant tool-call messages without content." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero probabilities is unrelated to processor.tokenizer/images regression." + "reason": "Different model classes and failures: zero probabilities in SigLIP2 vs processor/tokenizer regression in v4.57.5." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Different tokenizer issues: BigBird special-token registration vs mdeberta add_special_tokens not adding BOS/EOS." + "reason": "Both tokenizer issues, but one is BigBird special-token registration and the other is mdeberta BOS/EOS addition in v5." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Different model behaviors: BLIP2 dtype stays float32 vs RTDetrV2 output changes in v5." + "reason": "Different bugs: BLIP2 dtype casting vs RTDetrV2 output changes in Transformers v5." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Both involve generation/cache, but one is compiled-forward chunked generation inconsistency and the other is Qwen3 recurrent-state reset." + "reason": "Both generation/cache related, but one is chunked generation with compiled forward and the other is Qwen3 modular recurrent-state reset." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Unexpected key with position id is unrelated to sinq quantized save/load breakage." + "reason": "Different serialization/loading problems: unexpected position-id keys vs sinq quantization roundtrip breakage." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Offline model_info() call and a tokenizer-loading failure for EMBEDDIA/est-roberta have different root causes." + "reason": "Both tokenizer-loading regressions, but one is offline model_info() access and the other is a specific EMBEDDIA/est-roberta load failure." }, { "left": "issue:29159", "right": "issue:33357", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to a MacOS bus error loading a CLIP community model." + "reason": "Tokenizer inconsistency vs MacOS bus error on a CLIP model are unrelated." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Sliding_window Qwen3-MoE bug is unrelated to MobileLLM-125M key error during loading." + "reason": "Different model-specific issues: sliding_window in Qwen3-MoE vs loading MobileLLM-125M key error." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBird mask-token registration and DiaConfig custom token-ID persistence are separate tokenizer/config bugs." + "reason": "Different failures: BigBird special-token decode bug vs DiaConfig losing custom token IDs after save/load." }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and the reproduction issue 38071 are not the same failure." + "reason": "Different problems and code paths: auto device mapping cuda assert vs a reproduction of another issue." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Different generation regressions: compiled forward chunking vs Qwen3.5 flash-attention behavior." + "reason": "Both affect generation, but compiled-forward chunking and Qwen3.5 flash-attention failures are not the same concrete bug." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Torch.compile/SDPA issue on Bamba is unrelated to DiaConfig save/load token ID loss." + "reason": "Unrelated: torch.compile/SDPA failure in Bamba vs DiaConfig save/load token-id loss." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Different subsystems: compiled SDPA on Bamba vs tokenizer AttributeError on MLukeTokenizer." + "reason": "Completely different: Bamba compile/SDPA issue vs tokenizer task AttributeError." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Flash-attention generation issue and FP8 inference failure on B200 are distinct hardware/code-path problems." + "reason": "Different runtime failures: Qwen3.5 flash-attention generation vs FP8 inference on B200." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both involve trust_remote_code, but one is missing fields from AutoConfig and the other is a v5 removal breaking remote-code models." + "reason": "Both involve trust_remote_code, but one is missing fields from AutoConfig and the other is removal of is_torch_fx_available breaking models." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "BLIP2 dtype propagation and flash_attention.py crashing on s_aux=None are unrelated." + "reason": "Different code paths: BLIP2 dtype loading vs flash_attention.py crash on s_aux=None." } ] }, @@ -6578,8 +6737,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -6641,8 +6800,8 @@ "issue:43504|issue:44610", "issue:45491|issue:46002", "issue:32090|issue:33666", - "issue:44466|issue:46050", "issue:33666|issue:36010", + "issue:44466|issue:46050", "issue:43742|issue:44821", "issue:43688|issue:45440", "issue:45823|issue:46050", @@ -6665,220 +6824,220 @@ "issue:39692|issue:43493", "issue:43577|issue:44554" ], - "prepared_review_unit_hash": "0e447508d8da8d4eac61d1ff95baeff9a98c61eeffa93da3156cd539cd0d5b4b", + "prepared_review_unit_hash": "637eb13b9fb0799d4536528772077f5920a3029c66ca0f970685ee96f40c3c3d", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37128, - "estimated_input_tokens": 9282, - "estimated_eval_tokens": 18820 + "serialized_chars": 37132, + "estimated_input_tokens": 9283, + "estimated_eval_tokens": 18822 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs are false positives: they share at most a broad subsystem or symptom, but not the same concrete bug/change. No PRs are present, and there is no single canonical issue that subsumes the set.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "The soft pairs are mostly superficial similarities across unrelated bugs; none look like the same underlying issue or a mergeable duplicate PR. This set is a grab-bag of model-loading, trainer, config, and inference regressions rather than one coherent defect cluster.", + "confidence": 0.27, + "canonical_issue_reason": "Issue 43504 is a clean, self-contained load-time bug report with a concrete failure mode, making it the best anchor among these mostly unrelated reports.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 43504 is the strongest representative issue here because it is specific, reproducible, and centered on a concrete code path rather than a vague symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "Different failures: GPU broadcast rejects NoneType in Trainer, while the other is a TrainerState token-count update bug." + "reason": "Trainer GPU broadcast NoneType error and token-count tracking bug are different training subsystems and failure modes." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "Unrelated issues: missing PyTorch ImportError versus token-embedding resize reinitializing lm head on post_init." + "reason": "Missing PyTorch import and output embedding reinitialization in post_init are unrelated bugs." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Legacy field loading bug for Beit semantic segmentation is unrelated to Bamba torch.compile/SDPA failure." + "reason": "Legacy-field model loading failure vs torch.compile/SDPA failure in Bamba are different code paths and models." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Both involve loading, but one is a Beit config legacy-field preset issue and the other is BitNet packed-weight unpacking during accelerate load." + "reason": "BEiT preset loading bug and BitNet packed-weight accelerate loading bug are distinct model-loading issues." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Different code paths: Qwen3 auto device mapping causes cudaErrorAssert, while chunked generation diverges under compiled forward." + "reason": "Auto device mapping cuda assert and chunked-generation compiled-forward inconsistency are separate runtime bugs." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Beit pretrained preset loading and OmDet-Turbo processor/model size mismatch are separate bugs." + "reason": "Pretrained preset load failure is unrelated to OmDet processor/model input-size mismatch." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on mixed-length batches and the other is classification-head NaNs after checkout/init." + "reason": "Both involve NaNs, but one is a Gemma3 sliding-window GPU issue and the other is a classification-head init/checkout issue." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error is unrelated to Qwen2-VL multi-GPU training support." + "reason": "Custom Trainer broadcast failure and Qwen2-VL multi-GPU training are not the same bug." }, { - "left": "issue:44466", - "right": "issue:46050", + "left": "issue:33666", + "right": "issue:36010", "accept": false, - "reason": "One is v5 tied-weight serialization inconsistency for lm_head.weight; the other is sinq quantized save/load breakage." + "reason": "Multi-GPU training issue and GenerationMixin import error are unrelated." }, { - "left": "issue:33666", - "right": "issue:36010", + "left": "issue:44466", + "right": "issue:46050", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL is unrelated to the GenerationMixin import regression." + "reason": "lm_head serialization inconsistency and sinq quantization save/load breakage are different persistence bugs." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "MobileLLM loading key error and AutoImageProcessor URL loading are different loader failures." + "reason": "MobileLLM checkpoint load failure and AutoImageProcessor URL loading are different object-loading paths." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Aux-loss normalization in OLMoE/GPT Oss is not the same bug as DeepseekV3MoE diverging from remote implementation." + "reason": "Aux-loss normalization bug and DeepSeekV3 implementation divergence are different correctness issues." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device mismatch under device_map=auto is unrelated to sinq quantized save/load corruption." + "reason": "Gemma4 auto device-map mismatch and quantized model save/load breakage are unrelated." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "BitNet packed-weight unpacking on accelerate load and DiaConfig custom token-ID persistence are unrelated failures." + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are different models and failure modes." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and processor.tokenizer/image-passing regression are separate processor bugs." + "reason": "Audio chat-template sampling-rate default and processor.tokenizer regression are distinct processor bugs." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU-offload device mismatch in bark-small is unrelated to DiaConfig losing custom token IDs on save/load." + "reason": "CPU offload device mismatch and config token-ID loss on save/load are unrelated." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Missing PyTorch ImportError and Trainer GPU broadcast NoneType error are unrelated." + "reason": "Missing torch import and Trainer broadcast TypeError are completely different errors." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file breaking AutoTokenizer/AutoModel loading is a different failure from ScandiBERT load errors." + "reason": "gguf_file loading breakage and inability to load a specific model are different load failures." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug is unrelated to FP8 inference failure on B200/SM100 GPUs." + "reason": "Qwen3.5 hidden-state bug and FP8 inference failure on B200 are unrelated." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "PyTorch-not-found ImportError and MacOS bus error on pretrained CLIP are different issues." + "reason": "PyTorch import error and MacOS bus error on CLIP are not the same underlying bug." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 incorrect outputs and return_tensors being ignored in text_kwargs are unrelated behaviors." + "reason": "SigLIP2 wrong outputs and return_tensors being ignored are separate API regressions." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are missing pad-token-related config fields, but they affect different models and different attribute names, so not the same bug." + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are similar only superficially; they affect different configs and fields." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Qwen2-VL multi-GPU training is unrelated to CustomTrainer.compute_loss receiving num_items_in_batch." + "reason": "Multi-GPU training issue and custom compute_loss signature mismatch are unrelated trainer problems." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are different bugs." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Smollm3 RoPE-layer count mismatch and GPTNeoX rotary_pct reload regression are different config bugs." + "reason": "Smollm3 RoPE-layer-count bug and GPTNeoXConfig rotary_pct reload bug are different configuration issues." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 HF-vs-JAX implementation discrepancy is not the same as pipeline() advertising unsupported translation tasks." + "reason": "SigLIP2 implementation discrepancy and pipeline translation-task messaging are unrelated." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Qwen3.5 hidden-states bug and incorrect visual-encoder keys in save_pretrained are separate regressions." + "reason": "Qwen3.5 hidden-states bug and save_pretrained visual-encoder key regression are different failure modes." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression in Qwen VL variants and num_labels propagation in Qwen3.5 are different problems." + "reason": "Video-input regression and num_labels propagation bug are separate Qwen3.5 issues." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Loading Phi-4 multimodal instruct and bumping MIN_PEFT_VERSION to an unreleased version are unrelated." + "reason": "Model load failure and PEFT version pinning are unrelated." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both are about SigLIP2, but one is a docs/example mismatch plus quantization failure and the other is an implementation parity issue with JAX." + "reason": "SigLIP2 doc-example failures and SigLIP2 implementation divergence are not the same concrete bug." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype staying float32 and an MPS attention correctness issue are unrelated." + "reason": "Blip2 dtype-loading bug and MPS attention correctness issue are unrelated." } ] }, @@ -6897,8 +7056,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -6981,208 +7140,208 @@ "issue:44423|issue:44734", "issue:44464|issue:45584" ], - "prepared_review_unit_hash": "8fce7c3338a5ae0a31871d2c5194c9442b23d22d0c29d3a80f9437e93b4ac045", + "prepared_review_unit_hash": "ed7803c2025dd1bca8dcb49e579a4537139bdeadc30724fa9ef6778271285ba3", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35982, - "estimated_input_tokens": 8996, - "estimated_eval_tokens": 18248 + "serialized_chars": 35988, + "estimated_input_tokens": 8997, + "estimated_eval_tokens": 18250 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose grab bag of Transformers regressions across tokenization, model loading, generation, training, and device/memory behavior. The soft pairs share broad themes but not the same concrete bug, so they should all be kept separate.", - "confidence": 0.57, + "summary": "The items are mostly unrelated standalone issues; the soft edges are largely false positives based on broad surface similarity (tokenizers, Qwen, Gemma, loading regressions, CI/device issues) rather than the same underlying bug.", + "confidence": 0.95, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 45701 is the broadest representative here because it captures a version-wide tokenization regression theme, but it is still not a true umbrella for the many unrelated loading/training/device bugs in the cluster.", + "best_issue_reason": "issue:45812 is the strongest standalone report here: it is concrete, well-scoped, has a clear regression boundary (v4\u2192v5), and describes a high-impact tokenizer bug with a specific affected family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Aux-loss normalization in MoE models and an FSDP CPU RAM-efficient loading break are different training-path bugs." + "reason": "Different problems: auxiliary-loss normalization vs FSDP RAM-efficient loading failure." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "One is special-token insertion for mdeberta-v3; the other is wrong token IDs across Granite models. Related area, different bug." + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenization behavior." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Video-input processing failure in Qwen3OmniMoe is unrelated to Qwen3.5 flash-attention generation behavior." + "reason": "Different code paths: Qwen3OmniMoe video preprocessing vs Qwen3.5 flash-attention generation." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Defaulting mm_token_type_ids and multi-GPU CUDA OOM are distinct problems with different causes and fixes." + "reason": "One is a missing default for mm_token_type_ids; the other is a multi-GPU inference OOM." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Loading a specific model preset and an unconditional hub model_info() call in offline mode are different failure modes." + "reason": "Both are model-loading failures, but one is a specific model compatibility issue and the other is an offline-mode hub lookup regression." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both mention missing pad-token config, but they affect different models and config fields, so they are not the same bug." + "reason": "Different model/config attributes and different failure modes; only superficial similarity around missing token-id fields." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "A SigLIP2 documentation example error is not the same as a regression in processor.tokenizer and tokenizer image handling." + "reason": "Docs-example errors vs a runtime regression in custom processor/tokenizer code; not the same bug." }, { "left": "issue:43756", "right": "issue:45083", "accept": false, - "reason": "RoPE-layer dropping and a feature-extraction length helper behaving unexpectedly are unrelated model internals." + "reason": "Unrelated model architecture/blog discrepancy vs a helper-function bug in qwen3_omni_moe." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL and not unwrapping text_config in AutoModelForCausalLM.from_config are different loading bugs." + "reason": "Different Qwen multimodal loading issues with different missing/unwrap logic." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "MoE auxiliary-loss scaling and a T5 apex RMSNorm memory leak are unrelated code paths." + "reason": "Loss normalization bug is unrelated to the T5 RMSNorm memory leak." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and tied-weight serialization differing by device are separate v5 regressions." + "reason": "Both are v5 regressions, but they hit different code paths: init behavior vs weight serialization." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 implementation drift and a gradient blow-up under HfDeepSpeedConfig are not the same underlying issue." + "reason": "Different subsystems: SigLIP2 implementation mismatch vs DeepSpeed/ZeRO gradient explosion." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "An import error for Qwen2_5_VLForConditionalGeneration and a Float8 storage lookup failure are unrelated compatibility issues." + "reason": "Importing a Qwen model class is unrelated to the Float8 storage lookup failure." }, { "left": "issue:45229", "right": "issue:45823", "accept": false, - "reason": "Multi-GPU inference OOM and a PLE device mismatch with device_map=\"auto\" are different runtime failures." + "reason": "Both involve Gemma4, but one is OOM and the other is a device-map mismatch during forward." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "A MacOS CLIP bus error and Qwen2-VL multi-GPU training are clearly unrelated." + "reason": "Completely different models and failure modes: MacOS bus error vs multi-GPU training question." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Token ID loss on save/load and an AttributeError in MLukeTokenizer are different tokenizer bugs." + "reason": "Different tokenization/model API bugs with no shared underlying code path." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field model loading failure and CPU-offload device mismatch are different bugs in different execution paths." + "reason": "Legacy-field model loading bug vs CPU-offload device mismatch; not the same issue." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage resolution and flash_attn import handling on Python 3.13 are unrelated optional-dependency issues." + "reason": "Different failures: missing Float8 storage type vs missing flash_attn key on Python 3.13." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and an unexpected _is_hf_initialized argument in init_empty_weights are separate config/init regressions." + "reason": "A missing config field after 5.0 update is not the same as an init_empty_weights argument mismatch." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift between train/eval and a ZeRO-3 gradient anomaly are not the same problem." + "reason": "Training/eval dtype drift is unrelated to the ZeRO-3 gradient anomaly." }, { "left": "issue:44625", "right": "issue:45759", "accept": false, - "reason": "num_labels propagation and text_config unwrapping are both Qwen3.5 config issues, but they affect different fields and behaviors." + "reason": "Related model family, but different propagation bugs and different affected APIs." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "MOE weight conversion with accelerate+deepspeed and Qwen3VL MoE loading breakage are related family-wise but not the same concrete bug." + "reason": "Both mention Qwen3 MoE, but one is accelerate/deepspeed weight conversion and the other is dev loading breakage; not the same concrete bug." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "BEiT legacy-field loading and BigBirdTokenizer special-token registration are unrelated." + "reason": "Different models and failures: Beit legacy field loading vs BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "A real CPU-offload device mismatch is not the same as a stale CI test failure." + "reason": "A real CI test failure on GraniteSpeech is not the same as a CPU-offload device mismatch in bark-small." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "A TP OOM on Qwen2-57B and an unreleased MIN_PEFT_VERSION bump are unrelated." + "reason": "OOM during tensor-parallel loading is unrelated to a dependency version bump." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "MoE auxiliary-loss normalization and a tokenization regression are different subsystems." + "reason": "The titles only share a broad theme of model behavior changes; the actual bugs are different." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A Janus model test failure and a dependency-version bump are unrelated." + "reason": "A Janus test failure is unrelated to the PEFT version pin issue." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Continuous-batching multimodal crash and KV-cache continuation indexing are different serve bugs." + "reason": "Both are server/runtime issues, but one is a multimodal continuous-batching crash and the other is KV-cache continuation indexing." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Compiled-forward chunked generation mismatch and Whisper empty-transcription handling are unrelated generation issues." + "reason": "Chunked generation compiled-forward inconsistency is unrelated to Whisper empty-transcription handling." } ] }, @@ -7201,8 +7360,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -7288,20 +7447,20 @@ "issue:45200|issue:45325", "issue:43295|issue:43873" ], - "prepared_review_unit_hash": "b14a563ac4f0f03abec29272518c61d820825a786226e6f40e3d3ec342890c97", + "prepared_review_unit_hash": "f4e56951dfa82192a6270529af8b6b4ab9acc9a3961aebf6856a8b3164251451", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37032, + "serialized_chars": 37030, "estimated_input_tokens": 9258, "estimated_eval_tokens": 18772 }, @@ -7310,198 +7469,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft pairs share broad subsystem vocabulary but not the same underlying bug or fix. No soft edge should be accepted. The strongest standalone issue is the Gemma4 processor/tokenizer regression because it is concrete, user-facing, and broader than the other highly model-specific reports.", - "confidence": 0.89, - "canonical_issue_reason": "No clear canonical issue for deduping: these items span unrelated failures across tokenizer, generation, quantization, loading, and model-specific regressions. If one must be highlighted, issue 43295 is the broadest actionable regression, but it is not a duplicate of the others.", - "canonical_pr_reason": "No pull requests are present in this cluster.", - "best_issue_reason": "issue:43295 is the best standalone representative because it reports a concrete regression with clear impact and a reproducible processor/tokenizer breakage, making it more actionable than the other narrowly scoped or orthogonal issues.", + "summary": "These items are a heterogeneous set of distinct bugs and regressions across tokenization, generation, multimodal processing, loading, and training. The soft-similarity pairs are mostly broad-subs\u200bystem overlaps rather than the same concrete underlying issue, so none should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Different failures: Flash Attention 2 CI vs MLflow reporting on job failure; no shared code path." + "reason": "Unrelated failures: Flash Attention test regressions vs MLflow reporting after a failed training job." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Both mention model behavior, but one is SigLIP2 output quality and the other is a position-id key warning; not the same bug." + "reason": "Both mention model correctness, but one is a SigLIP2 inference-quality bug and the other is a generic position-id key warning across many models." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "Flash-attention integration AttributeError vs quantized save/load breakage; unrelated paths." + "reason": "Different problems: flash_attention sink-less crash on s_aux=None vs quantized model save/load corruption." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient growth and multimodal model loading are distinct issues." + "reason": "DeepSpeed ZeRO-3 gradient growth and Phi-4 multimodal loading regression are separate issues." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing bug vs TrainerState accounting bug; different subsystems and symptoms." + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState's num_input_tokens_seen tracking." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is unrelated to a trainer GPU broadcast NoneType error." + "reason": "Tokenizer slow/fast inconsistency does not match the Trainer NoneType broadcast crash." }, { "left": "issue:43688", "right": "issue:43873", "accept": false, - "reason": "Aux loss normalization and quantization offloading are different features with different failure modes." + "reason": "Aux-loss normalization and quantization/offloading behavior are different code paths and symptoms." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both involve chat templates, but one is assistant mask generation and the other is tool-call tokenization crash; not the same underlying defect." + "reason": "Both touch apply_chat_template, but one is all-zero assistant masks for multimodal inputs and the other is a crash on tool-call messages with no content." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness issue vs flex_attention version-check/import bug; unrelated code paths." + "reason": "Attention correctness on MPS is unrelated to a PyTorch version check in flex_attention import logic." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video processing error in Qwen3OmniMoe vs Gemma4 mm_token_type_ids defaulting; different model families and bugs." + "reason": "Video input validation in Qwen3OmniMoe is not the same as Gemma4 mm_token_type_ids defaulting." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Template-node compilation TypeError and flash_attn import KeyError are unrelated." + "reason": "Compilation of non-template nodes and flash_attn import KeyError are distinct failures." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript generator-expression export failure vs SigLIP2 zero probabilities; no overlap in bug mechanism." + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability inference." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is missing token-type defaults and the other is a device_map forward mismatch." + "reason": "Gemma4 token-type-id defaulting and device-map forward placement issues are different bugs." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Both touch quantization, but offloading behavior and sinq save/load corruption are distinct problems." + "reason": "Quantization offloading issues and sinq save/load breakage do not describe the same defect." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Bfloat16 dtype mismatches in inference vs Wav2Vec2 save_pretrained/tokenization failure; unrelated." + "reason": "bfloat16 dtype mismatches in inference are unrelated to Wav2Vec2 save_pretrained/tokenization failures." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Embedding dtype regression and resume_from_checkpoint key mismatch are separate training behaviors." + "reason": "BF16-to-FP32 dtype drift in eval and resume_from_checkpoint key mismatch are separate training-state problems." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation kwargs after sync_gpus vs hidden-state output bug in Qwen3.5; different generation issues." + "reason": "Generation kwargs sync bug and Qwen3.5 output_hidden_states bug affect different paths." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "Tokenizer crash on NER padding/truncation vs return_tensors being ignored when text_kwargs is set; not the same bug." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are not the same as return_tensors being ignored with text_kwargs." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and rope-parameter JSON list validation are unrelated." + "reason": "Gemma4Processor _tokenizer attribute regression is unrelated to rope-param list type handling." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Gemma4 device mismatch during forward vs tokenizer argument precedence bug; different failure modes." + "reason": "Gemma4 device placement and return_tensors precedence are different behaviors." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy and checkpoint resume key mismatch do not share a code path." + "reason": "SigLIP2 implementation mismatch and checkpoint resume key mismatch are unrelated." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer added-space bug vs MacOS bus error on CLIP model; no common underlying issue." + "reason": "Added-token spacing and CLIP MacOS bus error are unrelated tokenizer vs runtime crashes." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer regression and batched video-shape bug are both multimodal, but they affect different APIs and code paths." + "reason": "Processor.tokenizer regression and batched video shape bug are different multimodal issues." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Processor/tokenizer access regression is not the same as an image-size mismatch between processor and model." + "reason": "Both are processor/model input regressions, but one is an attribute/API break and the other is a resolution mismatch." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and loading physical-intelligence/fast on 5.1.0 are different version-specific issues." + "reason": "Missing pad_token_idx in StableLmConfig is not the same as a specific model-load failure on transformers>=5.1.0." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Qwen2.5-VL video position-id bug and FP8 inference failure on B200 are unrelated." + "reason": "Qwen2.5-VL vision_position_ids bug and FP8 inference failure on B200 are unrelated." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Both are loading/version regressions, but one is a missing config field and the other is meta-device loading breakage." + "reason": "Config missing pad_token_idx and meta-device loading regressions are different loading problems." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "Attention correctness on MPS is unrelated to tokenizer codec/regression handling in Kimi-K2.5." + "reason": "MPS attention correctness issue is unrelated to Kimi-K2.5 tokenizer codec/regression handling." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "Broken docs example and return_tensors precedence bug are unrelated." + "reason": "SigLIP2 documentation/example errors are not the same as return_tensors being ignored." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma4 token-type defaults and Qwen2.5-VL temporal position scaling are different multimodal bugs." + "reason": "Gemma4 token-type-id defaults and Qwen2.5-VL temporal position-id scaling are distinct multimodal bugs." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to quantization offloading behavior." + "reason": "Processor API regression and quantization offloading failure are not the same underlying bug." } ] }, @@ -7607,20 +7766,20 @@ "issue:43335|issue:43550", "issue:37428|issue:43425" ], - "prepared_review_unit_hash": "07bc3d2663c881bab35eb3cdbebf32ce2eb1734eabc589229164a7207eba713a", + "prepared_review_unit_hash": "aa067815102bf5afc9630624fa74b6d56f6b27f7fd14f25c43ad07fff86b5532", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37168, + "serialized_chars": 37167, "estimated_input_tokens": 9292, "estimated_eval_tokens": 18840 }, @@ -7629,198 +7788,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most items are unrelated model-specific bugs, import errors, docs regressions, or performance issues. No strong duplicate cluster emerges; the only near-match is within multimodal chat-template handling, but even that is not concrete enough to merge confidently.", - "confidence": 0.34, - "canonical_issue_reason": "Issue 44521 is the broadest and most central API-level bug in the set: it targets `apply_chat_template` on multimodal inputs, has active references, and is the best representative if one issue must stand in for the cluster.", + "summary": "This cluster is highly heterogeneous: the soft-similarity pairs span unrelated import errors, model-specific regressions, processor bugs, docs issues, and performance/OOM reports. None look like the same concrete underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No clear duplicate cluster exists; if a representative issue must be chosen, issue 44521 is the most active open item with the broadest follow-up potential around `apply_chat_template` multimodal behavior.", "canonical_pr_reason": null, - "best_issue_reason": "44521 is the most representative issue here because it touches a core shared user-facing path (`apply_chat_template`) and is phrased as a general multimodal bug rather than a narrow model-specific symptom.", + "best_issue_reason": "Issue 44521 is the best representative because it is open, has the most discussion/inbound interest, and concerns a user-facing multimodal templating path likely to draw continued attention.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Different subsystems and symptoms: Perceiver resizing/interpolation vs RT-DETR memory release." + "reason": "Different bugs: Perceiver non-default resolution failure vs RT-DETR memory not being released." }, { "left": "issue:43550", "right": "issue:43844", "accept": false, - "reason": "Different code paths: torch.compile with SDPA vs DeepSpeed ZeRO-3 gradient growth." + "reason": "Different failure modes: torch.compile+SDPA crash vs gradient growth under ZeRO-3." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is serialization/device-dependent and the other is LoRA merge collapse; not the same bug." + "reason": "Different problems: lm_head serialization/tied weights vs LoRA merge collapse with extended vocab." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Same symptom class (`pad_token_id` missing) but on different configs/models; not enough to treat as one underlying defect." + "reason": "Same symptom class (`pad_token_id` missing) but different model families and config paths; not the same bug." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Both affect loading/import, but one is a missing flash-attn helper import and the other is meta-device loading regressions." + "reason": "Import helper removal/import error vs meta-device loading regression; different code paths." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Unrelated issues: batched video processor shape vs tensor-parallel OOM on Qwen2." + "reason": "Unrelated issues: batched video processor shape vs tensor-parallel OOM." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Different backend/import failures (`flash_attn` vs `tensorflow_text`) with no shared concrete code path." + "reason": "Both are import/platform issues, but they involve different backends and different root causes." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Import error for a model class vs a detokenization regression; unrelated failures." + "reason": "Import error for a VL model class vs incorrect detokenization in DeepSeek coder; unrelated." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Local-path import warning and a GPT-oss CUDA OOM are unrelated." + "reason": "Local-path trailing-slash warning vs GPT-oss-20b CUDA OOM; unrelated." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Different problems: dtype mismatch in eval vs MoE training memory usage." + "reason": "Embedding dtype regression in eval vs Qwen3 MoE memory inefficiency; different bugs." }, { "left": "issue:43526", "right": "issue:46192", "accept": false, - "reason": "Different processor bugs: label reduction in Beit fast processor vs `return_tensors` handling." + "reason": "BeitImageProcessorFast label reduction bug vs `return_tensors` being ignored with `text_kwargs`; different processors and behaviors." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "FineGrainedFP8Config failure and a Qwen2.5-VL import error are unrelated." + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to a Qwen2.5-VL import error." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token registration vs a stale CI device override test." + "reason": "Tokenizer special-token registration bug vs stale GraniteSpeech CI test; unrelated." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Both are SigLIP2-related, but one is zero probabilities and the other is a broader HF-vs-JAX discrepancy; not clearly the same concrete bug." + "reason": "Both are SigLIP2-related, but one is zero probabilities and the other is HF/JAX parity mismatch; not the same concrete bug." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Different tokenizer/model families and different failure modes." + "reason": "Different ASR/tokenizer issues: FastSpeech2ConformerTokenizer load failure vs Wav2Vec2 save/tokenization failure." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Detokenization regression vs Python 3.13 flash-attn import/backend issue." + "reason": "DeepSeek detokenization bug vs import_utils flash_attn KeyError on Python 3.13; unrelated." }, { "left": "issue:43295", "right": "issue:43493", "accept": false, - "reason": "Processor/tokenizer regression and SigLIP2 implementation discrepancy are unrelated." + "reason": "Processor.tokenizer / image-passing regression is not the same as SigLIP2 implementation discrepancy." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both are missing import-utils helpers, but they gate different features and different remote-code failures." + "reason": "Different removed/missing import_utils symbols (`is_torch_fx_available` vs `is_falcon_mamba_ssm_available`) and different breakages." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Different missing symbols/classes in different import paths." + "reason": "Different import errors for different missing symbols/models; not one underlying bug." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 probability bug and RT-DETR memory leak are unrelated." + "reason": "SigLIP2 output discrepancy and RT-DETR memory retention are unrelated." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both involve `apply_chat_template`, but one is a batched-padding crash and the other is incorrect assistant masks; not clearly one fix." + "reason": "Both hit `apply_chat_template`, but one crashes on batched padding=False and the other returns zero assistant masks; different bugs." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate default and batched video shape are different processor bugs." + "reason": "Audio sampling-rate default bug vs batched video input shape bug; unrelated." }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Different hardware/model/runtime failures: CUDA assert on A800 vs FP8 inference on B200." + "reason": "Qwen3 auto device-map CUDA assert vs FP8 inference failure on B200; different hardware/runtime issues." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Different platform/dependency import failures with no shared code path." + "reason": "Both are import-time/platform problems, but the affected dependencies and failure points differ." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC forward/from_latents mismatch and a GraniteSpeech CI device override failure are unrelated." + "reason": "DAC quantization/STE mismatch vs GraniteSpeech CI device override test; unrelated." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor output-size bug vs AMP/CUDA loss crash are distinct issues." + "reason": "Processor output-size bug vs AMP float16 loss crash; different subsystems and causes." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs regressions, but they concern different APIs and different removals." + "reason": "Both are docs issues, but they concern different pipelines and different deprecated parameters." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "Different model/component logic: DAC latent mismatch vs qwen3_omni_moe helper behavior." + "reason": "DAC latent conversion mismatch vs qwen3_omni_moe feature-length helper behavior; unrelated." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Feature-extractor crash on unsupported arg vs Whisper empty-transcription generation bug." + "reason": "Audio feature-extractor regression vs Whisper empty-transcription generation bug; different code paths." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformers config bug and Bamba torch.compile/SDPA failure are unrelated." + "reason": "SwitchTransformers sparse-layer configuration bug vs Bamba torch.compile+SDPA failure; unrelated." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Missing flash-attn symbol import and Torch 2.10 incompatibility are not the same defect." + "reason": "ImportError for a missing flash-attn helper vs generic Torch 2.10 incompatibility; not the same issue." } ] }, @@ -7925,226 +8084,226 @@ "issue:42490|issue:43854", "issue:43606|issue:44857" ], - "prepared_review_unit_hash": "94df5672862af146a8c3b3b045bfa11f8b3361f6e05c324904bc04ab4f70551a", + "prepared_review_unit_hash": "373ada10a182d310e21ab9cfdb9b2bf2ad9c89d60e1be3867e7f6829bef5543b", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36704, - "estimated_input_tokens": 9176, - "estimated_eval_tokens": 18608 + "serialized_chars": 36707, + "estimated_input_tokens": 9177, + "estimated_eval_tokens": 18610 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "One clear duplicate pair stands out: issue 45310 and 45478 describe the same Qwen3.5 MoE from_pretrained regression on transformers>=5.4.0. The rest of the soft pairs are only superficially similar and appear to concern different code paths or symptoms, so they should stay separate.", + "summary": "Most items in this cluster are unrelated bug reports that only share broad Transformers subsystem vocabulary. The only clear duplicate pair is the Qwen3.5 MoE from_pretrained regression: issues 45310 and 45478. No pull requests are present.", "confidence": 0.93, - "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is earlier than 45478, has the same title/problem statement, and is the cleaner anchor for the duplicate pair.", + "canonical_issue_reason": "Issue 45310 is the clearest canonical issue because it is the earlier report for the same Qwen3.5 Moe from_pretrained failure on transformers>=5.4.0, and 45478 is a near-verbatim follow-up of the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "45310 is the strongest representative issue in the set because it is a direct duplicate target with the clearest, most specific failure description.", + "best_issue_reason": "45310 is the best issue to keep as the representative duplicate target: it matches 45478 exactly on the failing model/version and appears slightly earlier and more canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Different bugs: AMP/CUDA crash in LwDetrImageLoss vs missing auxiliary losses in D-FINE." + "reason": "Different detection bugs: AMP/CUDA loss crash vs auxiliary-loss omission when denoising is off." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "Both mention model/runtime behavior, but one is a CPU offload device mismatch and the other is a feature-length helper bug in qwen3_omni_moe." + "reason": "Unrelated code paths: CPU offload/device mismatch in Bark vs a qwen3_omni_moe feature-length helper." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Different symptoms and code paths: Janus image generation test failure vs stale generate warning for num_return_sequences." + "reason": "One is a Janus image generation test failure; the other is a paged generate warning for num_return_sequences." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Different problems: Pegasus v4/v5 output drift vs Qwen3 modular recurrent state reset with cache." + "reason": "Different models and failures: Pegasus v4/v5 output drift vs Qwen3 recurrent-state reset with cache." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge/tied-embeddings collapse is unrelated to the category cardinality runtime error." + "reason": "LoRA/extended-vocab collapse is unrelated to the 2^24 categories runtime limit." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "apply_chat_template tool-call crash and qwen2.5-vl video position_ids issue affect different APIs and failure modes." + "reason": "Chat-template tool-call crash and Qwen2.5-VL video position-id bug are separate features/paths." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Gemma4 text-only mm_token_type_ids defaulting and flash_attention s_aux=None are different model/integration bugs." + "reason": "Gemma4 text-only mm_token_type_ids issue is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to BigBirdTokenizer special-token registration." + "reason": "SwitchTransformers sparse-layer config bug and BigBirdTokenizer special-token registration are different problems." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "LasrFeatureExtractor center-arg regression and Gemma4 PLE device mismatch are different forward-path failures." + "reason": "Feature extractor center-arg regression is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "Base_model_prefix loading regression is not the same as Qwen3-VL weight-shape mismatch." + "reason": "Base_model_prefix loading breakage is not the same as Qwen3-VL weight-shape mismatch." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba SDPA/torch.compile issue and GraniteSpeech stale device override failure are unrelated CI/model bugs." + "reason": "torch.compile/SDPA failure in Bamba and stale GraniteSpeech device-override test are different bugs." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different import paths." + "reason": "Both are import failures, but for different missing symbols and different modules/causes." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 TorchScript parsing/IndentationError is unrelated to qwen3_omni_moe feature-length behavior." + "reason": "Python 3.13 indentation/JIT parsing issue is unrelated to the qwen3_omni_moe helper behavior." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "DAC latent conversion mismatch is a different code path from the position-id/key warning issue." + "reason": "DAC from_latents/STE mismatch has nothing in common with position-id key warnings." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformers is unrelated to BEiT pretrained preset loading with a legacy field." + "reason": "Sparse-layer creation bug in SwitchTransformers is unrelated to BEiT semantic-segmentation legacy-field loading." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Flash attention auxiliary-attribute crash and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are different runtime failures." + "reason": "Flash-attention sinkless-model crash and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are distinct." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "FlashAttention test failures and Trainer _gpu_broadcast_one NoneType errors are unrelated CI issues." + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate failure modes." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype drift in evaluate and a pipeline error message regression are not the same bug." + "reason": "Embedding dtype drift during evaluate is unrelated to the pipeline translation-task warning regression." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "gather_for_metrics truncation in tuple labels is unrelated to BeitImageProcessorFast reduce_labels behavior." + "reason": "Metric gathering drops tuple labels, while BeitImageProcessorFast reduce_labels returns only one label; not the same bug." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe length computation are different subsystems." + "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe feature-length helper are unrelated." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer protobuf error masking and missing import_utils helper are different import/exception-handling issues." + "reason": "Tokenizer error masking without protobuf and missing is_falcon_mamba_ssm_available are different import_utils issues." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 doc example errors are unrelated to BeitImageProcessorFast label reduction." + "reason": "SigLIP2 docs/model mismatch is unrelated to BEiT fast processor label reduction." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "Documentation/model-processor mismatch is unrelated to unexpected position-id keys." + "reason": "Documentation example errors do not match the position-id warning issue." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Exact duplicate: same title, same transformers>=5.4.0 Qwen3.5 MoE from_pretrained error, so they describe the same bug." + "reason": "Same concrete bug: Qwen3.5 Moe from_pretrained fails on transformers>=5.4.0; 45478 is a duplicate report of 45310." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "BitNet packed-weight unpacking during accelerate loading is unrelated to GraniteSpeech stale device override failures." + "reason": "BitNet packed-weights loading and GraniteSpeech stale device override are unrelated." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer added-space bug and embedding reinitialization after resize are different behaviors in different components." + "reason": "Tokenizer extra-space behavior and embedding reinitialization after resize are different code paths." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus output drift and Deepseek tokenizer incorrect results are different model-specific regressions." + "reason": "Pegasus output differences and Deepseek tokenizer regressions are distinct model behaviors." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "SwitchTransformers config bug and GraniteSpeech CI device override bug do not share the same underlying cause." + "reason": "SwitchTransformers sparse-layer config bug is unrelated to GraniteSpeech device override failures." }, { "left": "issue:43335", "right": "issue:45072", "accept": false, - "reason": "SwitchTransformers config bug is unrelated to bfloat16 dtype mismatches in inference." + "reason": "Config-layer creation bug is unrelated to bfloat16 dtype mismatches in inference." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "Local-path trailing-slash import warning and GLM-4.7-Flash unit-test load failure are different loading issues." + "reason": "Slash-path import warning and GLM-4.7-Flash unit-test loading failure are different loading issues." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "Both involve runtime mismatches, but one is CPU offload and the other is float16 AMP/CUDA loss computation; not the same bug." + "reason": "Device mismatch in Bark CPU offload is not the same as LwDetrImageLoss crashing under float16 AMP." } ] }, "evaluator_result": { "accept": true, - "feedback": "The canonical duplicate choice is well grounded: 45310 and 45478 have the same title/problem statement, and selecting the earlier issue as canonical is reasonable. The soft-edge verdicts are conservative overall, mostly staying at the level of title-based dissimilarity rather than asserting stronger evidence than the packet provides. The summary does not overstate the evidence." + "feedback": "The summary is well grounded in the packet: the only strongly supported duplicate pair is 45310/45478, which have identical titles and clearly describe the same Qwen3.5 MoE from_pretrained regression. The other soft-edge verdicts are conservative and generally supported by the issue titles, and the note that there are no PRs is correct." }, "evaluator_used": true, "retried": false, @@ -8160,8 +8319,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -8246,214 +8405,214 @@ "issue:45245|issue:45563", "issue:45459|issue:45712" ], - "prepared_review_unit_hash": "5a8a500586b956e19966c90c7e6be3acad6c1ef48fa8da9827c5027e4fbf9c71", + "prepared_review_unit_hash": "bb2afbdf2d45c7f91ff8b627236528b4e32c6ac93ca93f795524766078bcad9d", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36945, - "estimated_input_tokens": 9237, - "estimated_eval_tokens": 18730 + "serialized_chars": 36943, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most soft pairs are only loosely similar at a subsystem level and do not look like the same underlying bug/change. I would not merge any of the suggested pairs as duplicates; the only partial thematic overlap is among a few v5 tokenizer/regression reports, but they remain distinct failures.", - "confidence": 0.93, - "canonical_issue_reason": "If a single representative is needed, issue 45920 is the broadest open regression report in the set and has the clearest user impact across multiple models; however, the cluster does not actually contain a true duplicate group.", + "summary": "No soft pair appears to be the same underlying bug/change; the set spans distinct regressions across tokenization, loading, generation, attention, training, and export. I\u2019d use issue 45920 as the best representative issue because it is open, broad, and describes a multi-model v5 tokenizer regression. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45920 is the broadest, most representative open issue here: it summarizes a v5 tokenizer regression affecting multiple model families, making it a stronger canonical anchor than the narrower single-model reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the best anchor only because it is broad, active, and well-scoped as a v5 tokenizer regression affecting several models, even though the other items are not duplicates of it.", + "best_issue_reason": "issue:45920 is the best global issue candidate because it is open, high-impact, and framed as a cross-model regression rather than a model-specific edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "Tokenizer-ID regression vs. return_tensors being ignored with explicit text_kwargs; different API paths and failure modes." + "reason": "Both concern tokenizer/image processing inputs, but one is a multi-model token-id regression and the other is a `return_tensors` parameter precedence bug; different failure modes." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to position-id key warnings; different models and bugs." + "reason": "Siglip2 zero probabilities and unexpected position-id keys are unrelated inference bugs with different symptoms and code paths." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA/ReLU discussion is unrelated to the v5 import break from removing is_torch_fx_available." + "reason": "DSA ReLU discussion and removal of `is_torch_fx_available` are unrelated; no shared concrete bug." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Flash-attention generation failure and multi-GPU CUDA OOM are different concrete problems." + "reason": "Qwen3.5 flash-attention generation failure and Gemma4 multi-GPU OOM are different issues and would not be merged into one fix." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "AMP float16 crash in LwDetrImageLoss is not the same as bfloat16 dtype mismatches in other models." + "reason": "One is an AMP/CUDA crash in a loss module; the other is a dtype mismatch during bfloat16 inference." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus v4/v5 output drift is a separate regression from Qwen3.5 flash-attention generation issues." + "reason": "Transformers v4/v5 output drift for Pegasus and Qwen3.5 flash-attention generation are unrelated model behaviors." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Model loading breakage for Llama 3.2 Vision is unrelated to token-embedding resizing reinitializing lm_head." + "reason": "Model loading breakage for Llama 3.2 vision and post-init reinitialization of embeddings are separate bugs." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint resume key mismatch is a different failure than the Qwen3 Omni feature-length helper behavior." + "reason": "Checkpoint resume key mismatch and `_get_feat_extract_output_lengths` behavior do not describe the same underlying defect." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "LayoutLMv3 box-content error messaging and Trainer GPU broadcast NoneType are unrelated." + "reason": "LayoutLMv3 box-validation messaging and Trainer `_gpu_broadcast_one` NoneType handling are unrelated." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile + SDPA failure in Bamba is not the same bug as the qwen3_omni_moe length helper." + "reason": "Bamba `torch.compile`/SDPA failure is not the same as the qwen3_omni_moe feature-length helper issue." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to BeitImageProcessorFast reduce_labels returning only one label." + "reason": "IterableDataset multi-image finetuning failure and BeitImageProcessorFast `reduce_labels` bug affect different pipelines and code paths." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Custom token IDs being lost in config save/load is a different serialization issue than loading the wrong custom module." + "reason": "Broken custom token IDs after save/load and wrong custom module selection after save_pretrained are both persistence-related, but they are distinct bugs." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures are not the same concrete bug as post_init reinitializing output embeddings." + "reason": "Flash Attention 2 test failures and embedding reinitialization are not the same code-path issue." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Trailing-slash import warning and tensor-parallel OOM are unrelated load-time issues." + "reason": "Trailing-slash import warning and tensor-parallel OOM while loading Qwen2-57B are unrelated." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Removed helper import breaking trust_remote_code models is unrelated to DeepSeek tokenizer output regression." + "reason": "`is_torch_fx_available` removal breaking trust-remote-code models and DeepSeek tokenizer regressions are different regressions." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Flash-attention generation bug and stale num_return_sequences warning are different code paths." + "reason": "Flash-attention generation correctness and a stale warning about `num_return_sequences` are not the same bug." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP crash in LwDetrImageLoss and flash_attention.py s_aux=None crash are separate failures." + "reason": "Loss-module AMP crash and flash_attention `s_aux=None` crash involve different model components and failure conditions." }, { "left": "issue:43746", "right": "issue:46209", "accept": false, - "reason": "PEFT checkpoint loading problem is unrelated to FP8 inference failure on B200." + "reason": "PEFT adapter local checkpoint loading and FP8 inference failure on B200 are unrelated." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token regression and Gemma3 NaN embeddings are different bugs." + "reason": "Tokenizer special-token addition regression and Gemma3 NaN embeddings under sliding-window attention are different issues." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import break from mistral_common is unrelated to the flash_attention s_aux=None crash." + "reason": "Gemma4 processor import failure and flash_attention `s_aux=None` crash are separate problems." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export support issues do not match the sink-less flash_attention AttributeError." + "reason": "ExecuTorch export constraints for Gemma3n and flash-attention sink-less crash do not share a concrete fix." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure is unrelated to the bare except swallowing errors in FuyuBatchFeature." + "reason": "Wav2Vec2 save/tokenization failure and a bare-except swallowing errors in Fuyu are distinct bugs." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Embedding dtype drift between train/eval and gradient checkpointing in eval are different behaviors." + "reason": "Embedding dtype drift in evaluation and gradient checkpointing in eval mode are related only by eval, not by root cause." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "flash_attention.py s_aux=None crash is unrelated to Gemma4 PLE device mismatch." + "reason": "Flash_attention `s_aux=None` handling and Gemma4 PLE device placement mismatch are different runtime errors." }, { "left": "issue:36010", "right": "issue:37428", "accept": false, - "reason": "Missing GenerationMixin import and missing _flash_supports_window_size import are separate API breakages." + "reason": "Missing `GenerationMixin` import and missing `_flash_supports_window_size` import are separate import regressions." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Generation-output differences and lm_head serialization inconsistency are distinct v5 regressions." + "reason": "Pegasus v4/v5 output differences and `lm_head.weight` serialization inconsistency are different model state bugs." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to Qwen3.5 recurrent state reset." + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting change and GPU broadcast NoneType error are unrelated issues." + "reason": "Pipeline `top_k` nesting change and Trainer broadcast of NoneType are not the same underlying issue." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count limit error and stale generate() warning are not the same underlying defect." + "reason": "Category-cardinality limit and stale `num_return_sequences` warning are unrelated generation/configuration issues." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Protobuf decode-error masking and leftover dummy class cleanup are unrelated issues." + "reason": "Tokenizer error masking without protobuf and leftover dummy classes failing repo checks are different maintenance issues." } ] }, @@ -8472,8 +8631,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -8559,220 +8718,220 @@ "issue:29942|issue:36010", "issue:43278|issue:43329" ], - "prepared_review_unit_hash": "5ab9e8e6751ad943b2a07707fc5fb694fcc97535afd2e2643835a0b1d07d9256", + "prepared_review_unit_hash": "25084b9c9a816f144968988dbf9098f1e24bf5f206ac886266ff83a6c406f010", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37059, - "estimated_input_tokens": 9265, - "estimated_eval_tokens": 18786 + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose cluster of Transformers issues with superficial text similarity, but the pairs describe different model, tokenizer, processor, serving, and training failures. I would not merge any soft edge; the cluster is better treated as separate bugs that happen to share subsystem vocabulary.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 43873 is the broadest and most centrally framed bug report in the set, with active discussion and inbound references, so it is the best triage anchor if a single representative is needed. That said, the surrounding items are heterogeneous and do not appear to be duplicates of it.", + "summary": "The cluster is mostly a grab bag of unrelated single-issue reports. None of the soft pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue cleanly represents this cluster; the items span unrelated subsystems, so there is no strong canonical duplicate anchor.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the most generic and discussion-rich report, making it the most suitable single issue to represent the cluster for navigation and triage. It should be used as an anchor only, not as evidence that the other items are duplicates.", + "best_issue_reason": "issue:45406 is the most self-contained and actionable concrete bug report in the set, with a specific crash path and direct repro.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Both involve model components and memory-like behavior, but one is a dummy Conv2d in Sam3PixelDecoder while the other is RT-DETR cleanup/memory release; different bugs and code paths." + "reason": "Different subsystems and symptoms: a dummy conv layer in a decoder vs a model memory-leak cleanup issue." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "Both mention tokenization-related failures, but one is LayoutLMv2 NER/padding crashes and the other is Wav2Vec2 save_pretrained/tokenization failure; not the same defect." + "reason": "Tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are unrelated code paths." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "One is a Transformers 5 loading regression around base_model_prefix, the other is a rope-parameter validation TypeError on JSON-loaded lists; unrelated failure modes." + "reason": "Model loading with base_model_prefix is unrelated to rope-parameter validation type handling." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id inconsistency is a config mismatch, while Qwen3.5 flash-attention generation issues are an attention/runtime bug; no shared underlying code path." + "reason": "Different models and failures: eos_token_id config mismatch vs flash-attention generation bug." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError and Wav2Vec2 save/tokenization problems both touch tokenizers, but they affect different tokenizer implementations and different crashes." + "reason": "MLuke tokenizer AttributeError and Wav2Vec2 save/tokenization issues do not share the same bug." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "LasrFeatureExtractor unsupported center arg is an extractor API regression; qwen2_5_vl/qwen3_vl video input regression is a multimodal processor issue, not the same bug." + "reason": "Feature extractor argument incompatibility is unrelated to the video-input regression in Qwen VL models." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "Both are Mac/CLIP-ish vision issues at a high level, but one is a bus error on community CLIP models and the other is a SigLIP2 docs example mismatch/quantization failure." + "reason": "One is a MacOS bus error on a community CLIP model; the other is a documentation example with different errors." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma4 processor loading fails because of a missing mistral_common import, while the protobuf issue hides tokenizer errors when protobuf is absent; distinct root causes." + "reason": "Gemma4 processor import failure and protobuf-error masking are distinct failure modes with different causes." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are different image/model pipeline problems with no concrete shared bug." + "reason": "Flash Attention test failures and void segmentation map handling are unrelated." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error messaging for insufficient boxes and TrainerState num_input_tokens_seen not updating are unrelated; one is user-facing validation, the other is trainer accounting." + "reason": "Error-message clarity in LayoutLMv3 and Trainer token-count tracking are different bugs." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility regression and DeepseekV3MoE divergence are separate broad regressions with different triggers and affected subsystems." + "reason": "Python-version compatibility regression is unrelated to DeepSeekV3MoE implementation divergence." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Qwen3MoE FineGrainedFP8Config failure and quantized model saving failure both involve quantization, but one is loading/configuration and the other is serialization/saving." + "reason": "Both mention quantization, but one is an FP8 config failure and the other is quantized saving; different concrete problems." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 nonsensical outputs via AutoModel/pipeline and RT-DETR memory release are different model correctness vs lifecycle issues." + "reason": "Nonsensical SigLIP2 outputs and RT-DETR memory retention are separate issues." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export.export with torch_compilable_check and Wav2Vec2 save/tokenization failure are unrelated; one is export compatibility, the other is tokenizer/model save behavior." + "reason": "torch.export failures from torch_compilable_check are unrelated to Wav2Vec2 tokenizer/save issues." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Embedding dtype mismatch in evaluate and Qwen2-57B-A14B-Instruct tp OOM are different performance/correctness issues with no shared fix." + "reason": "Embedding dtype drift during eval and tensor-parallel OOM are not the same underlying bug." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "German DistilBertModel issue and GeneratorExp torch.jit.script export failure are entirely separate legacy bugs." + "reason": "German DistilBertModel issue and TorchScript generator-expression incompatibility are unrelated." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong tokenizer decoder type in Transformers v5 and AutoTokenizer ignoring tokenizer.json are both tokenizer-loading related, but they concern different logic and symptoms." + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are different tokenizer-loading bugs." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Local path trailing-slash import warning and quantization offloading not working are unrelated; one is path parsing, the other is model offload behavior." + "reason": "Local-path import warning and quantization offloading malfunction are not the same issue." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Both involve loading/resource behavior, but offloading with quantization is not the same as Gemma4 multi-GPU CUDA OOM; different models and failure mechanisms." + "reason": "Offloading/quantization behavior and Gemma4 multi-GPU OOM have different root causes and fixes." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "KV cache continuation wrong tensor indexing in transformers serve is a server-side runtime bug, while Gemma4Processor missing _tokenizer is an initialization/API bug." + "reason": "KV-cache continuation indexing in serve is unrelated to Gemma4Processor missing _tokenizer." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings with mixed-length batching are different tokenizer vs attention-masking defects." + "reason": "Runtime post-processor updates and Gemma3 NaN embeddings are different tokenizer/model bugs." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBirdTokenizer mask token registration is a specific tokenizer special-token bug, whereas FastSpeech2ConformerTokenizer loading failure is a different AutoTokenizer resolution problem." + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are distinct tokenizer issues." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and SigLIP2 AutoModel/pipeline nonsense outputs are unrelated model-specific issues." + "reason": "A dummy conv in Sam3PixelDecoder is unrelated to SigLIP2 incorrect outputs." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are separate training-flow bugs; not the same underlying defect." + "reason": "Eval-mode gradient checkpointing and resume checkpoint key mismatch are different training-state problems." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "OmDet-Turbo processor output size mismatch and RT-DETR memory release are different vision-model issues with different code paths." + "reason": "Processor size mismatch for OmDet-Turbo is unrelated to RT-DETR memory release." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and Gemma4 multi-GPU OOM are both GPU deployment issues, but they are distinct failures with different causes." + "reason": "cudaErrorAssert with auto device mapping is not the same as Gemma4 multi-GPU OOM." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error and Mac bus error on CLIP are unrelated runtime failures." + "reason": "Trainer NoneType broadcast error and MacOS bus error are unrelated runtime failures." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE layer drop and DeepseekV3MoE divergence are model-architecture discrepancies in different families, not the same bug." + "reason": "Smollm3 RoPE-layer count mismatch and DeepSeekV3MoE implementation divergence are different model-specific bugs." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor loading ImportError and transformers serve missing _tokenizer are both Gemma4-related, but one is dependency import failure and the other is a processor attribute bug." + "reason": "Both involve Gemma4 loading, but one is an external import dependency problem and the other is an internal processor attribute crash." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention 2 test failures and GenerationMixin import error are unrelated issues with different components and triggers." + "reason": "Flash Attention test failures and GenerationMixin import errors are unrelated." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "BF16-to-FP32 embedding dtype change during evaluate and undefined video multimodal token counting are unrelated bugs." + "reason": "Embedding dtype changes and undefined video-branch helpers in multimodal token counting are unrelated code paths." } ] }, @@ -8791,8 +8950,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -8859,8 +9018,8 @@ "issue:43575|issue:44387", "issue:43824|issue:44279", "issue:44442|issue:45072", - "issue:44743|issue:44811", "issue:29127|issue:35141", + "issue:44743|issue:44811", "issue:42915|issue:43931", "issue:43646|issue:44077", "issue:43262|issue:44206", @@ -8872,190 +9031,190 @@ "issue:28218|issue:32090", "issue:44442|issue:44898" ], - "prepared_review_unit_hash": "f7c68ca10fa952112fb56de2e4889b9df210de62a3a23d79c9d8f6978726d6c3", + "prepared_review_unit_hash": "752b27a85a78fdfff8ac2d8ef90a787cff16798e6e8f4b1d3a90a760c5970632", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 34969, - "estimated_input_tokens": 8743, - "estimated_eval_tokens": 17742 + "serialized_chars": 34979, + "estimated_input_tokens": 8745, + "estimated_eval_tokens": 17746 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are broadly similar only at a superficial level (regressions, loading errors, tokenizer/generation bugs) but they do not describe the same underlying defect or fix path. None of the soft pairs look like true duplicates or mergeable PR-equivalents.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This cluster is not a duplicate set; it mixes many unrelated bugs across tokenization, model loading, attention, audio, and training. No soft edge looks like the same underlying change or bug.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated subsystems and failure modes, so there is no single representative duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "No issue is a strong global canonical for this cluster; any choice would be arbitrary because the issues are thematically unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "Different bugs in different processors: BEiT Fast image label reduction vs Kosmos2.5 OCR long-input index error." + "reason": "Both are vision/ocr-related at a high level, but one is a BeitImageProcessorFast label-reduction bug and the other is a Kosmos2.5 long-OCR index error; different code paths." }, { "left": "issue:43761", "right": "issue:45859", "accept": false, - "reason": "Unrelated model behaviors: CLIPVision hidden_states regression vs Qwen3.5Moe rotary embedding CPU offload incompatibility." + "reason": "Unrelated failures: CLIPVisionModel hidden_states handling versus Qwen3_5Moe rotary embedding CPU offload compatibility." }, { "left": "issue:43723", "right": "issue:44554", "accept": false, - "reason": "Tokenizer loading failure in v5 is unrelated to an MPS attention correctness bug with mismatched value/query head dims." + "reason": "Tokenizer loading in v5 and an MPS attention correctness issue are different problems in different subsystems." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Both are import-time Python issues, but one is auto-docstring annotation handling and the other is Gemma 4 processor loading via mistral_common ReasoningEffort." + "reason": "auto_docstring annotation parsing crash is unrelated to a mistral_common import error in Gemma 4 processor loading." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "Different model-specific problems: DeepSeekV3 native implementation divergence vs D-FINE auxiliary loss omission when denoising is off." + "reason": "Different model implementations and bugs: DeepSeekV3 divergence versus D-FINE auxiliary-loss behavior." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Both involve image/processor code, but one is a label reduction bug and the other is LayoutLMv2Tokenizer crashing on NER padding/truncation." + "reason": "Both involve processor/tokenizer behavior, but the concrete bugs differ: label reduction versus NER/batched padding crashes." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Different Qwen3-VL/Qwen3-MoE failures: missing pad_token_id on text config vs sliding_window behavior in MoE models." + "reason": "Same broad Qwen prefix only; pad_token_id absence and sliding_window behavior are distinct bugs." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "One is tensor-parallel OOM loading a large Qwen2 model; the other is a resume_from_checkpoint key mismatch." + "reason": "Qwen2 large-model TP OOM and resume_from_checkpoint key mismatch are unrelated training/loading issues." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "Different detection/image bugs: BEiT label reduction vs LwDetrImageLoss AMP CUDA crash." + "reason": "Image processor label reduction and LwDetrImageLoss AMP/CUDA crash are different code paths and symptoms." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Qwen3.5 flash-attention generation issue is unrelated to the 2^24 category-count limit runtime error." + "reason": "Flash-attention generation bug and category cardinality limit error are unrelated." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Local path import warning and gradient checkpointing eval-mode restriction are unrelated behaviors." + "reason": "Local path import warning and gradient checkpointing eval-mode restriction are unrelated." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Large-model TP OOM is not the same as increased CUDA reserved memory under int4 quantization." + "reason": "Both mention memory/OOM, but one is TP loading OOM for a specific Qwen2 model and the other is int4 CUDA reserved-memory regression." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Qwen2.5-VL import error and a generic dependency issue with transformers do not describe the same failure." + "reason": "ImportError for a specific model class and a generic dependency issue are not the same underlying bug." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel bfloat16 inference." + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are separate failures." }, { - "left": "issue:44743", - "right": "issue:44811", + "left": "issue:29127", + "right": "issue:35141", "accept": false, - "reason": "Different code paths: Qwen3.5 recurrent state reset with cache vs Whisper processor.batch_decode skip_special_tokens handling." + "reason": "LayoutLMv3 error messaging and embedding reinitialization on resize are different issues." }, { - "left": "issue:29127", - "right": "issue:35141", + "left": "issue:44743", + "right": "issue:44811", "accept": false, - "reason": "Both are library bugs, but one is LayoutLMv3 error messaging and the other is token embedding resize/post_init reinitialization." + "reason": "Recurrent cache reset in Qwen3.5 modular code and Whisper batch_decode skip_special_tokens handling are unrelated." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Qwen3Moe FP8 config failure is a different issue from Qwen3-VL weight shape mismatch during loading." + "reason": "FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are different model-loading problems." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Transformers 5 custom model init breakage is not the same as patchtsmixer no-longer-allowed optional post_init behavior." + "reason": "Framework-wide custom initialization breakage is not the same as a patchtsmixer-specific post_init policy issue." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio processor chat template sampling-rate defaulting and LasrFeatureExtractor center-arg crash are unrelated processor regressions." + "reason": "Audio chat-template sample-rate defaulting and unsupported center arg in LasrFeatureExtractor are unrelated audio-processing bugs." }, { "left": "issue:44977", "right": "issue:45584", "accept": false, - "reason": "Both involve generation, but one is flash-attention behavior in Qwen3.5 and the other is Whisper empty transcription after align_special_tokens." + "reason": "Qwen3.5 flash-attention generation issue and Whisper empty-transcription failure after align_special_tokens are different bugs." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "Tokenizer/model loading with gguf_file and flash_attention s_aux=None AttributeError are different failures in different subsystems." + "reason": "GGUF loading/AutoTokenizer behavior and flash_attention s_aux=None crash are unrelated." }, { "left": "issue:33357", "right": "issue:38175", "accept": false, - "reason": "MacOS bus error on pretrained CLIP models is not the same as unexpected zero probabilities in siglip2-base-patch16-224." + "reason": "MacOS bus error with CLIP and zero-probabilities in SigLIP2 are different model/runtime failures." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "AddedToken keyword duplication and list_repo_files revision forwarding are unrelated API bugs." + "reason": "AddedToken special-argument duplication and list_repo_files revision forwarding are unrelated API bugs." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is distinct from Gemma3 NaN embeddings with sliding-window all-padding windows." + "reason": "Chunked generation under compiled forward and Gemma3 NaNs with mixed-length sliding windows are different correctness issues." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer extra-space after added token is unrelated to Trainer/_gpu_broadcast_one NoneType handling." + "reason": "Tokenizer added-token spacing bug and Trainer _gpu_broadcast_one NoneType error are unrelated." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading failure and Perceiver non-default-resolution interpolation failure are different model/input-path issues." + "reason": "FastSpeech2ConformerTokenizer loading failure and Perceiver non-default-resolution failure are different code paths." } ] }, @@ -9152,19 +9311,19 @@ "issue:44077|issue:44479", "issue:44568|issue:45479", "issue:44855|issue:45137", - "issue:43650|issue:46202", "issue:33290|issue:34689", + "issue:43650|issue:46202", "issue:44190|issue:44496", "issue:29127|issue:33357" ], - "prepared_review_unit_hash": "fd3087ff478a757181b3ae5286e11fa90c010c95c0ba65e80f249c103b3e38d8", + "prepared_review_unit_hash": "c2dabccf7bf08a3a55d78aba11249d3ad6762f50c5895b50c9a25c9c0c98a0ff", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, @@ -9179,180 +9338,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of unrelated Transformers issues. The soft-similarity pairs share surface themes like loading, configs, or regressions, but they do not look like the same concrete bug/change, so no duplicates are accepted.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "The set is mostly heterogeneous Transformers issues with only loose thematic overlap around loading/config/serialization. No pair appears to be the same underlying bug or change, so I rejected all soft edges. If a single representative issue is needed, the broadest fit is the from_pretrained/save_pretrained roundtrip issue.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45698 is the broadest and most representative open issue in the set: it concerns a general from_pretrained/save_pretrained loading-path bug, has inbound references, and is more central than the narrower model-specific or config-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong canonical fit because the set is heterogeneous. If forced, issue:45698 is the closest representative: it is an active, clearly scoped `from_pretrained`/`save_pretrained` custom-module loading bug, but it still does not unify the rest of the cluster.", + "best_issue_reason": "issue:45698 is the best single issue to represent the cluster because it is general enough to anchor a serialization/loading regression theme, whereas the others are mostly unrelated one-off model, backend, or training failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "Different failures: DeepSpeed/Adafactor OOM versus TrainerState token counter not updating." + "reason": "Different subsystems and symptoms: DeepSpeed/optimizer OOM vs TrainerState token counter not updating." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Both touch training behavior, but one is eval-mode gradient checkpointing and the other is quantization/offloading; different root causes." + "reason": "Eval-mode gradient checkpointing limitation is unrelated to quantization offloading behavior." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both are config API mismatches, but for different models and different missing fields; not the same bug." + "reason": "Both are config API issues, but they affect different fields and different model families; not the same bug." }, { "left": "issue:44479", "right": "issue:45290", "accept": false, - "reason": "One is a video-input regression across Qwen VL variants; the other is a chat-template crash on tool-call assistant messages." + "reason": "Different multimodal/chat-template regressions with different failure modes and code paths." }, { "left": "issue:44964", "right": "issue:45440", "accept": false, - "reason": "Separate model-loading divergence reports for different architectures and code paths." + "reason": "Different model-loading/divergence bugs; one is a loading failure, the other is implementation parity with remote code." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "Unrelated areas: BitNet packed-weight loading versus tokenizer loading failure." + "reason": "BitNet accelerate-loading unpacking bug is unrelated to AutoTokenizer loading FastSpeech2ConformerTokenizer." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Both are loading regressions, but they affect different versions and different mechanisms." + "reason": "Both are model-loading regressions, but for different root causes and versions; not plausibly one merged fix." }, { "left": "issue:42175", "right": "issue:46234", "accept": false, - "reason": "One is backend packaging for TensorFlow in extras; the other is a missing import_utils symbol for a specific availability check." + "reason": "Packaging/backend inclusion problem is unrelated to a missing utility symbol in import_utils." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType error and DeepSpeed OOM are distinct runtime failures." + "reason": "Trainer broadcast NoneType error and DeepSpeed Adafactor OOM are distinct failure classes." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer class loading versus a tensor conversion exception-handling bug; no shared code-path problem." + "reason": "Tokenizer-loading failure and a bare-except bug in Fuyu tensor conversion are unrelated." }, { "left": "issue:43550", "right": "issue:44743", "accept": false, - "reason": "One is torch.compile/SDPA failure in Bamba, the other is recurrent state reset in modular_qwen3_5 cache handling." + "reason": "Different model-specific runtime bugs: torch.compile/SDPA vs recurrent state reset with cache." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Both involve Qwen models, but one is a shape mismatch on load and the other is a save_pretrained regression." + "reason": "Both involve Qwen models, but one is a weight-shape mismatch and the other is a bad checkpoint save regression." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "Model implementation parity versus documentation references; not the same underlying issue." + "reason": "Model parity discrepancy and docs pipeline-removal cleanup are unrelated." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Different loading/import breakages with different symptoms and code paths." + "reason": "Different loading failures: multimodal model breakage vs GenerationMixin import path regression." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both concern config handling, but one is API parity for Gemma3TextConfig and the other is text_config unwrapping in multimodal CausalLM creation." + "reason": "Both mention config structure, but one is a field-discrepancy issue and the other is missing text_config unwrapping in composite multimodal configs." }, { "left": "issue:45290", "right": "issue:46209", "accept": false, - "reason": "Chat-template tool-call crash and FP8 inference on B200 are unrelated." + "reason": "Chat template crash is unrelated to FP8 inference failure on B200 hardware." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "CLIP hidden_states regression and attention-kernel divergence under torch.compile are different bugs." + "reason": "CLIPVision hidden_states regression is not the same as the attention-kernel divergence under torch.compile." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior versus Gemma eos_token_id mismatch; no shared failure mode." + "reason": "Quantized offloading behavior and Gemma eos_token_id inconsistency are different issues." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Both mention save/load flows, but one is dequantized-model saving and the other is wrong custom-module loading after save_pretrained." + "reason": "Both touch save/load paths, but one is dequantized model saving and the other is custom-module loading after save_pretrained." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Config reload bugs, but one drops fields with trust_remote_code and the other resets GPTNeoX rotary_pct on reload." + "reason": "Both are config persistence problems, but they involve different mechanisms and model classes." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Loading a specific model under transformers>=5.1 versus a missing backend mapping entry are separate issues." + "reason": "Model loading regression and missing BACKENDS_MAPPING entry for tensorflow_text are unrelated." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "PatchTSMixer post_init API policy and Qwen video-input regression are unrelated." + "reason": "PatchTSMixer post_init validation has no relation to Qwen video-input regression." }, { "left": "issue:44568", "right": "issue:45479", "accept": false, - "reason": "Tokenizer special-token insertion and sequence-classification loss computation are different subsystems." + "reason": "Tokenizer special-token insertion bug and sequence-classification zero-loss bug are different code paths." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 import parsing error versus DeepSpeed ZeRO3 deque underflow; different code paths." - }, - { - "left": "issue:43650", - "right": "issue:46202", - "accept": false, - "reason": "Placeholder/low-signal issue versus Gemma roundtrip tokenization failure; not a duplicate." + "reason": "Python 3.13 decorator parsing issue is unrelated to DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM with optimizer/deepspeed and a model-loading regression are not the same bug." + "reason": "DeepSpeed optimizer OOM and multimodal model loading regression are not the same bug." }, { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:43650", + "right": "issue:46202", "accept": false, - "reason": "Local dataset loading in an example script versus unrecognized-model config loading error." + "reason": "Very different problems: low-information issue title vs Gemma tokenization roundtrip failure." }, { - "left": "issue:29127", - "right": "issue:33357", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "LayoutLMv3 box-information error and CLIP MacOS bus error are unrelated model/runtime failures." + "reason": "Local dataset loading in an example script is unrelated to model-type recognition for an OLMo checkpoint." } ] }, @@ -9453,196 +9606,196 @@ "issue:43232|issue:43388", "issue:43504|issue:45698" ], - "prepared_review_unit_hash": "7db45a1fbb99769d1ddda5933085c6a8f2dfaa961fc92e6b67c1df011a5d219f", + "prepared_review_unit_hash": "79bf1557f817774166aa57d2a9eefb70e976ea4be07aee7b1cc7bc80edd03a05", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35194, - "estimated_input_tokens": 8799, - "estimated_eval_tokens": 17854 + "serialized_chars": 35191, + "estimated_input_tokens": 8798, + "estimated_eval_tokens": 17852 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft pairs share broad Transformers ecosystem themes, but none look like the same concrete bug or change. I would not merge any of the suggested pairs as duplicates.", - "confidence": 0.88, - "canonical_issue_reason": "issue:45698 is the strongest representative of a concrete, user-facing regression with substantial discussion and active impact around save/load behavior for custom modules.", + "summary": "These are mostly unrelated standalone issues spanning different models, pipelines, and backend/runtime regressions. None of the soft pairs look like true duplicates or the same concrete fix path, so all soft edges should be rejected.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must anchor the cluster, issue 45698 is the most broadly framed open regression and the most representative of a general load/save model-resolution problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45698 is the most broadly useful canonical issue in this set because it is specific, actionable, and close to a common model persistence workflow.", + "best_issue_reason": "Issue 45698 is the best cluster representative: it is open, broadly applicable, and concerns a common from_pretrained/save_pretrained failure mode rather than a model-specific edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Unrelated symptoms: pipeline top_k nesting vs missing PyTorch import for AutoModel." + "reason": "Different failures: pipeline output nesting vs missing PyTorch import." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Both touch pipeline-related APIs, but one is a tokenizer/processor regression and the other is an error-message wording issue." + "reason": "Both touch v5 regressions, but one is custom processor/tokenizer access and the other is a pipeline error message issue." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "Different code paths: torch.compile/SDPA failure vs resume_from_checkpoint key mismatch." + "reason": "torch.compile/SDPA OOM for a model versus checkpoint key mismatch during resume are unrelated." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "Different problems: running 3d_parallel.py vs tensor-parallel OOM on a specific model." + "reason": "3d_parallel.py runtime failure and TP load OOM are different code paths and symptoms." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is unrelated to checkpoint key mismatches." + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both are multimodal/video-adjacent, but one is a video input ValueError and the other is a batched chat-template crash." + "reason": "Both involve video/vision processors, but the concrete bugs differ: video input validation vs batched chat template padding behavior." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Different model families and failures: Gemma config API discrepancy vs Mistral-3 dtype mismatch." + "reason": "Config API discrepancy for Gemma3 text config is unrelated to image preprocessor/model dtype mismatch." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Both involve token/decoding behavior, but one is eos_token config inconsistency and the other is Whisper empty-transcription generation failure." + "reason": "EOS token config mismatch and empty-transcription generation failure are different problems." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "No shared underlying bug; ESM2 breakage and a Janus test failure are distinct." + "reason": "ESM2 runtime breakage and janus image-generation test failure are unrelated." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Both are config-related, but one is a Gemma API discrepancy and the other is strict loading for a different model config." + "reason": "Different config issues: one API inconsistency, one strict config loading failure." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Deepspeed Adafactor OOM is unrelated to post_init reinitializing tied embeddings." + "reason": "Deepspeed AdaFactor OOM is not the same as embedding resize reinitialization." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Model TP OOM and a key error loading MobileLLM are different load-time failures." + "reason": "TP load OOM and a model-loading key error are distinct failure modes." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "OOM with Adafactor in deepspeed is not the same as optimizer.step being called twice." + "reason": "Both mention deepspeed, but one is OOM during optimization and the other is double optimizer stepping." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Bare except swallowing interrupts is unrelated to Windows/ROCm import dependency failures." + "reason": "Bare except swallowing errors is unrelated to Windows ROCm import/dependency issues." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Same model family, but weight tying and FlashAttention failures are different concrete bugs." + "reason": "Mistral3 weight tying and FlashAttention failures are separate bugs." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Custom processor/tokenizer regression is unrelated to eos_token_id mismatch." + "reason": "Custom processor/tokenizer regression does not match an EOS token configuration inconsistency." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Dummy Conv2d in a decoder is unrelated to a CUDA OOM in Sam3Video." + "reason": "Dummy Conv2d in Sam3PixelDecoder is a model implementation issue, not the same as Sam3Video OOM." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Pipeline processor lookup and a missing vision-encoder attribute are distinct failures." + "reason": "Pipeline choosing image processors and a missing vision encoder attribute are different defects." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage is broader than, but not the same as, GPTNeoX rotary_pct reload behavior." + "reason": "Transformers 5 custom initialization breakage is unrelated to GPTNeoX rotary_pct reload behavior." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "Tokenizer/model loading with gguf_file is unrelated to an unrecognized model config error." + "reason": "GGUF loading regression and unrecognized model/config.json validation are not the same bug." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Config inconsistency for eos_token_id is unrelated to quantized save/load breakage." + "reason": "EOS token mismatch and quantized model save/load breakage are different persistence/config issues." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen video-input regression are different model-specific bugs." + "reason": "CLIP hidden_states regression and Qwen video-input regression affect different models and code paths." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with extended vocab is not the same as missing mm_token_type_ids defaults." + "reason": "LoRA merge collapse with extended vocab is unrelated to mm_token_type_ids defaults for text-only fine-tuning." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "StableLm config field mismatch and tensorflow_text backend registration are unrelated." + "reason": "StableLmConfig pad_token_idx and tensorflow_text backend mapping are unrelated config/import issues." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "TensorBoard/TensorFlow backend handling is unrelated to a missing import_utils symbol for falcon mamba availability." + "reason": "TensorBoard/TensorFlow backend handling is unrelated to a missing import_utils helper." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs sync logic and metric gathering label trimming are different subsystems." + "reason": "Generation kwargs sync and metric gathering label truncation are different distributed-training bugs." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Both are loading-related, but one is a legacy preset field and the other is incorrect custom-module restoration after save_pretrained." + "reason": "Legacy field loading in Beit semantic segmentation is unrelated to incorrect custom module resolution after save/load." } ] }, @@ -9747,20 +9900,20 @@ "issue:44112|issue:45600", "issue:46082|issue:46121" ], - "prepared_review_unit_hash": "91c7bcaae9dbbc9254d572d00d2ca29d509cf45970e6dd28677c6e3ea34a276a", + "prepared_review_unit_hash": "f6aeace59e45c407cce127fad06c174fe09770a3aeb2a8f2f29738bdd417745e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36829, + "serialized_chars": 36832, "estimated_input_tokens": 9208, "estimated_eval_tokens": 18672 }, @@ -9769,192 +9922,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue reports; only issue:43644 and issue:44534 are a true duplicate pair about the same v5 non-persistent-buffer corruption regression. No pull requests are present.", - "confidence": 0.96, - "canonical_issue_reason": "issue:44534 is the clearest representative of the duplicate buffer-junk regression; it states the bug broadly and matches issue:43644's earlier report.", + "summary": "This cluster is mostly noisy: the soft-similarity pairs share broad themes (generation, multimodal, config, CI) but do not look like the same underlying bug or change. The only near-duplicate in the item set is the non-persistent buffers regression (43644/44534), which is the best anchor issue, but none of the provided soft edges should be merged.", + "confidence": 0.93, + "canonical_issue_reason": "issue:43644 is the clearest and earliest report of the recurring 'non-persistent buffers filled with junk' regression, and issue:44534 is effectively a reworded follow-up of the same bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44534 is the best global issue representative because it captures the concrete regression succinctly and is the cleanest duplicate target for issue:43644.", + "best_issue_reason": "issue:43644 is the best cluster anchor because it has the most precise title and a direct duplicate-style re-report exists in the set; it is the strongest representative of the only obvious duplicate theme here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Different bugs: patchtsmixer post_init policy vs non-persistent-buffer initialization corruption." + "reason": "Both mention model init/post-init behavior, but one is about optional post_init in patchtsmixer and the other about non-persistent buffers being initialized with junk; different bugs." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Different code paths: missing GenerationMixin cache vs recurrent-state reset with cache." + "reason": "Both involve cache/state handling, but one is a missing GenerationMixin cache in chunked_prefill and the other is recurrent state reset in modular_qwen3_5; not the same code-path failure." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "Hidden-states output regression is unrelated to router load-balancing loss omission." + "reason": "Different subsystems and symptoms: CLIP vision hidden_states output vs load-balancing loss computation in router logits." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Janus image-generation test failure vs Qwen3.5 flash-attention generation bug." + "reason": "Both are generation failures, but they affect different models and appear to come from different mechanisms (test failure vs flash-attention behavior)." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation failure." + "reason": "Completely different issues: Sam3Video OOM versus Whisper empty-transcription failure after align_special_tokens." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Different failures: buffer junk regression vs GPTNeoX rotary_pct not persisting on reload." + "reason": "One is a v5 buffer-initialization regression; the other is a GPTNeoXConfig reload bug for rotary_pct. No shared underlying fix." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Both are video/KV-cache adjacent but the concrete failures and code paths differ." + "reason": "Both are v5-era regressions, but one is multimodal video input handling and the other is KV-cache continuation indexing in transformers serve." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading failure is unrelated to Phi-4-multimodal model loading." + "reason": "Unrelated: local dataset loading in a training script versus loading a specific multimodal model." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Test failure for Janus vs stale auto_mappings references to removed configs." + "reason": "A failed janus test and stale auto_mappings references are both CI-related only in the broadest sense; not the same bug." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import parsing issue vs flash_attention crash on s_aux=None." + "reason": "Different failure modes: Python 3.13 IndentationError in torch.jit parsing versus flash_attention.py crashing on s_aux=None." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs mentioning pipeline() are unrelated to a SigLIP2 inference correctness bug." + "reason": "Docs still referencing pipeline() is not the same as SigLIP2 producing nonsensical outputs with AutoModel/pipeline." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 device-mapping cudaErrorAssert is unrelated to the 2^24 category limit error." + "reason": "Unrelated: CUDA assert on A800 with auto device mapping versus a category-count limit error." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Dummy Conv2d implementation in Sam3PixelDecoder is not the same as the 640px-vs-224px processor mismatch." + "reason": "Different model/component bugs: a dummy Conv2d in Sam3PixelDecoder versus a processor producing the wrong image resolution for OmDet-Turbo." }, { "left": "issue:45803", "right": "issue:45987", "accept": false, - "reason": "Bare except swallowing errors is unrelated to StaticCache.get_seq_length returning a tensor." + "reason": "Bare except swallowing interrupts is unrelated to StaticCache.get_seq_length returning the wrong type." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Weight-shape mismatch on load is unrelated to the Qwen3-Next tied-weights warning." + "reason": "Both concern Qwen3-family weight handling, but one is a model loading shape mismatch and the other is an incorrect tied-weights warning across layers." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding reinit after resize and trailing-slash local path import warnings are different issues." + "reason": "Embedding reinit after resize and path-with-trailing-slash import warning are unrelated bugs." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "Sparse-layer creation bug is unrelated to the DebertaV2 import IndentationError." + "reason": "SwitchTransformers sparse-layer creation and DebertaV2 Python 3.13 import parsing are unrelated." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Buffer junk regression and GPTNeoX rotary_pct reload behavior are different bugs." + "reason": "The first is a buffer initialization regression; the second is a config value not persisting on reload. Different root causes." }, { "left": "issue:45632", "right": "issue:46234", "accept": false, - "reason": "Local-model cache collision is unrelated to the missing import_utils symbol." + "reason": "Cache-path collision for local trust_remote_code models is unrelated to the missing falcon-mamba availability helper." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Different config-related failures: patchtsmixer post_init policy vs granite_speech strict loading." + "reason": "Optional post_init in patchtsmixer and strict config blocking granite_speech are different config/init problems." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "torch.compile SDPA failure in Bamba is unrelated to the local dataset loading error." + "reason": "Bamba torch.compile+SDPA failure is unrelated to loading a local dataset in the image classification script." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "Python version support issue is unrelated to a JIT parsing/IndentationError." + "reason": "Python version compatibility problem versus an IndentationError in a scripted model import; not the same issue." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError is not the same as the KV-cache continuation crash." + "reason": "Video-input ValueError in Qwen3OmniMoe is not the same as serve-time KV-cache continuation tensor indexing." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks and wrong vision_position_ids are different multimodal bugs." + "reason": "Both are multimodal-template regressions, but assistant_masks all zero and wrong vision_position_ids are different bugs and likely different fixes." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge collapse with extended vocab is unrelated to incorrect visual encoder keys on save." + "reason": "LoRA merge collapse with extended vocabulary is unrelated to incorrect visual encoder keys saved in Qwen3.5." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "Tokenizer.json selection bug is unrelated to list_repo_files not forwarding revision." + "reason": "Repository tokenizer resolution and list_repo_files revision forwarding are separate hub API behaviors." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "CPU device placement in XLNet is unrelated to the bare-except error swallowing bug." + "reason": "XLNet CPU-side torch.arange and Fuyu bare-except error handling are unrelated implementation bugs." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformers is unrelated to local dataset loading." + "reason": "SwitchTransformers config behavior and a dataset-loading script issue do not share the same underlying defect." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "Stale CI test around device overrides is unrelated to auto_mappings referencing removed configs." + "reason": "Both are CI failures, but one is a stale test and the other is stale generated auto_mappings entries; different maintenance issues." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "LlamaConfig head_dim validation and rope-key list handling are separate configuration issues." + "reason": "LlamaConfig head_dim validation and rope-parameter JSON list handling are unrelated config-validation bugs." } ] }, @@ -9973,8 +10126,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -10029,8 +10182,8 @@ ], "soft_pairs": [ "issue:43116|issue:44190", - "issue:44938|issue:45542", "issue:45276|issue:45584", + "issue:44938|issue:45542", "issue:43295|issue:45276", "issue:43479|issue:45375", "issue:45245|issue:45584", @@ -10059,214 +10212,214 @@ "issue:44496|issue:45042", "issue:43673|issue:44336" ], - "prepared_review_unit_hash": "0f37a27bf2dc9ee032a05ea72083fb2f9b23bc06de5f2de9916335338cacc629", + "prepared_review_unit_hash": "8f450094836d89b70db97e7500dbe86d9f2c475bea56f591631d27c49f82797a", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36581, - "estimated_input_tokens": 9146, - "estimated_eval_tokens": 18548 + "serialized_chars": 36575, + "estimated_input_tokens": 9144, + "estimated_eval_tokens": 18544 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items form many unrelated bug reports across import errors, model configs, generation, pipelines, and docs. None of the soft pairs look like the same underlying issue closely enough to merge, so all soft edges are rejected. If one issue must anchor the cluster, the Phi-4 multimodal loading report is the clearest standalone bug report.", - "confidence": 0.92, - "canonical_issue_reason": "issue:44964 is a well-scoped, concrete model-loading regression with a specific model, failure mode, and version context; it is more representative than the other disparate reports.", + "summary": "These are mostly unrelated Transformers bug reports with a few topical similarities, but none of the soft pairs look like the same concrete bug/change. I would not merge any of the proposed pairs.", + "confidence": 0.82, + "canonical_issue_reason": "issue:41628 is the broadest, most user-facing regression in the set: a top-level import failure affecting the core `transformers` entrypoint.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44964 is the strongest standalone issue here: clear repro domain, explicit impacted model, and actionable bug description.", + "best_issue_reason": "issue:41628 is the best single anchor because it is generic, high-impact, and clearly described compared with the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Different example scripts and different failure modes; not the same bug." + "reason": "Both are example-script failures, but one is multi-label classification output behavior and the other is local dataset loading in image-classification; different code paths." }, { - "left": "issue:44938", - "right": "issue:45542", + "left": "issue:45276", + "right": "issue:45584", "accept": false, - "reason": "Python 3.14 import failure vs TensorFlow/tensorboard backend error; unrelated causes." + "reason": "Unrelated model behaviors: Gemma4 embedding resize propagation vs Whisper empty-transcription generation after token alignment." }, { - "left": "issue:45276", - "right": "issue:45584", + "left": "issue:44938", + "right": "issue:45542", "accept": false, - "reason": "Different model families and different generation/token embedding behaviors." + "reason": "One is a Python 3.14 import/load failure; the other is a TensorBoard-without-TensorFlow backend issue. Different failures and subsystems." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Both involve embedding/processor changes, but they affect different models and code paths." + "reason": "Both mention tokenizer/embedding-related regressions, but the concrete bug is different: processor/tokenizer API breakage vs token-embedding resize not propagating in Gemma4." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both are config-handling bugs, but on different model configs and different missing/defaulted fields." + "reason": "Both are config-shape issues, but they affect different model families and different missing/default fields." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Category-count limit error is unrelated to Whisper empty-transcription generation." + "reason": "Category-count runtime error is unrelated to Whisper empty transcription generation." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Both are generation-related, but one is output inconsistency and the other is a stale warning." + "reason": "Both are generation-related, but one is inconsistent outputs under compiled forward and the other is a stale warning about `num_return_sequences`." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Different import symbols and different modules; same broad symptom only." + "reason": "Both are import errors, but for different symbols and modules; no evidence they stem from the same export bug." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "Tokenizer NER/padding crash is unrelated to serve KV-cache continuation indexing." + "reason": "Tokenizer crashes on NER/padding are unrelated to KV-cache tensor indexing in the serve `/v1/responses` path." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Placeholder issue title vs audio feature extraction bug; no overlap." + "reason": "No meaningful overlap: one is a vague data request, the other is an audio feature extraction bug." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import latency and sys.modules safety are different problems." + "reason": "Import performance regression vs unsafe `sys.modules` access; related to import time, but not the same bug or fix." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both touch flex_attention, but one is a compile/runtime failure and the other is a version-check/import gate bug." + "reason": "Both involve `flex_attention`, but one is compiled flex attention failing on torch>=2.9 and the other is an incorrect version check for AuxRequest import." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Weight-tying bug in Mistral3 is unrelated to Qwen3.5 save_pretrained corruption." + "reason": "Different model-specific regressions: tied `lm_head` weights in Mistral3 vs incorrect save_pretrained checkpoint state in Qwen3.5." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Image-utils import error vs Qwen3.5 chat crash; no shared underlying bug." + "reason": "Nonexistent `image_utils` import and Qwen3.5 chat crash are unrelated failures." }, { "left": "issue:38175", "right": "issue:43065", "accept": false, - "reason": "Siglip2 zero probabilities and Sam3PixelDecoder dummy conv are unrelated model bugs." + "reason": "Different model families and symptoms: zero probabilities in SigLIP2 vs dummy Conv2d in Sam3PixelDecoder." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Both involve config loading, but one loses fields with trust_remote_code while the other lacks model_type entirely." + "reason": "Config loading with `trust_remote_code` missing fields is distinct from an unrecognized model missing `model_type`." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI stale-device test failure is unrelated to Phi-4 multimodal model loading." + "reason": "CI failure in GraniteSpeech tests is unrelated to Phi-4-multimodal model loading." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Both concern loading models, but one is a specific Phi-4 load failure and the other is a broken auto-mapping reference." + "reason": "Phi-4-multimodal load failure and stale auto_mappings references to removed Sam3LiteText configs are separate issues." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Different generation bugs with different triggers and outcomes." + "reason": "A stale generation warning does not share the same code path as Whisper failing on empty transcription." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different models and different loading failure modes." + "reason": "Both are model loading issues, but they concern different checkpoints and different failure modes." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Two distinct import errors in different modules; not the same missing export." + "reason": "Both are import errors, but they involve different missing names and different modules; not the same regression." }, { "left": "issue:44734", "right": "issue:45632", "accept": false, - "reason": "Server KV-cache crash and trust_remote_code cache-path collision are unrelated." + "reason": "Server KV-cache continuation crash is unrelated to local trust_remote_code cache-path collisions." }, { "left": "issue:44561", "right": "issue:45632", "accept": false, - "reason": "Removal of a helper API breaking trust_remote_code models is not the same as local cache path collisions." + "reason": "Removal of `is_torch_fx_available` breaking remote-code models is unrelated to cache-path collisions for local models." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and Sam3LiteText auto-mapping cleanup are unrelated." + "reason": "Qwen3-MoE sliding-window behavior and stale Sam3LiteText auto-mapping references are different code paths." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Both mention resize_token_embeddings, but they affect different model implementations and symptoms." + "reason": "Both are embedding/processor-style regressions, but one is processor/tokenizer API breakage and the other is decoder embedding resize propagation in t5gemma." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Saving dequantized Ministral models and FlashAttention failures are different bugs." + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral-3 are separate issues." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "ImageProcessorFast label reduction bug vs pipeline documentation mismatch; not the same issue." + "reason": "BEiT image processor label reduction bug is not the same as a documentation mismatch in text classification pipeline behavior." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification pipeline bug and notebook syntax error are unrelated." + "reason": "Video-classification pipeline image-processor lookup and a notebook syntax error are unrelated." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Model recognition/config error vs image processor torchvision dependency issue." + "reason": "Unrecognized model/config loading and PIL backend torchvision dependency regression are different failures." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Generation cache regression and ANSI color output in loading_report are unrelated." + "reason": "Generation cache missing in v5 chunked prefill is unrelated to ANSI codes emitted by loading_report." } ] }, @@ -10368,20 +10521,20 @@ "issue:44016|issue:44062", "issue:43295|issue:44779" ], - "prepared_review_unit_hash": "3bb1ca519485302860c8475a412a338f913ff47ed262e84926294e3691614a13", + "prepared_review_unit_hash": "a41bd36a97de021b518022b58d1be54f8c9ac9ddcbfa2db133258604bf477892", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35610, + "serialized_chars": 35609, "estimated_input_tokens": 8903, "estimated_eval_tokens": 18062 }, @@ -10390,192 +10543,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly loose similarity matches across unrelated bugs (imports, tokenizer behavior, model configs, docs, training utilities). None of the soft pairs looks like the same concrete underlying bug/change, so all are rejected. No PRs are present; I chose a broad, user-facing regression issue as the best issue representative.", - "confidence": 0.82, - "canonical_issue_reason": "issue:43295 is a clear, high-signal regression report with concrete reproduction impact and a broad API breakage surface.", + "summary": "The cluster is heterogeneous: mostly unrelated Transformers bugs spanning model loading, tokenizer behavior, config/init issues, docs, and dependency/import regressions. None of the soft pairs look like the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43295 is the best anchor only in a weak sense: it is a concrete, user-facing regression report with clear reproduction, while the rest of the cluster is a mix of unrelated model-, config-, and docs-level problems.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43295 is the most representative cluster member because it is a direct regression affecting existing code paths and has a clearly stated failure mode.", + "best_issue_reason": "Issue 43295 is the strongest single triage candidate because it is a specific regression with direct breakage and clearer scope than the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "Different model families and failure modes: sparse-layer config construction vs missing auxiliary losses when denoising is off." + "reason": "Different bugs: SwitchTransformers sparse-layer construction vs D-FINE auxiliary-loss omission." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both mention flash attention, but the bugs are different: one is a Ministral-3 FlashAttention failure, the other is an AttributeError on s_aux=None in sink-less models." + "reason": "Different flash-attention failures with different root causes and code paths." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Unrelated dependency/package issues; one is missing TensorFlow in an extra install, the other is a generic transformers dependency complaint." + "reason": "Backend install/dependency packaging issue vs a generic transformers dependency problem." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Different code paths: metric gathering drops tuple labels vs incorrect auxiliary-loss normalization." + "reason": "Metrics label truncation vs auxiliary-loss normalization are unrelated." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "Optimizer argument handling is unrelated to a tie_word_embeddings warning in Qwen3.5 fine-tuning." + "reason": "Unused SGD optimizer args is unrelated to a Qwen tie_word_embeddings warning." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "A notebook syntax error and a feature-extraction function bug are not the same underlying issue." + "reason": "Notebook syntax error vs audio feature extraction bug; not the same issue." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both involve Qwen models, but one breaks model loading and the other corrupts save_pretrained output; different code paths." + "reason": "Both are Qwen regressions, but one is model loading and the other is save_pretrained checkpoint corruption." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to auxiliary-loss usage in D-FINE." + "reason": "Gradient checkpointing in eval mode is unrelated to auxiliary-loss behavior." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Documentation mismatch for a pipeline is not the same as an unrecognized-model config loading error." + "reason": "Docs mismatch for return_all_scores vs an unrecognized-model config error." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "No shared underlying bug; one is a non-informative issue title, the other is an FP8 inference failure on B200." + "reason": "No meaningful relation; one title is malformed, the other is an FP8 inference failure." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Different failures: unknown model type in config vs AutoModelForCausalLM not unwrapping text_config for composite multimodal configs." + "reason": "Both involve model config/loading, but one is unknown-model detection and the other is AutoModel text_config unwrapping." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Token embedding resize behavior in T5Gemma is unrelated to Whisper generation failing after align_special_tokens." + "reason": "Different model behaviors: T5Gemma embedding resize vs Whisper empty-transcription generation." }, { "left": "issue:45372", "right": "issue:46234", "accept": false, - "reason": "Both are import-related, but they are different missing symbols from different modules and likely different fixes." + "reason": "Different missing imports/symbols affecting different subsystems and models." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error and embedding reinitialization after resize_token_embeddings are different bugs." + "reason": "Trainer GPU broadcast TypeError vs token-embedding reinitialization after resize are unrelated." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Different configs and symptoms: default multimodal sub-config initialization vs missing pad_token_id on Llama4Config." + "reason": "Multimodal config defaults being created vs a missing pad_token_id attribute are different config bugs." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Auxiliary load-balancing loss omission and missing mm_token_type_ids default are distinct model/training behaviors." + "reason": "MoE load-balancing loss handling vs Gemma 4 mm_token_type_ids defaulting are different concrete issues." }, { "left": "issue:38175", "right": "issue:44617", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and CUDA OOM in Sam3Video are unrelated." + "reason": "SigLIP2 zero probabilities and Sam3Video OOM are unrelated failures." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Dtype propagation in Blip2 loading is unrelated to incorrect checkpoint saving for Qwen3.5." + "reason": "Dtype propagation in Blip2 vs Qwen3.5 save_pretrained checkpoint bugs are different code paths." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both involve Qwen3.5/3VL, but one is a load-time failure and the other is a save_pretrained serialization bug." + "reason": "Both are Qwen3.5 multimodal, but one is loading failure and the other is incorrect visual-encoder keys on save." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "A pipeline processor lookup bug is not the same as remote_code breakage across recent versions." + "reason": "Video-classification processor lookup bug vs remote_code compatibility regressions are unrelated." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Different import errors from different namespaces and causes." + "reason": "Different import failures from different modules and symbols." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Video-classification processor lookup and Gemma4 device placement are unrelated." + "reason": "Pipeline processor-selection bug vs Gemma4 device-map mismatch are unrelated." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and likely different tokenizer code paths." + "reason": "Both are tokenizer regressions, but they affect different model families and likely different implementations." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "A generic dependency issue is not the same as a version floor bump to an unreleased PEFT release." + "reason": "Generic dependency issue vs an overly new MIN_PEFT_VERSION requirement are not the same bug." }, { "left": "issue:44351", "right": "issue:45278", "accept": false, - "reason": "A missing HybridCache symbol and a broad collection of import errors are not the same concrete bug." + "reason": "Single missing HybridCache export vs broad import errors after upgrade are not the same underlying change." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both are documentation-related, but one is a broken example and the other is stale pipeline docs." + "reason": "Different documentation problems in different examples and API surfaces." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken keyword-argument conflict are unrelated." + "reason": "Notebook syntax error vs AddedToken keyword collision are unrelated." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Both are regression-style tokenizer complaints, but they concern different models and different failure mechanisms." - }, - { - "left": "issue:43295", - "right": "issue:45812", - "accept": false, - "reason": "Different affected models and symptoms: processor/tokenizer API breakage vs wrong token IDs for Granite models." - }, - { - "left": "issue:43388", - "right": "issue:45216", - "accept": false, - "reason": "Metric gathering/label truncation is unrelated to incorrect checkpoint saving." + "reason": "Different regression scopes: custom multimodal processor/tokenizer handling vs DeepSeek tokenizer output correctness." } ] }, @@ -10594,8 +10735,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -10681,220 +10822,220 @@ "issue:43525|issue:43881", "issue:43901|issue:44246" ], - "prepared_review_unit_hash": "1ce5a2d442dcde25ccb5b098c83d57a433acba72179c60a400ab5410799b99c5", + "prepared_review_unit_hash": "75d501850e0977cd81c33423de60caa7a08c4ce636a3a92a5ab6504baed26f1d", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36719, - "estimated_input_tokens": 9180, - "estimated_eval_tokens": 18616 + "serialized_chars": 36711, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft pairs span unrelated tokenizer, loading, attention, CI, docs, and performance regressions. None of the candidate edges looks like the same underlying bug/change, so all soft links are rejected.", - "confidence": 0.92, - "canonical_issue_reason": "No single issue is a safe canonical representative; the set mixes unrelated bugs and regressions rather than one duplicate theme.", + "summary": "The cluster is highly heterogeneous: the soft-linked pairs mostly share only broad vocabulary like loading, tokenizer, or model regressions, but they point to different subsystems and root causes. I would not merge any of the suggested soft edges.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue stands out because the items span unrelated bugs, regressions, docs, CI, and dependency problems rather than one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44462 is the broadest representative of the tokenizer-loading subset and is the closest fit among the available issues, but it still does not cover the whole cluster.", + "best_issue_reason": "No issue is a strong global canonical anchor for this cluster; the set is too diverse to represent one concrete bug or change.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Different failure classes: compiled flex_attention on torch>=2.9 vs an optional TensorFlow/backend import issue." + "reason": "Different failures: compiled flex_attention on torch>=2.9 vs missing TensorFlow backend handling for tensorboard-only installs." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python version support breakage and a PEFT version bump are unrelated release/dependency problems." + "reason": "Unrelated dependency/version problems: Python 3.9 runtime incompatibility vs an unreleased PEFT minimum-version bump." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Both are model-loading reports, but they target different models and likely different load paths." + "reason": "Both are model-loading regressions, but for different models and distinct code paths; not the same underlying bug." }, { "left": "issue:43650", "right": "issue:46132", "accept": false, - "reason": "A vague data request is unrelated to AttentionInterface.register changing function behavior." + "reason": "Completely unrelated: a vague data-related issue vs AttentionInterface.register changing behavior." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "A Python-version compatibility issue is not the same as a model-specific loading failure." + "reason": "Different issues: Python version support vs loading a specific model repository." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is AutoTokenizer ignoring tokenizer.json while the other is a Mistral regex-patching crash." + "reason": "Tokenizer-related, but separate bugs: tokenizer.json selection vs a Mistral regex patch crash." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Video processor shape handling is unrelated to a generic dependency issue." + "reason": "Video processor shape bug vs generic dependency issue; no shared code path." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python compatibility and dependency problems are different root causes." + "reason": "Different classes of problems: interpreter support vs dependency resolution/installation." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Save-time tokenizer_class mismatch and load-time tokenizer.json resolution are different serialization paths." + "reason": "Both involve tokenizers, but one is save_pretrained serialization metadata and the other is AutoTokenizer loading behavior." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex patching and list_repo_files revision forwarding are unrelated code paths." + "reason": "Tokenizer regex crash vs repo file listing not forwarding revision; unrelated subsystems." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Tied-weight serialization depends on device placement, while this gemma4 issue is about resize_token_embeddings not updating related layers." + "reason": "Both touch weights/vocab resizing, but one is serialization inconsistency and the other is embedding resizing propagation." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "All-zero assistant masks in multimodal chat templating is unrelated to Qwen2.5-VL rope index scaling." + "reason": "Different multimodal bugs: assistant mask generation vs rope index temporal scaling." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention failure in Ministral-3 and a Kosmos2.5 OCR index error are different model-specific bugs." + "reason": "Different model-specific inference bugs: FlashAttention failure vs OCR index error." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both touch repo hygiene/CI, but one is stale auto_mappings references and the other is leftover dummy classes." + "reason": "Both are cleanup/CI-related, but one is stale auto_mappings references and the other is dummy class leakage." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model saving and dequantized offload loading are opposite workflows with different failure modes." + "reason": "Quantization/offload related, but one is saving quantized models and the other is loading dequantized models with device_map=auto." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Tokenizer special-token registration vs D-FINE auxiliary-loss logic; unrelated." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor save files and gemma-4 ZeRO-3 from_pretrained are different loading/saving problems." + "reason": "Different save/load issues: processor.save_pretrained missing files vs gemma-4 ZeRO3 from_pretrained behavior." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Processor save file loss is unrelated to a tokenizer/filtering performance regression." + "reason": "Processor saving bug vs unrelated data-filtering slowdown; no common defect." }, { "left": "issue:43065", "right": "issue:44898", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and Perceiver non-default resolution failure are different vision-model bugs." + "reason": "Different vision-model issues: dummy Conv2d in Sam3PixelDecoder vs Perceiver non-default resolution failure." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "Missing rms_norm_eps precision drift and torch.split return values in a MoE indexer are distinct code-path bugs." + "reason": "Unrelated bugs: missing rms_norm_eps precision issue vs torch.split return-value handling." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error and gemma4 PLE device mismatch are unrelated." + "reason": "Notebook syntax error vs Gemma4 device mismatch during forward; unrelated." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated issues." + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; unrelated." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM prediction errors and Deepseek tokenizer regressions affect different models and logic." + "reason": "Both are v5 regressions, but one is CamemBERT masked-LM output and the other is DeepSeek tokenizer behavior." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer loading problems and ESM2 model breakage are not the same underlying bug." + "reason": "Different loading regressions for different model families; no evidence they are the same bug." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "Both mention attention, but Ministral FlashAttention and compiled flex_attention on new torch versions are different implementations." + "reason": "Both mention attention kernels, but one is a model-specific FlashAttention issue and the other is compiled flex_attention on newer torch." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs mentioning return_all_scores and incorrect SigLIP2 outputs are different problem types." + "reason": "Docs mismatch vs incorrect SigLIP2 model outputs; unrelated." }, { "left": "issue:43450", "right": "issue:44617", "accept": false, - "reason": "Incorrect batched video-processor shapes and Sam3Video OOM are not the same bug." + "reason": "Video processor shape bug vs Sam3Video memory exhaustion; different failure modes." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer config serialization mismatch and gemma-4 ZeRO-3 loading are unrelated." + "reason": "Tokenizer save metadata issue vs gemma-4 ZeRO3 loading issue; unrelated." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "A missing nn import in sentence-transformers and a TensorBoard-without-TF backend error are different import-time dependency issues." + "reason": "Import-time NameError in sentence-transformers integration vs missing TensorFlow backend handling; unrelated." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Llama4Config missing pad_token_id and glm-4v-9b loading failure are distinct model/config problems." + "reason": "Different model-loading/config issues: missing pad_token_id on Llama4Config vs GLM-4V loading failure." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Pipeline docs drift and intermittent import slowness are unrelated." + "reason": "Docs/behavior mismatch vs import performance regression; unrelated." } ] }, @@ -10913,8 +11054,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -10998,208 +11139,208 @@ "issue:44162|issue:45200", "issue:43901|issue:45003" ], - "prepared_review_unit_hash": "68e3cd471b47eb6971dfb5e7674839ed11ff6f8bc9fc3e67314803802eb015f5", + "prepared_review_unit_hash": "551a385fb73b4e9cb3b4d961f0f81cd1dcd47b73a05ecf4b17d1862e40dd43e1", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36330, - "estimated_input_tokens": 9083, - "estimated_eval_tokens": 18422 + "serialized_chars": 36333, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of mostly unrelated Transformers issues: import/export regressions, model-specific loading bugs, cache/offline behavior, performance complaints, and a few doc/notebook problems. I do not see a single underlying duplicate bug tying them together, so all soft edges should be rejected.", - "confidence": 0.29, - "canonical_issue_reason": "Issue 44521 is the strongest standalone bug report in the set: it is concrete, recent, has multiple inbound references, and describes a reproducible multimodal chat-template failure.", + "summary": "This cluster is loosely related around Transformers v5 regressions, especially import/export and model-loading breakages, but the individual items are mostly distinct bugs or docs issues. I do not see any true duplicate pairs here.", + "confidence": 0.37, + "canonical_issue_reason": "issue:41628 is the broadest and most generic user-facing regression in the set: a top-level import failure from `transformers`, which matches the general import/export theme better than the more model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44521 is the best representative issue overall because it is specific, user-impacting, and well-scoped to a concrete code path rather than a vague symptom or docs-only problem.", + "best_issue_reason": "issue:41628 is the best representative issue because it is concrete, broadly scoped, and closest to the common import/export regression pattern seen across the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Different failures: missing `pad_token_id` on a config vs `trust_remote_code` returning an incomplete config object." + "reason": "Both are config-related, but one is a missing `pad_token_id` field on Llama4 and the other is missing fields from `AutoConfig` with `trust_remote_code`; different failure modes and code paths." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "A notebook syntax error is unrelated to the `init_empty_weights` TypeError during model loading." + "reason": "A notebook syntax error and an `init_empty_weights` TypeError are unrelated." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Both involve chat templating, but one is an audio sampling-rate default and the other is multimodal assistant-mask generation." + "reason": "Both involve `apply_chat_template`, but one is an audio sampling-rate default and the other is multimodal assistant mask generation; not the same bug." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Torch.compile attention divergence is unrelated to ANSI codes emitted by `loading_report`." + "reason": "Attention-kernel divergence on `torch.compile` is unrelated to ANSI codes in loading reports." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "An import/export error for `AutoImageProcessor` is not the same bug as Python-version incompatibility in 5.1.0." + "reason": "Top-level import failure vs Python-version compatibility regression; not the same underlying issue." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "These are different code paths: a Float8 storage lookup error vs a Qwen chat-model crash." + "reason": "Different model/chat failures with no shared concrete code path." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Chunked prefill cache handling and CLIPVision hidden-state output are separate regressions." + "reason": "Chunked-prefill cache handling and CLIP hidden-states behavior are separate regressions." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "ANSI code leakage in loading reports is unrelated to a wrong docstring for `position_ids`." + "reason": "Terminal ANSI formatting and a docstring problem are unrelated." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "A model max-length question for DeBERTa is unrelated to the missing `transformers.onnx` module import." + "reason": "A `model_max_length` question on DeBERTa-V3 is unrelated to a missing `transformers.onnx` module import." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen/MoE, but one is a cross-version LoRA compatibility issue and the other is a specific model-loading regression in v5." + "reason": "Both mention Qwen3 MoE, but one is LoRA portability across versions and the other is model loading on v5; too different to be the same bug." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Reading `position_ids` after buffer registration is not the same as `from_pretrained` failing for gemma-4 zero3." + "reason": "A `position_ids` buffer access error is unrelated to the Gemma-4 zero3 `from_pretrained` issue." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Slow import performance and an unrecognized-model/config error are unrelated bug classes." + "reason": "Import slowness is not the same as an unrecognized model/config loading failure." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Missing load-balancing loss and missing auxiliary losses are both loss-related, but they affect different models and mechanisms." + "reason": "Both concern losses not being applied, but one is MoE load-balancing loss and the other is D-FINE auxiliary losses; different model families and mechanisms." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import latency has nothing to do with inconsistent EOS token configuration." + "reason": "Import performance and Gemma eos-token configuration are unrelated." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "A pipeline dispatch bug for video classification is unrelated to `_torch_extract_fbank_features()`." + "reason": "Video pipeline image-processor lookup and `_torch_extract_fbank_features()` are different subsystems." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "Loading-report ANSI output is unrelated to a GLM-5 RoPE implementation discussion." + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "A missing exported model class is not the same issue as Python 3.9+ incompatibility." + "reason": "Missing Qwen2.5 VL export and Python 3.9 compatibility are different issues." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Different missing symbols from different packages; these are not the same underlying import bug." + "reason": "Both are import errors, but for different symbols from different packages and likely different fixes." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different missing exports and likely different fixes." + "reason": "Missing model-specific export vs missing `HybridCache` export; similar symptom, not the same concrete bug." }, { "left": "issue:43976", "right": "issue:44964", "accept": false, - "reason": "Python-version support is unrelated to loading a specific multimodal model." + "reason": "Python support regression and inability to load a specific multimodal model are not the same underlying problem." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline cache loading and `AttentionInterface.register` behavior are unrelated." + "reason": "Offline cache loading and `AttentionInterface.register` behavior change are unrelated." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer output regressions and embedding-resize propagation are different bugs." + "reason": "Tokenizer correctness and `resize_token_embeddings` output-weight syncing are different bugs." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "A buffer access error is not the same as a major post-upgrade filtering slowdown." + "reason": "`position_ids` buffer access and a reported post-upgrade slowdown are not the same issue." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Generation cache handling and removal of `is_torch_fx_available` are separate regressions." + "reason": "Both are v5 regressions, but one is chunked-prefill cache handling and the other is `trust_remote_code` breakage from a removed helper; not mergeable as one fix." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Missing auxiliary loss is unrelated to an AMP/CUDA crash in LwDetrImageLoss." + "reason": "MoE load-balancing loss omission and an AMP/CUDA crash in LwDetrImageLoss are unrelated." }, { "left": "issue:45081", "right": "issue:45362", "accept": false, - "reason": "A Mistral tokenizer regex patch crash is unrelated to a Qwen chat loading crash." + "reason": "Mistral tokenizer regex patching and Qwen3.5 chat crashes are different code paths." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "A model-specific Flash Attention unsupported warning is not the same as a general Flash Attention 2.0 issue." + "reason": "The first is a specific unsupported-model warning for Flash Attention 2.0; the second is just a generic topic title with no evidence of the same bug." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 multimodal token-type IDs are unrelated model-specific bugs." + "reason": "ESM2 breakage and Gemma 4 token-type defaults are unrelated." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "A documentation mismatch for `return_all_scores` is unrelated to unsafe `sys.modules` access in code." + "reason": "A pipeline docs mismatch and unsafe `sys.modules[]` access are separate concerns." } ] }, @@ -11218,8 +11359,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -11303,30 +11444,30 @@ "issue:43931|issue:44368", "issue:44060|issue:44077" ], - "prepared_review_unit_hash": "1e21b76f8d93b82d70bd34c470f258a06e484dd4dfb8d22979dfb7ebde524252", + "prepared_review_unit_hash": "836ecc2a4da635fd2b18b8fb60d4386ff055ef6cfd79b464e2778510b7c9789f", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36607, - "estimated_input_tokens": 9152, - "estimated_eval_tokens": 18560 + "serialized_chars": 36613, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs span unrelated bugs across Qwen/SAM/Deepspeed/tokenizer/config/loading paths, so none look like true duplicates of the same underlying change or defect.", - "confidence": 0.97, + "summary": "The soft pairs are mostly broad theme matches (DeepSpeed, Qwen, resize/tokenization, config handling) but not the same concrete bug or change. I would reject all proposed duplicate edges and keep these issues separate.", + "confidence": 0.94, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -11336,175 +11477,175 @@ "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "ImportError for Qwen2.5-VL image processor is unrelated to a Deepspeed/Bert zero-sized tensor crash." + "reason": "Different failures in unrelated code paths: a Qwen2.5-VL import error vs a DeepSpeed ZeRO3/BERT index error." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BEiT label reduction bug and unsafe sys.modules access are different code paths and failure modes." + "reason": "One is a BeitImageProcessorFast label-reduction bug; the other is unsafe sys.modules access in modeling_utils. No shared underlying defect." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config error is not the same issue as modeling_utils sys.modules access." + "reason": "Unrecognized-model loading for a specific OLMo checkpoint is distinct from the generic sys.modules access bug." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "RT-DETR memory retention on delete is unrelated to MaskGenerationPipeline dropping final partial batch results." + "reason": "RT-DETR memory not being released is unrelated to MaskGenerationPipeline dropping the final partial batch." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Token embedding resize mismatch is unrelated to Whisper empty-transcription generation failure." + "reason": "resize_token_embeddings/output_embeddings mismatch is unrelated to Whisper empty transcription after align_special_tokens." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio processor chat-template sample-rate default is distinct from slow import-time performance." + "reason": "Audio sampling-rate defaults and import latency are unrelated problems." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Missing Qwen vision config field is not the same as AutoModelForCausalLM failing to unwrap text_config." + "reason": "Both involve Qwen 3.5/3.6 multimodal configs, but one is a missing strict field and the other is AutoModelForCausalLM.from_config text_config unwrapping; not the same bug." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is different from a ZeRO3 deque underflow." + "reason": "Both mention DeepSpeed/ZeRO, but one is Qwen3 MoE weight conversion and the other is an empty-deque IndexError. Different code paths and fixes." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "resize_token_embeddings propagation bug is unrelated to version-dependent tokenization changes." + "reason": "resize_token_embeddings behavior is unrelated to tokenizer-version-dependent tokenization changes." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Granite speech strict-config loading and Qwen3.5 MoE vision config dropping a field are separate model-config bugs." + "reason": "Both are config/strictness related, but they concern different models and different failure modes." }, { "left": "issue:43232", "right": "issue:44336", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to ANSI codes in loading_report." + "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI codes in loading_report." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "StaticCache sequence-length return type and LlamaConfig head_dim validation are different defects." + "reason": "StaticCache return type contract and LlamaConfig head_dim validation are separate issues." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "pytest-xdist file race and stale num_return_sequences warning are unrelated testing/runtime issues." + "reason": "A pytest-xdist file race is unrelated to a stale num_return_sequences warning." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model-load failure on transformers>=5.1.0 is not the same as tokenizer error masking when protobuf is missing." + "reason": "A model-loading regression for physical-intelligence/fast is unrelated to protobuf error masking in tokenizer decoding." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Video-input regression in v5.3.0 and Windows ROCm torch.distributed import issues are separate platform/model problems." + "reason": "A video-input regression in Qwen models is unrelated to Windows ROCm/torch.distributed import issues." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and sam3_video/sam3_tracker warning are related to SAM3 but not the same concrete bug." + "reason": "Both are Sam3-related, but a dummy Conv2d in the pixel decoder is not the same bug as a sam3_video vs sam3_tracker model-type warning." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "GlmMoeDsa stale indexer cache is unrelated to Gemma 4 requiring zeroed mm_token_type_ids." + "reason": "GlmMoeDsa stale cache crash and Gemma 4 mm_token_type_ids defaults are unrelated." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Missing Sam3VisionEncoderOutput field and Qwen3OmniMoe video-input ValueError affect different models and code paths." + "reason": "SAM3 VisionEncoderOutput missing a field and Qwen3OmniMoe video input validation are different failures." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Generic tied-weight serialization inconsistency and t5gemma-specific decoder.embed_tokens resize issue are not the same bug." + "reason": "Both touch embeddings/weight tying, but one is serialization inconsistency and the other is resize_token_embeddings not propagating to decoder.embed_tokens." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "gemma_speech zero3 loading and duplicate optimizer.step under deepspeed are different Deepspeed behaviors." + "reason": "Both mention DeepSpeed, but one is Gemma-4 zero3 loading and the other is double optimizer stepping. Different concrete problems." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "Chat/completions request rejection and Gemma pipeline roundtrip failure on \u2581 are unrelated API/tokenization problems." + "reason": "API request rejection and Gemma pipeline roundtrip failure on '\u2581' are unrelated." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "Zero3 deque underflow is not the same as auto_mappings referencing removed Sam3LiteText configs." + "reason": "An empty-deque ZeRO3 crash is unrelated to stale auto_mappings references to removed Sam3LiteText configs." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "tie_word_embeddings warning noise and flash-attention generation failure are separate Qwen3.5 issues." + "reason": "A tied-word-embeddings warning is unrelated to a flash-attention generation failure." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to the broader ESM2 breakage report." + "reason": "A tokenizer special-token registration bug is unrelated to the broad ESM2 breakage report." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "GlmMoeDsa cache corruption and flash_attention s_aux=None crash are different failures." + "reason": "A second-forward stale indexer cache crash is unrelated to a flash_attention.py s_aux=None AttributeError." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Qwen3.5 weight-tying warning and paged generate stale warning do not describe the same defect." + "reason": "The tied-weights warning and the stale num_return_sequences warning are different issues." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Granite speech strict config loading and ROCm/Windows distributed import errors are unrelated." + "reason": "Granite speech strict config loading and Windows ROCm import problems do not share a concrete bug." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight shape mismatch and Qwen3.5 tie_word_embeddings warning are different loading/config issues." + "reason": "A Qwen3-VL weight-shape mismatch is unrelated to tied-word-embedding warnings." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init policy change are unrelated model/config topics." + "reason": "Incorrect tied-weights warning in Qwen3-Next and patchtsmixer post_init policy are unrelated." } ] }, @@ -11523,8 +11664,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -11609,14 +11750,14 @@ "issue:45081|issue:45103", "issue:45310|issue:45375" ], - "prepared_review_unit_hash": "db6f01d26686a0cba96e01805ae15bb60c6ec60ac31629000bbdf8662a94818a", + "prepared_review_unit_hash": "b9806d58f2825e31e82e8e7664d57a4640c9acdb36cf816dff9051ee5bf23a42", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, @@ -11631,120 +11772,120 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers bug reports with similar wording around loading, config, generation, tokenizer, and vision/model-init regressions, but I do not see a true duplicate group here. Most soft edges are false positives from broad title similarity rather than the same underlying bug.", - "confidence": 0.38, - "canonical_issue_reason": "Issue 43646 is the broadest representative of the cluster\u2019s recurring theme of v5 model-initialization regressions, even though the cluster is not a tight duplicate set.", + "summary": "This cluster is heterogeneous: mostly Transformers v5-era regressions around model/config loading, tokenizers, generation, and save/load paths, but the soft pairs generally point to different concrete bugs rather than true duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43646 is the broadest, most central regression report in the set: it covers v5 model initialization breakage and can act as a representative umbrella for several adjacent load/config failures, even though the cluster is not a clean duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the best single issue to represent the cluster because it is the most general and central to the apparent v5 breakage theme; however, it should be treated as a representative, not as a confirmed duplicate hub.", + "best_issue_reason": "issue:43646 is the best single issue to anchor the cluster because it is the most general and broadly applicable bug report among otherwise distinct failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both involve generation/cache behavior, but they are different failures: missing GenerationMixin cache in chunked prefill vs wrong indexing in /v1/responses KV-cache continuation." + "reason": "Both are generation/cache-related, but one is chunked prefill cache handling and the other is KV-cache continuation indexing; different code paths and symptoms." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Tokenizer offline cache-loading failure and Gemma4 CUDA OOM are unrelated problems." + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are unrelated failure modes." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils and a pytest-xdist file-write race are unrelated." + "reason": "Unsafe sys.modules access in modeling_utils and an xdist race on captured_info.txt are different bugs." }, { "left": "issue:44530", "right": "issue:46170", "accept": false, - "reason": "PagedAttentionCache linear_attention crash and Windows ROCm torch.distributed import/type errors are different subsystems and bugs." + "reason": "PagedAttentionCache group-type crash on Qwen3.5 models is unrelated to Windows ROCm import/dependency issues." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Both concern save_pretrained APIs, but one is tokenizer metadata mismatch and the other is processor missing files; not the same concrete bug." + "reason": "Both involve save_pretrained, but tokenizer config mismatch and missing processor files are different save-time defects." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer task AttributeError and a data-filtering slowdown regression are unrelated." + "reason": "MLukeTokenizer task AttributeError and a Chinese performance regression are not the same bug." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper timestamp decoding crash and Kimi-K2.5 codec/fix_mistral_regex regression are different tokenizer issues." + "reason": "Both are tokenizer-related, but one is Whisper timestamp decoding and the other is Kimi-K2.5 codec/fix_mistral_regex behavior; not the same underlying issue." }, { "left": "issue:43452", "right": "issue:45003", "accept": false, - "reason": "gguf_file loading failures and unsafe sys.modules access do not look like the same code-path problem." + "reason": "gguf_file loading failures and unsafe sys.modules access do not share the same concrete code path." }, { - "left": "issue:43643", + "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "trust_remote_code missing fields in AutoConfig is not the same as the ESM2 breakage described here." + "reason": "AutoConfig trust_remote_code field loss and ESM2 breakage are different model families and failure mechanisms." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "processor.save_pretrained missing files and deepspeed optimizer double-step are unrelated." + "reason": "Processor save_pretrained missing files and deepspeed optimizer double-step are unrelated." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "A filtering-performance regression and an optimizer-step bug are unrelated." + "reason": "A throughput regression is not the same bug as an optimizer step being called twice." }, { "left": "issue:43673", "right": "issue:43827", "accept": false, - "reason": "Generation cache bug and docs still referencing removed pipeline() APIs are not the same issue." + "reason": "Generation cache corruption and stale docs referencing removed pipeline APIs are unrelated." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint key mismatch and data-processing slowdown are unrelated." + "reason": "Checkpoint resume key mismatch and a data-processing slowdown are different problems." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are different bugs." + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior changes affect different subsystems." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "torch.compile attention-kernel divergence and Gemma pipeline roundtrip failure are unrelated." + "reason": "torch.compile branching in attention kernels and Gemma pipeline roundtrip failure are distinct issues." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Issue 45230 is too generic to establish duplication with the Gemma-4 zero3 from_pretrained bug." + "reason": "Generic bug report provides no concrete match, and gemma-4 zero3 from_pretrained is a specific loading regression." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "Both are video/vision model issues, but one is a missing attribute in SAM 3 video and the other is wrong vision_position_ids in qwen2.5-vl; not the same bug." + "reason": "SAM 3 Video missing vision output fields and Qwen2.5-VL video position-id errors are different multimodal bugs." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Missing attribute in SAM 3 video and trust_remote_code missing fields are unrelated." + "reason": "SAM 3 Video output-field failure and trust_remote_code AutoConfig field loss are unrelated." }, { "left": "issue:42994", @@ -11756,67 +11897,67 @@ "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe from_pretrained error and processor.save_pretrained missing files are not the same underlying bug." + "reason": "Qwen3-VL-Moe v5 breakage and processor save_pretrained missing files are not the same bug." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "A missing PyTorch ImportError and a hang while loading Sentence Transformers are different failures." + "reason": "Missing PyTorch ImportError and Sentence Transformers loading hang are unrelated." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Gemma4 causal-mask construction and MaskGenerationPipeline final-batch handling are unrelated." + "reason": "Gemma4 causal-mask construction and partial-batch pipeline result dropping are different behaviors." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both are generation-related, but they target different code paths: sync_gpus kwargs update vs chunked_prefill cache handling." + "reason": "Generation kwargs update after sync_gpus and missing cache during chunked_prefill are separate generation bugs." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 configuration docs and remote-code meta-initialization compatibility are unrelated." + "reason": "TF32 settings API guidance and InternVL2 meta-init incompatibility are unrelated." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "SAM 3 video attribute error and add-new-model-like tokenizer-mapping failure are unrelated." + "reason": "SAM 3 Video output shape bug and add-new-model-like mapping failure are different model-registration problems." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import errors around image-processing APIs, but they refer to different missing symbols/modules and do not clearly describe the same bug." + "reason": "Both mention import issues, but one is missing AutoImageProcessor export and the other is a nonexistent image_utils import; not the same defect." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized saving failure and missing pad_token_id on Llama4Config are unrelated." + "reason": "Quantized model saving and missing pad_token_id on Llama4Config are unrelated." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5-Moe related, but one is a missing config field silently dropped by @strict and the other is a generic from_pretrained error; not enough to call them the same bug." + "reason": "Both are Qwen3.5 Moe-related, but one is a missing vision-config field dropped by strict serialization and the other is a generic from_pretrained error; too little evidence they are the same bug." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Mistral regex patch crash and auto_docstring annotation handling crash are unrelated." + "reason": "Mistral regex patch tokenizer crash and auto_docstring future-annotations crash are separate failures." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Both mention Qwen3.5-Moe, but one is a broad from_pretrained error and the other is a specific config-field omission; the evidence is too weak for a duplicate." + "reason": "Both concern Qwen3.5 Moe, but one is a generic from_pretrained error and the other is a missing deepstack_visual_indexes field in a vision config; not clearly the same underlying issue." } ] }, @@ -11835,8 +11976,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -11922,20 +12063,20 @@ "issue:43874|issue:45092", "issue:44734|issue:45103" ], - "prepared_review_unit_hash": "3c771669f0e423b1c8ed76135f1df1ce32bb5b8219ccbcbc081daf2d4d5e9531", + "prepared_review_unit_hash": "9f99593b328e4a7ca99cac58dbbe212dde7eeb7d99e7609bb5e69a92c15e5d50", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36963, + "serialized_chars": 36962, "estimated_input_tokens": 9241, "estimated_eval_tokens": 18738 }, @@ -11944,204 +12085,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated singleton issues. The only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report (42673/43704); the rest share broad themes at best, but not the same concrete bug or code path.", - "confidence": 0.93, - "canonical_issue_reason": "issue:42673 is the earlier, fuller report of the same Qwen3ForCausalLM multi-dataloader VRAM leak; issue:43704 is an exact-title duplicate.", + "summary": "The cluster is mostly a heterogeneous mix of unrelated Transformers bugs and regressions, with only one clear duplicate pair. Most soft edges should be rejected because they share only a broad subsystem or release/version theme, not the same concrete failure mode.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 42898 is the strongest anchor in this mixed set: it describes a concrete, reproducible v5 tokenizer regression with clear impact, while most other candidates are narrower model-specific breakages or unrelated failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the strongest canonical issue in this set because it has the clearest duplicate match and the most actionable description of the underlying bug.", + "best_issue_reason": "42898 is the best standalone issue here because it is a well-scoped framework-level regression with a specific symptom and broader user impact than the other, more fragmented reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Different Gemma-related failures: a server crash from missing `_tokenizer` versus a pipeline roundtrip issue on the underscore token." + "reason": "Both are Gemma-related, but one is a missing processor tokenizer attribute and the other is a token roundtrip/pipeline mismatch; different bugs and code paths." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different symbols (`AutoImageProcessor` vs `HybridCache`) and likely different fixes." + "reason": "Both are import errors, but they involve different symbols (`AutoImageProcessor` vs `HybridCache`) and different underlying breakages." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact same Qwen3ForCausalLM VRAM leak report, with identical title and matching symptom." + "reason": "Exact same reported bug: Qwen3ForCausalLM VRAM leak when used in multiple dataloader threads." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Model load failure for Phi-4 multimodal is unrelated to the DeepSpeed ZeRO3 deque underflow error." + "reason": "One is a model load failure for Phi-4 multimodal, the other is a DeepSpeed ZeRO3 deque error; unrelated failures." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized model saving failure is unrelated to an import error for a specific Qwen2.5-VL class." + "reason": "Quantized model saving failure versus a missing model import; different functionality and failure modes." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Both mention MoE/FP8, but one is a `save_pretrained` FP8 issue and the other is a `from_pretrained` error for Qwen3.5; different code paths." + "reason": "Both mention MoE/FP8-era models, but one is `save_pretrained` on FP8 MoE and the other is a `from_pretrained` load error for Qwen3.5 MoE; not the same bug." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "An accuracy regression in FineGrainedFP8 is not the same bug as FP8 MoE save/pretrained handling." + "reason": "Accuracy regression in FineGrainedFP8 versus serialization failure in FP8 MoE; different concrete problems." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Different missing-import regressions for different API symbols; same general symptom, not the same issue." + "reason": "Different missing imports (`GenerationMixin` vs `AutoImageProcessor`) and different parts of the API." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Different model families and failure modes: Qwen3-VL-Moe vs Qwen3.5 MoE." + "reason": "Both are Qwen MoE-related, but they describe different models/versions and different failure points." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/backend extra selection is unrelated to leftover dummy class cleanup in `dummy_pt_objects.py`." + "reason": "Packaging dependency omission is unrelated to leftover dummy classes and repo-check leakage." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Unrecognized model/config loading is unrelated to a pytest-xdist race in testing utilities." + "reason": "A model config/load compatibility issue is unrelated to a pytest-xdist file race." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "Missing padding-mask support in MIMI encoding is unrelated to router-logit/load-balancing-loss handling in MoE." + "reason": "MIMI padding-mask batched-vs-single mismatch is unrelated to MoE load-balancing loss being skipped when router logits are off." }, { "left": "issue:45561", "right": "issue:45600", "accept": false, - "reason": "A test-file race and a broken auto-mapping entry are different CI issues." + "reason": "A concurrent test-file race is not the same as stale `auto_mappings.py` references to removed configs." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwargs sync logic is unrelated to a Janus image-generation test failure." + "reason": "Generation kwargs/sync_gpus bug versus a Janus image-generation test failure; different code paths and symptoms." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is `clean_up_tokenization_spaces` behavior and the other is BOS/EOS insertion for a specific tokenizer." + "reason": "Both are tokenizer regressions in v5, but one is about `clean_up_tokenization_spaces` and the other about `add_special_tokens` not inserting BOS/EOS; not the same bug." }, { "left": "issue:43901", "right": "issue:45561", "accept": false, - "reason": "Docs mismatch for `TextClassificationPipeline` is unrelated to the xdist race on `captured_info.txt`." + "reason": "Documentation mismatch for `TextClassificationPipeline` is unrelated to the xdist captured-info race." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "Broken VitPose models and Gemma4 `mm_token_type_ids` behavior are different model-specific bugs." + "reason": "VitPose model breakage and Gemma 4 token-type-id handling are different model families and different failure mechanisms." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "A loading hang for Sentence Transformers is unrelated to embedding reinitialization during resize." + "reason": "A loading hang for Sentence Transformers is unrelated to embedding reinitialization after resizing token embeddings." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Embedding dtype drift in eval is unrelated to the ZeRO3 deque error." + "reason": "Training/eval dtype drift is unrelated to a DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing files in `processor.save_pretrained` is a different bug from Qwen3.5 MoE load errors." + "reason": "Missing files in `processor.save_pretrained(...)` is a persistence bug, not the same as a model load error." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Torch.compile attention-kernel divergence is unrelated to a docstring typo." + "reason": "Attention-kernel divergence under `torch.compile` is unrelated to a docstring typo for `position_ids`." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "Dummy Conv2d cleanup in Sam3PixelDecoder is unrelated to final-partial-batch handling in MaskGenerationPipeline." + "reason": "A dummy `nn.Conv2d` placeholder is unrelated to the final-partial-batch `is_last` handling bug." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed jobs is unrelated to an OOM with Adafactor/DeepSpeed." + "reason": "MLflow reporting failure after training job failure is unrelated to an Adafactor/DeepSpeed OOM." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation-state bookkeeping and attention-kernel branching are different concrete problems." + "reason": "Generation-state update bug and attention-kernel branching divergence are different code paths." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Different tokenizer regressions affecting different models and behaviors." + "reason": "Both are tokenizer regressions, but one is cleanup-space behavior and the other is Granite token-ID mapping; different underlying issues." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Both are import failures, but they concern different missing names and different integration points." + "reason": "A missing `nn` import during sentence-transformers import is unrelated to a Qwen2.5-VL import error." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading regressions and BOS/EOS special-token behavior are not the same underlying bug." + "reason": "Tokenizer loading in v5 and special-token insertion behavior in v5 are related only at a high level; they are distinct bugs." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Different tokenizer regressions: cleanup-space behavior versus Kimi-K2.5 codec handling / warning message." + "reason": "`clean_up_tokenization_spaces` regression is unrelated to Kimi tokenizer codec handling and warning behavior." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both involve tokenizer regressions, but for different models and root causes." + "reason": "A cleanup-space regression is not the same as Deepseek tokenizer producing incorrect results." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image-processor API and remote-code meta-init incompatibility are unrelated multimodal issues." + "reason": "Missing image-processor helper methods and remote-code/meta-init incompatibility are different multimodal loading issues." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache continuation crash in serve is unrelated to an auto-docstring annotation crash." + "reason": "KV-cache continuation tensor indexing in `transformers serve` is unrelated to the auto-docstring crash under future annotations." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet, and the soft-edge verdicts are appropriately conservative overall. The duplicate call on 42673/43704 is supported by the identical titles, earlier timestamp, and larger body on 42673. No mergeability issues apply since this is an issue-only set." + "feedback": "The clustering is well-grounded in the packet: the only clear duplicate pair is 42673/43704, and the other soft edges are rejected conservatively based on distinct titles, symptoms, or code paths. The choice of 42898 as the anchor issue is reasonable given the title-level evidence." }, "evaluator_used": true, "retried": false, @@ -12157,8 +12298,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -12245,232 +12386,232 @@ "issue:44222|issue:44484", "issue:44295|issue:44623" ], - "prepared_review_unit_hash": "84c4427cbd86ef2b5599bab5b9af56b77aeb45556d5491ac1b87412e726f5a6d", + "prepared_review_unit_hash": "b91d287695169573644d24a8895dc7d346dd06f6891444ca2b2f50f20345a9e1", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 36765, - "estimated_input_tokens": 9192, - "estimated_eval_tokens": 18640 + "serialized_chars": 36771, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated import errors, tokenizer/model bugs, trainer issues, config regressions, and docs problems. None of the soft edges look like the same underlying defect, so all candidate pairs should be rejected.", - "confidence": 0.92, - "canonical_issue_reason": "No single canonical issue; the items span unrelated bugs and regressions across different code paths.", + "summary": "This is a loose cluster of unrelated Transformers issues rather than a duplicate set. The soft-similarity pairs mostly share broad keywords like v5 regressions, imports, save_pretrained, or model loading, but they do not appear to describe the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.24, + "canonical_issue_reason": "No clear duplicate canon emerges; if one representative is needed, issue 43873 is the broadest active report and best cluster anchor, but it is still not a true duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44297 is one of the more concrete code-path bugs, but it still doesn't represent the rest of the mixed cluster well.", + "best_issue_reason": "Issue 43873 is the most general and still-open report in the set, so it is the least misleading representative, even though the surrounding issues are mostly distinct bugs or documentation problems.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error vs. a docs/parameter-name mismatch; different problems." + "reason": "Notebook syntax error vs documentation parameter-name mismatch; different artifacts and different failure modes." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Optimizer args being ignored is unrelated to a `set_training()` doc mismatch." + "reason": "Optimizer args being ignored is unrelated to a docs typo about set_training() parameters." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Position-ids buffer access bug vs. optimizer step being called twice; unrelated code paths." + "reason": "Position_ids buffer access error and double optimizer-step under deepspeed are separate code paths." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Tied-weights warning bug is unrelated to missing `_tp_plan` for tensor parallelism." + "reason": "Different Qwen model-specific bugs: tied-weights warning vs missing tensor-parallel plan." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Both mention v5 behavior, but one is tokenization cleanup and the other is model output drift for Pegasus; not the same bug." + "reason": "Both involve v5 behavior changes, but one is tokenization-space cleanup and the other is Pegasus output differences; not the same bug." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "PEFT version pinning issue vs. leftover dummy classes in `dummy_pt_objects.py`; unrelated." + "reason": "PEFT version pinning issue is unrelated to leftover dummy classes leaking into dir(transformers)." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Missing TensorFlow backend after `pip install .[torch]` is unrelated to a missing `image_utils` import." + "reason": "Backend/extra installation packaging issue differs from a missing image_utils import export." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Pegasus output change vs. `AttentionInterface.register` behavior change; different subsystems and failures." + "reason": "Model output regression and attention registration behavior change are distinct problems." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 `save_pretrained` bug is unrelated to tokenizer class metadata mismatch." + "reason": "FP8 save_pretrained bug and tokenizer_config class mismatch are unrelated save_pretrained issues." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Multimodal config default initialization bug vs. `ModelOutput` key assignment bug; unrelated." + "reason": "Phi4 multimodal config initialization is unrelated to ModelOutput key assignment." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Missing import from `image_utils` vs. dummy classes leaking into `dir(transformers)`; unrelated." + "reason": "Both touch import surface area, but one is a missing symbol and the other is dummy class leakage without torch." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer metadata mismatch vs. optimizer double-step; not the same defect." + "reason": "Tokenizer config mismatch and deepspeed double optimizer step have no shared underlying cause." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version compatibility problem vs. missing import; unrelated." + "reason": "Python version compatibility regression is not the same as a missing import from image_utils." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "FP8 MoE save bug vs. processor save_pretrained missing files; different save paths." + "reason": "FP8 MoE save_pretrained bug is different from missing files in processor.save_pretrained()." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading failure vs. LlamaConfig head-dim validation; unrelated." + "reason": "Local dataset loading and LlamaConfig head_dim validation are unrelated." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "MLflow reporting on failed jobs vs. Llama 3.2 Vision model loading regression; unrelated." + "reason": "MLflow reporting on failed training jobs and Llama 3.2 Vision model loading are different subsystems and failures." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FP8 accuracy issue vs. processor serialization missing files; different failures." + "reason": "FineGrainedFP8 accuracy issue is unrelated to processor save_pretrained missing files." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Config defaulting bug vs. Whisper timestamp decode crash; unrelated." + "reason": "Multimodal config default-init bug and Whisper timestamp decode crash are separate model-path issues." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLflow job-status bug vs. custom trainer signature change; unrelated." + "reason": "Training job reporting and custom trainer compute_loss signature mismatch are unrelated." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-count overflow vs. attention registration behavior change; unrelated." + "reason": "Category-limit runtime error and attention registration behavior change are not the same underlying bug." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "MLflow reporting bug vs. embedding resize/reinit bug; unrelated." + "reason": "MLflow job-state reporting and token embedding reinitialization are unrelated." }, { "left": "issue:28282", "right": "issue:30333", "accept": false, - "reason": "PyTorch missing import error vs. MLflow reporting bug; unrelated." + "reason": "AutoModel missing PyTorch import error is unrelated to failed-job MLflow reporting." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Type-checking regression vs. LlamaConfig head_dim validation; unrelated." + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are different config-layer problems." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different modules; both packaging-related, but not the same issue." + "reason": "Different missing imports from different packages; both are import errors but not the same symbol or fix." }, { "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 and Qwen3.5 MoE from_pretrained regression are unrelated." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Feature-extractor center-arg regression vs. auto-docstring crash with future annotations; unrelated." + "reason": "LastFeatureExtractor center-arg regression and auto_docstring annotation crash are different failure modes." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading regression vs. Qwen3.5 flash-attention generation bug; unrelated." + "reason": "Tokenizer loading issue in v5 and Qwen3.5 flash-attention generation bug do not share a concrete root cause." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenization change across versions vs. FP8 inference failure on B200; unrelated." + "reason": "Tokenization drift across versions and FP8 inference failure on B200 are unrelated." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "`torch.split` return-value handling bug vs. remote-code/meta-init incompatibility; unrelated." + "reason": "torch.split return-value handling in one model is unrelated to remote-code checkpoint meta-init incompatibility." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs update bug vs. Kosmos2.5 long-OCR index error; unrelated." + "reason": "Generation kwarg update after sync_gpus and Kosmos2.5 OCR index error are different code paths." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization/offloading bug vs. pipeline-parallelism output bug; unrelated code paths." + "reason": "Quantization offloading issue and incorrect pipeline-parallel output are not the same bug." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "FP8 MoE save bug vs. `max_shard_size` question; not the same defect." + "reason": "FP8 save_pretrained MoE bug and max_shard_size question are unrelated save_pretrained topics." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Position-ids buffer read bug vs. processor save_pretrained missing files; unrelated." + "reason": "Registering position_ids as a buffer and processor missing files are separate serialization/model-state issues." } ] }, @@ -12489,8 +12630,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -12573,14 +12714,14 @@ "issue:42371|issue:44263", "issue:43646|issue:46002" ], - "prepared_review_unit_hash": "61b79087ceaf101823b122aafe9de34be888f7258bd9de5befb12ef35306192d", + "prepared_review_unit_hash": "c0121a48f777bee04abcbb2e4441ce17dd59e1a495f8b333d35ea863550960c0", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, @@ -12595,180 +12736,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: mostly separate Transformers v5 regressions around imports, tokenizer/config behavior, quantization, and a few model-specific bugs. None of the soft pairs look like true duplicates. I chose the broad v5 import-regression report as the best issue anchor.", - "confidence": 0.66, - "canonical_issue_reason": "issue:45278 is the broadest umbrella item here: it describes many import errors after upgrading to 5.5.0 and is closest to the repeated import/load-failure reports in the set.", + "summary": "This set is highly heterogeneous: most items are unrelated import/API regressions, tokenizer/model-specific bugs, or saving/loading issues. I did not find any soft pair that looks like the same concrete bug/change, so all soft edges should be rejected. If forced to anchor the cluster, 42886 is the cleanest narrow regression report to use as a representative issue.", + "confidence": 0.92, + "canonical_issue_reason": "issue:42886 is the most concrete, self-contained regression report with a narrow repro (tokenizer cache loading under HF_HUB_OFFLINE=1) and clear scope; the other issues are mostly unrelated bug classes.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most representative issue to anchor this cluster because it captures a recurring class of upgrade/import breakages rather than a single model-specific symptom.", + "best_issue_reason": "issue:42886 is the most triage-friendly representative because it describes one specific failure mode and environment, making it a better anchor than the broader or more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Docstring mismatch vs GLM-5 RoPE implementation bug; different problem and code path." + "reason": "Docstring typo vs GLM-5 RoPE behavior; different bug classes and code paths." }, { "left": "issue:43761", "right": "issue:46082", "accept": false, - "reason": "Hidden-states regression in CLIPVisionModel vs LlamaConfig validation on head_dim; unrelated." + "reason": "CLIP hidden_states regression vs LlamaConfig validation; unrelated functionality." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Multimodal config defaulting bug vs non-persistent buffer initialization issue; not the same bug." + "reason": "Multimodal config initialization vs non-persistent buffer junk; not the same underlying defect." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "Chat-completions request rejection vs Gemma4Processor missing _tokenizer; distinct failures." + "reason": "Chat/completions request rejection vs Gemma4Processor missing _tokenizer; different processor/server failures." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Meta-initialization incompatibility for old InternVL2 checkpoints vs Qwen3.5 save_pretrained key bug; different symptoms." + "reason": "InternVL2 meta-init incompatibility vs Qwen3.5 save_pretrained visual key bug; same broad area, different concrete failures." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer load failure in v5 vs eos_token_id config inconsistency; related area but not same issue." + "reason": "Tokenizer loading regression vs eos_token_id mismatch; related to tokenization but not the same bug." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Both concern DSA indexer internals, but one is torch.split output handling and the other is a missing ReLU; different bugs." + "reason": "Two different DSA/indexer issues with different symptoms (split outputs vs missing ReLU)." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading stall vs zero-probability output on SIGLIP2; unrelated." + "reason": "Loading stall vs zero-probability output; different model/runtime failures." }, { "left": "issue:44514", "right": "issue:45103", "accept": false, - "reason": "Batched apply_chat_template crash vs auto-docstring crash with future annotations; different code paths." + "reason": "Qwen2_5_VL batched template crash vs auto_docstring annotation crash; unrelated subsystems." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Precision/error from missing rms_norm_eps vs remote-code meta-init incompatibility; not the same underlying bug." + "reason": "Layernorm epsilon precision issue vs meta initialization incompatibility; not the same code-path problem." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both involve quantized save/load, but one is dequantized Ministrals/Devstrals save failure and the other is sinq round-trip breakage; not identical." + "reason": "Dequantized model save failure vs sinq quantization save/load breakage; same broad area, different quantization flow." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Flash-attention-3 training degeneration vs NaNs after checkout+init; different failure modes and likely different causes." + "reason": "Flash-attention training degeneration vs NaNs during checkout/init; unrelated training/init bugs." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Config default initialization bug vs rotary_pct not persisting on reload; separate config regressions." + "reason": "Default multimodal config init vs GPTNeoX rotary_pct reload behavior; different config bugs." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "Generic FineGrainedFP8 accuracy issue vs Qwen3.5 from_pretrained error; different scope and symptom." + "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 Moe from_pretrained error; different model and failure mode." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing transformers.onnx module vs missing PyTorch dependency for AutoModel; both import errors but unrelated." + "reason": "Missing transformers.onnx module vs missing PyTorch dependency; both import errors but not the same root cause." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Cannot import AutoImageProcessor vs cannot import Qwen2_5_VLForConditionalGeneration; different missing symbols." + "reason": "Two separate missing imports for different symbols; no evidence they are the same underlying change." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "DSA indexer runtime issue vs wrong docstring for position_ids; one is functional, the other documentation." + "reason": "Indexer tensor-splitting bug vs docstring mistake; clearly unrelated." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "Accuracy issue associated with FineGrainedFP8 vs Qwen3-VL-Moe bug; different models and failures." + "reason": "FP8 accuracy problem vs Qwen3-VL-Moe bug; different model behavior and code path." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Qwen3 tokenizer offset_mapping bug vs mDeBERTa add_special_tokens/BOS-EOS regression; different tokenizer behaviors." + "reason": "Tokenizer offset mapping bug vs BOS/EOS special-token regression; both tokenizer-related but distinct." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline parallelism produces wrong output vs infinite loop on a specific image; separate model/runtime bugs." + "reason": "Pipeline parallelism output error vs image-specific Qwen3_5 infinite loop; different symptoms and fixes." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure vs wrong tokenizer decoder type in v5; related tokenizer area but not same bug." + "reason": "Offline cache load failure vs wrong decoder type in v5; separate tokenizer/config issues." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "Specific missing is_offline_mode import vs broad many-import-errors upgrade report; overlapping theme only." + "reason": "Single missing hub symbol vs broad upgrade-related import failures; 45278 is an aggregate report, not the same concrete bug." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Tokenizer cache/offline failure vs quantized model save failure; different subsystems." + "reason": "Offline tokenizer cache load bug vs quantized model saving failure; different operations and root causes." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None assumption vs AutoProcessor kwargs not forwarded; separate from_pretrained/caching issues." + "reason": "Tokenizer mapping None handling vs AutoProcessor kwargs forwarding; different from_pretrained internals." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "Broken transformers env command vs NameError importing sentence-transformers with latest transformers; different entry points." + "reason": "Broken env command vs sentence-transformers import NameError; unrelated CLI vs import regression." }, { "left": "issue:43550", "right": "issue:46132", "accept": false, - "reason": "torch.compile+SDPA failure for Bamba-9B-v2 vs AttentionInterface.register behavior change; unrelated code paths." + "reason": "torch.compile SDPA bug vs AttentionInterface.register behavior change; different APIs and code paths." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 settings API guidance issue vs DSA indexer torch.split bug; not the same change or bug." + "reason": "TF32 settings warning/behavior vs DSA indexer tensor split issue; unrelated." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model init regression in Transformers 5.0 vs NaNs in classification heads on init; different manifestations and causes." + "reason": "Custom model initialization regression vs NaNs after checkout/init; different failure mechanisms and scope." } ] }, @@ -12787,8 +12928,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -12876,232 +13017,232 @@ "issue:43784|issue:45446", "issue:22355|issue:28218" ], - "prepared_review_unit_hash": "21345a071d6165be02f22439bb5406047b2dcb22dd360a472cbd9a5cab0f460b", + "prepared_review_unit_hash": "16f7fe824ce783ac7981a43f34b5e530f6046d5168300e5eecab44e19ae5a36c", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37186, - "estimated_input_tokens": 9297, - "estimated_eval_tokens": 18850 + "serialized_chars": 37193, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a collection of unrelated Transformer issues across tokenization, import/loading, multimodal processing, generation, and model-saving paths. I do not see a safe canonical duplicate representative, and none of the soft pairs look like the same underlying bug/change.", - "confidence": 0.91, - "canonical_issue_reason": "No single issue is a safe canonical representative; the items span unrelated bugs and regressions rather than one duplicate thread.", + "summary": "This cluster is a loose grab-bag of Transformers bug reports, mostly around imports, tokenizers, model loading, and generation. The soft pairs look superficially similar by keyword, but they do not describe the same concrete underlying bug.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical issue emerges: the items are mostly distinct bugs with only superficial overlap in subsystem or wording.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue: none of these issues cleanly subsumes the rest or represents the whole cluster.", + "best_issue_reason": "Issue 43723 is the broadest tokenizer-loading regression and is the closest thing to a thematic anchor, but it still does not represent the rest as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Both are bugs, but one is about reading a buffered position_ids tensor and the other about tokenizer_class metadata on save_pretrained; different code paths." + "reason": "Different bugs: position_ids buffer access vs tokenizer_class metadata mismatch in save_pretrained." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Sentence Transformers loading hang vs a MacOS bus error in a CLIP model; different symptoms and likely different root causes." + "reason": "Unrelated failures: Sentence Transformers loading hang vs MacOS bus error on a CLIP model." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation kwargs after sync_gpus is unrelated to D-FINE auxiliary loss behavior." + "reason": "Generation kwargs/sync_gpus bug is unrelated to D-FINE auxiliary loss behavior." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Passing kwargs to cached_file in AutoProcessor is a loading/path issue, while apply_chat_template with tool calls is a templating/runtime crash." + "reason": "Different code paths: AutoProcessor kwargs forwarding vs chat template crashing on tool-call messages." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "Precision drift from missing rms_norm_eps is unrelated to a docstring correction for position_ids." + "reason": "MLA layernorm epsilon precision issue is unrelated to a docstring problem for position_ids." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and incorrect pipeline-parallel outputs are different problems." + "reason": "Different problems: dequantized save_pretrained failure vs incorrect output under pipeline parallelism." }, { "left": "issue:43723", "right": "issue:44442", "accept": false, - "reason": "A generic AutoTokenizer loading regression is not the same as one specific tokenizer class failing to load." + "reason": "Both involve AutoTokenizer loading, but one is a v5 generic issue and the other is a specific FastSpeech2Conformer tokenizer failure." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Missing nn during import of sentence-transformers is unrelated to stale auto_mappings references to removed Sam3LiteText configs." + "reason": "Importing sentence-transformers with missing nn is unrelated to stale auto_mappings references breaking CI." }, { "left": "issue:43723", "right": "issue:44779", "accept": false, - "reason": "Tokenizer loading failure in v5 is not the same as Deepseek producing incorrect tokenization results." + "reason": "Generic tokenizer loading regression is not the same as Deepseek tokenizer producing wrong results." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE-layer loss and Qwen3-Next tied-weights warnings are different model-specific bugs." + "reason": "Different model-specific logic: Smollm3 RoPE layer count vs Qwen3-Next tied-weights warning." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "A single missing AutoImageProcessor import does not match a broad set of import errors after version upgrade." + "reason": "A specific missing AutoImageProcessor import is not the same as broad import errors after version upgrade." }, { "left": "issue:43122", "right": "issue:44514", "accept": false, - "reason": "Tokenization differences across versions and a batched apply_chat_template crash are separate paths." + "reason": "Tokenizer behavior change across versions is unrelated to Qwen2_5_VLProcessor batched chat-template crashes." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Model save/pretrained dequantization is unrelated to an invalid GenerationConfig error." + "reason": "Model saving failure is unrelated to GLM-5 invalid GenerationConfig." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "A model-type warning for sam3_tracker is not the same as a missing fpn_position_embeddings attribute in SAM 3 Video." + "reason": "These are different SAM 3 video issues: model-type warning vs missing fpn_position_embeddings attribute." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Empty multi-label classification output in an example script is not the same as missing mm_token_type_ids defaults for Gemma 4 fine-tuning." + "reason": "Multi-label classification example output bug is unrelated to Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Version-dependent tokenization changes are different from assistant_masks being all zero for multimodal chat templates." + "reason": "Tokenizer version drift is not the same as assistant_masks being all zero for multimodal chat templates." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "A dequantized save_pretrained bug is unrelated to loading a specific multimodal Phi-4 model." + "reason": "Dequantized save_pretrained failure and Phi-4 multimodal load failure are separate model-loading/saving bugs." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "A missing ReLU in the DSA indexer is unrelated to GLM-5 RoPE implementation concerns." + "reason": "DSA indexer ReLU discussion is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe transformer behavior and tokenizer_config metadata mismatch are different issues." + "reason": "Qwen3-VL-Moe bug is not the same as tokenizer_class metadata mismatch on save_pretrained." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation for Qwen3.5 and flash_attention_3 training degeneration do not share a concrete code path." + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 degenerate training are unrelated." }, { "left": "issue:43232", "right": "issue:45245", "accept": false, - "reason": "Generation-state handling after sync_gpus is unrelated to a category-cardinality limit error." + "reason": "Generation kwargs/sync_gpus behavior is unrelated to the 2^24 categories runtime limit." }, { "left": "issue:44038", - "right": "issue:44222", + "right": "issue:44295", "accept": false, - "reason": "A Qwen3-VL-Moe bug and FP8 save_pretrained for MoE are both MoE-adjacent but not the same concrete failure." + "reason": "Different issues: Qwen3-VL-Moe behavior vs reading position_ids after registering a buffer." }, { "left": "issue:28218", "right": "issue:30990", "accept": false, - "reason": "An extra space added after an added token is unrelated to a loading hang in Sentence Transformers." + "reason": "Tokenizer adds a space after an added token is unrelated to Sentence Transformers loading getting stuck." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Dequantized save_pretrained and gguf_file incompatibilities affect different APIs and likely different fixes." + "reason": "Dequantized save_pretrained failure is unrelated to gguf_file breaking AutoTokenizer/AutoModel loading." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Different tokenization across versions and Deepseek-specific incorrect results are not the same bug." + "reason": "Version-to-version tokenization drift is not the same as Deepseek tokenizer producing incorrect results." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor-parallel training bugs and multi-label example-script output are unrelated." + "reason": "Tensor-parallel training bug is unrelated to the multi-label classification example script returning empty results." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Qwen3-VL-Moe runtime behavior and FP8 save_pretrained MoE serialization are distinct problems." + "reason": "Qwen3-VL-Moe bug and FP8 save_pretrained moe are distinct MoE-related problems, not one concrete fix." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Wrong offset_mapping for Qwen3 tokenizer is unrelated to failing to load cjvt/sleng-bert." + "reason": "Qwen3 tokenizer offset_mapping mismatch is unrelated to failing to load cjvt/sleng-bert." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "A bad save_pretrained checkpoint and a from_pretrained error for Qwen3.5 MoE are related by model family but not the same concrete bug." + "reason": "Both mention Qwen3.5, but one is save_pretrained checkpoint regression and the other is from_pretrained load error." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Missing Qwen2_5_VLForConditionalGeneration and missing HybridCache are separate import failures." + "reason": "Different missing-symbol import errors: Qwen2_5_VLForConditionalGeneration vs HybridCache." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Tokenizer offset mapping and multimodal assistant mask generation are different code paths." + "reason": "Tokenizer offset mapping bug is unrelated to assistant_masks being all zero for multimodal inputs." }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "An import NameError in sentence-transformers is unrelated to a PyTorch version check bug in flex_attention." + "reason": "Sentence-transformers import NameError and flex_attention version-check bug are unrelated." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "A missing transformers.onnx module and an added-token spacing regression have no shared underlying cause." + "reason": "Missing transformers.onnx module import and tokenizer added-space behavior are unrelated." } ] }, @@ -13120,8 +13261,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -13205,20 +13346,20 @@ "issue:42175|issue:44938", "issue:43827|issue:45446" ], - "prepared_review_unit_hash": "e06622f2bb21146714b22bf45dcc60c346220a856f4155e22664e92debc076f0", + "prepared_review_unit_hash": "eda97432bc9648aaee3dfb393bf3a30d67d6981b095fb84aa2dd5d8ce101a2e8", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35907, + "serialized_chars": 35908, "estimated_input_tokens": 8977, "estimated_eval_tokens": 18210 }, @@ -13227,192 +13368,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are separate Transformers v5 regressions or model-specific bugs, not duplicates. The closest thing to a broad umbrella is the upgrade-related import-error report.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45278 is the broadest representative: it describes multiple import failures after upgrading Transformers, which best matches the cluster\u2019s overall upgrade-regression flavor.", + "summary": "The cluster is mostly heterogeneous: tokenizer regressions, import/API breakages, model-specific bugs, and unrelated docs/tests. I don\u2019t see any soft pair that clearly describes the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43122 is the broadest tokenizer-regression report in the set and best matches the recurring v4\u2192v5 behavior-change theme; the other issues are either narrower model-specific bugs or unrelated API/docs/import failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best single issue to anchor the cluster because it is user-facing, broad, and can absorb several import-regression reports without overcommitting to a narrow code path.", + "best_issue_reason": "Issue 43122 is the best umbrella issue for this cluster because it is the most general and central tokenizer behavior regression, whereas the rest are mostly distinct one-off failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols in different packages; not the same underlying bug." + "reason": "Both are import errors, but they involve different missing symbols from different packages (`huggingface_hub` vs `transformers`) and do not indicate the same bug." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "One is a missing module import, the other is a tokenizer behavior inconsistency; unrelated symptoms." + "reason": "One is a missing module import for `transformers.onnx`; the other is a tokenizer consistency complaint. Different codepaths and failures." }, { "left": "issue:43122", "right": "issue:43906", "accept": false, - "reason": "The second is only an isolated reproduction reference, and the visible titles do not establish the same concrete bug." + "reason": "Both mention tokenizer behavior, but one is a version-regression report and the other is an isolated reproduction of a different issue. Not the same concrete bug." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "A compiled-file loading error and a documentation parameter typo are not the same issue." + "reason": "Compiled Python file errors and a documentation parameter-name mismatch are unrelated." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "These concern different code paths: model weight loading vs. AutoProcessor kwargs forwarding." + "reason": "`UMT5Encoder` missing weights and `AutoProcessor` dropping kwargs are distinct loader/processor bugs." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different failure modes." + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes (decoder type vs incorrect results)." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Model loading with base_model_prefix is unrelated to NaNs in classification heads after init/checkpoint." + "reason": "A `base_model_prefix` loading regression and NaNs in classification heads are unrelated issues." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "Same model family, but one is GenerationConfig validation and the other is RoPE implementation; different bugs." + "reason": "These are separate GLM-5 problems: invalid generation config versus RoPE implementation behavior." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility and a PyTorch version guard are different compatibility problems." + "reason": "Python-version support failure and an incorrect PyTorch version check are different compatibility bugs." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF backend packaging and a missing image-patch helper are unrelated." + "reason": "TF32 API guidance and a missing image-patch helper in GLM46V are unrelated." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Different model-specific breakages: vitpose models vs. Ministral-3 FlashAttention failure." + "reason": "Different models and different failures: vitpose breakage versus FlashAttention support for Ministral-3." }, { "left": "issue:42831", "right": "issue:44297", "accept": false, - "reason": "FP8 accuracy and tokenizer save metadata mismatch are not the same defect." + "reason": "FineGrainedFP8 accuracy issues are unrelated to tokenizer save/load metadata mismatches." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Optimizer argument handling and trainer evaluate() failing after train() are distinct training bugs." + "reason": "Optimizer argument handling and `trainer.evaluate()` failing after training are separate trainer issues." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both involve Qwen3.5, but one is save_pretrained key serialization and the other is from_pretrained loading; not the same concrete bug." + "reason": "Both mention Qwen3.5, but one is a save/load key regression and the other is a `from_pretrained` error. Not the same bug." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Related model family, but the visible reports describe different load-time failures and do not show one shared code path." + "reason": "These are different Qwen/MoE model-loading problems with no clear shared codepath." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "An example-script classification issue and a pipeline() error message/docs mismatch are separate problems." + "reason": "A multi-label classification example bug and a misleading pipeline error message are unrelated." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "Different model families and different symptom types: implementation discrepancy vs. infinite loop on a specific image." + "reason": "SigLIP2 implementation drift and a Qwen3_5 infinite loop are different model-specific bugs." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Multi-label example behavior and SigLIP2 implementation discrepancy are unrelated." + "reason": "Classification example behavior and SigLIP2 implementation discrepancy do not share the same underlying failure." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "A helper-function behavioral bug and a version-check import bug are not the same underlying issue." + "reason": "A Qwen3_omni_moe helper-function issue and a flex_attention import check are unrelated." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "Both concern attention behavior, but they are different models and different masking/scaling logic." + "reason": "Both involve attention behavior, but they concern different models and different mechanics (scaling vs mask construction)." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "GenerationConfig validation for GLM-5 and a stale warning in paged generate() are different bugs." + "reason": "GenerationConfig validation and a stale warning about `num_return_sequences` are not the same issue." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Missing transformers.onnx module and unclear LayoutLMv3 error messaging are unrelated." + "reason": "Missing module import and an unclear LayoutLMv3 error message are unrelated." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping fallback logic and flash_attention.py handling of s_aux=None are different code paths." + "reason": "Tokenizer mapping assumptions and a flash-attention `s_aux=None` crash affect different subsystems." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Wrong offset mapping in Qwen3 tokenizer and Deepseek tokenization regression are distinct model-specific issues." + "reason": "Both are tokenizer regressions, but one is an offset-mapping bug in Qwen3 and the other is a broader incorrect-results report for Deepseek v5." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression and wrong offset mapping are related only broadly, not as the same concrete bug." + "reason": "Tokenizer loading regression and wrong offset mapping are distinct tokenizer bugs with different symptoms." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import-related, but one is a specific missing symbol while the other is a broad post-upgrade import-failure report; not enough evidence they are duplicates." + "reason": "A specific missing import and a broad set of import errors after upgrade are related only at the theme level, not the same bug." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Num_labels propagation in Qwen3.5 is unrelated to NaNs in classification heads." + "reason": "`num_labels` propagation and NaNs in classification heads are separate classification-related problems." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Backend packaging under pip install .[torch] and a Python 3.14 load failure are different environment issues." + "reason": "Torch-only install backend coverage and Python 3.14 loading failures are different compatibility/install issues." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() and a PyTorch version check bug are unrelated." - }, - { - "left": "issue:44038", - "right": "issue:45753", - "accept": false, - "reason": "Both mention Qwen3-related models, but one is a load/regression report and the other is an infinite-loop case on a specific image; not the same bug." + "reason": "Docs still referencing `pipeline()` and an incorrect PyTorch version check are unrelated changes." } ] }, @@ -13431,8 +13566,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -13513,190 +13648,190 @@ "issue:42371|issue:44261", "issue:43012|issue:43577" ], - "prepared_review_unit_hash": "8d06732e35095b31bf0d3802629ba2bc1983fb9c562b4622deea9965d6b8d118", + "prepared_review_unit_hash": "78b9a1e0f7a004ece0facdf09d4f1f1fa74118370f94adf9b8bedc0ad7d33420", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 35556, - "estimated_input_tokens": 8889, - "estimated_eval_tokens": 18034 + "serialized_chars": 35557, + "estimated_input_tokens": 8890, + "estimated_eval_tokens": 18036 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of standalone Transformers issues, with only a few loose thematic similarities (tokenization, dtype, v5 regressions) but no clear duplicate cluster. If a representative issue is needed, the broad v5 import/regression report is the best umbrella candidate.", - "confidence": 0.76, - "canonical_issue_reason": "Issue 45278 is the broadest and most representative regression report in the set: it summarizes multiple import failures after upgrading to v5.5.0, whereas most other items are narrower, model-specific, or unrelated bugs.", + "summary": "The cluster is highly heterogeneous: most pairs are only loosely related by theme (tokenizers, loading, training, video/multimodal, backend/dependency issues) and do not describe the same underlying bug or change.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report here (\u201cMany import errors after update from 4.57.0 to 5.5.0\u201d), so it is the best representative if one must be chosen, though the set is not a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 best captures the general upgrade/regression theme and would be the most useful anchor for triage, even though the cluster is not a true duplicate set.", + "best_issue_reason": "issue:45278 has the widest scope and is the most plausible sink for related version-regression reports, but it still does not cover the other concrete failures in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Both involve training/generation internals, but one is an optimizer-state bug under FSDP2+PEFT and the other is a cache omission in GenerationMixin; different code paths and fixes." + "reason": "Different bugs: optimizer-state corruption under FSDP2/PEFT vs missing generation cache during chunked prefill." }, { "left": "issue:43381", "right": "issue:44945", "accept": false, - "reason": "Eval-mode gradient checkpointing and pipeline-parallel output corruption are unrelated behaviors in different execution paths." + "reason": "Eval-mode gradient checkpointing limitation is unrelated to pipeline-parallel output corruption." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "One is an incorrect batched video-processor output shape, the other is a crash caused by missing torchvision during AutoProcessor loading." + "reason": "Batched video shape bug vs crash from missing torchvision dependency are different failure modes." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Both are generation failures, but one is a GenerationConfig validation issue and the other is a flash-attention-specific Qwen3.5 generation bug." + "reason": "Invalid GenerationConfig for GLM-5 is not the same as Qwen3.5 flash-attention generation breakage." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "Serialization of tied lm_head weights and NaNs after checkout/init are distinct defects with different symptoms and likely different fixes." + "reason": "Serialization of tied weights depends on device; NaNs after checkout/init is a separate model-init issue." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A dependency version bump issue and a PyTorch version-check bug in flex_attention are unrelated." + "reason": "Dependency version bump problem is unrelated to the flex_attention PyTorch version check bug." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leaking in Qwen3ForCausalLM threads and dtype mismatches in SwitchTransformers/TimmWrapperModel are not the same bug." + "reason": "VRAM leak in multithreaded Qwen3.5 inference is distinct from bfloat16 dtype mismatches." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "A tokenizer mapping fallback bug and a Mistral regex patch crash in tokenizer loading are different tokenizer subsystems." + "reason": "Tokenizer mapping None is a different issue from the Mistral regex patch crashing on backend_tokenizer access." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "A broken vitpose model import and a backend-mapping error for tensorflow_text are separate backend/model-loading issues." + "reason": "Broken vitpose models and missing tensorflow_text backend registration are unrelated." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing layer projections in qwen3next and a Qwen3.5 MoE from_pretrained error are different model-loading problems." + "reason": "Missing qwen3next projections is a different model-architecture bug from Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "FlashAttention-3 training degeneration and incorrect Qwen3.5 checkpoint saving are unrelated failures." + "reason": "flash_attention_3 training degradation is unrelated to the save_pretrained checkpoint regression." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Wrong offset_mapping in Qwen3 tokenizer and runtime special-token/post-processor update behavior are different tokenization bugs." + "reason": "Wrong offset_mapping and stale post-processor updates are separate tokenizer issues." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "These are both v5 strictness regressions, but one is an init_empty_weights argument error and the other is PretrainedConfig type-checking; not the same concrete code path." + "reason": "init_empty_weights argument handling differs from PretrainedConfig type-checking regression." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Import errors for different symbols ('layer_type_validation' vs 'AutoImageProcessor') are separate missing-export problems." + "reason": "Both are import errors, but they involve different missing symbols and code paths." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A model-type mismatch warning and a key error while loading a specific model are not the same defect." + "reason": "Model-type warning for sam3 is not the same as the MobileLLM key-error load failure." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading and a documentation parameter-name mismatch are unrelated." + "reason": "Strict config loading and a docs parameter-name mismatch are unrelated." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "A question about max_shard_size and a Qwen3.5 MoE from_pretrained failure do not share a code path." + "reason": "These both touch chat/completions, but one is a generic endpoint rejection and the other is a Qwen3.5 MoE model load error." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Invalid GenerationConfig and unrecognized model/config.json are different loading errors." + "reason": "Invalid GenerationConfig and unrecognized model/config.json are different loading failures." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "They are both tokenizer regressions, but the first is specifically about decoder type in v5 while the second is a broader tokenization-output change; not enough to treat as the same bug." + "reason": "Wrong tokenizer decoder type and version-to-version tokenization changes are related only broadly, not the same bug." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "Broken vitpose models and a TensorFlow/tensorboard backend install problem are distinct backend-related issues." + "reason": "Vitpose model breakage is unrelated to the undefined TensorFlow backend caused by tensorboard-only installation." }, { "left": "issue:44964", "right": "issue:45278", "accept": false, - "reason": "Loading Phi-4 multimodal instruct and a broad 'many import errors after upgrade' report are related only by release timing, not the same underlying bug." + "reason": "Loading a specific multimodal model is not the same as a broad import-error regression after upgrading transformers." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Both mention chat/completions, but one is a general API rejection and the other is a model-specific streaming inference failure." + "reason": "Both mention chat/completions, but the concrete failures differ: endpoint validation vs streaming inference on Qwen3.5-0.8B." }, { "left": "issue:42175", "right": "issue:42222", "accept": false, - "reason": "Missing TensorFlow in the torch extra and broken vitpose models are unrelated dependency/model issues." + "reason": "Missing TensorFlow in the torch extra is a packaging/backend issue, not the vitpose model breakage." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "Optimizer args ignored and optimizer step called twice under deepspeed are both optimizer-related, but they describe different defects." + "reason": "SGD args being ignored is a different optimizer bug from Deepspeed calling step twice." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 API guidance and MLA q_a_layernorm epsilon precision drift are unrelated." + "reason": "TF32 behavior settings and missing rms_norm_eps precision error are unrelated." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "A PyTorch precision warning during bfloat16 compilation and incorrect dtype persistence when loading BLIP-2 are not the same bug." + "reason": "PyTorch precision warning during compile is not the same as incorrect dtype propagation in Blip2 loading." } ] }, @@ -13715,8 +13850,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -13806,244 +13941,244 @@ "issue:43931|issue:44410", "issue:43502|issue:43992" ], - "prepared_review_unit_hash": "8401b3fcaad7dfb1206b362bd7b44c64a2873c4c0f665bdf50d01aa7637c17f4", + "prepared_review_unit_hash": "3f3bcf1d9ccb496ed817481eb1112286a4d52961fbe7ac15fba77cb4490c6fa1", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38082, - "estimated_input_tokens": 9521, - "estimated_eval_tokens": 19298 + "serialized_chars": 38090, + "estimated_input_tokens": 9523, + "estimated_eval_tokens": 19302 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated Transformers issues; none of the suggested soft pairs look like true duplicates or the same fix, so all soft edges should be rejected.", - "confidence": 0.58, - "canonical_issue_reason": "issue:43097 is a clear, user-facing regression tied to a specific v5 API change, with enough discussion/inbound attention to serve as a representative issue in this mixed set.", + "summary": "The cluster is heterogeneous: the soft-similarity pairs mostly reflect broad subsystem or symptom overlap, not the same underlying bug/change. I would not merge any of the proposed soft edges, and there is no clear canonical issue/PR for this set.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated failures across imports, model config, attention, tokenizer, and training paths.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the most concrete and broadly relevant regression report here, making it the strongest single issue to anchor triage despite the lack of a coherent duplicate cluster.", + "best_issue_reason": "No best issue for duplicate consolidation: none of the issues appears to subsume the others as the same concrete bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3-family failures, but one is a missing projection in qwen3next while the other is a Qwen3.5 MoE from_pretrained error; different code paths and symptoms." + "reason": "Both involve Qwen-family models, but one is a missing projection in qwen3next and the other is a Qwen3.5 MoE loading error; different code paths." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "These are different import/functionality bugs in unrelated areas: a missing configuration_utils symbol versus an audio feature extraction issue." + "reason": "Different failures: a missing transformers export vs an audio feature extraction issue." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "One is about a removed config flag in v5, the other about unexpected position-id keys; they do not describe the same underlying defect." + "reason": "One is a removed config API warning; the other is an unexpected position-id key during model loading." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different multimodal/model internals are failing: image patch counting versus MLA q_a_layernorm precision/config behavior." + "reason": "Unrelated model components: image processor method missing vs a layer-norm epsilon precision bug." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "An import error for AutoImageProcessor is not the same as a PretrainedConfig type-checking regression." + "reason": "Import of AutoImageProcessor is unrelated to PretrainedConfig type-checking breakage." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "These target different GLM bugs; one is torch.split handling in a DSA indexer, the other is RoPE implementation for GLM-5." + "reason": "Both touch GLM-related code, but the reported defects are different operators/modules and not the same bug." }, { "left": "issue:42673", "right": "issue:44977", "accept": false, - "reason": "A VRAM leak in multi-threaded dataloading is a different bug from flash-attention generation failures in Qwen3.5." + "reason": "VRAM leak in dataloader threads is unrelated to flash-attention generation failure." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Type-checking breakage in PretrainedConfig and a TensorFlow backend import/runtime issue are unrelated." + "reason": "Type-checking regression and TensorBoard/TF backend error are unrelated." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "The isolated reproduction of another issue is not enough to match the Qwen3ForCausalLM VRAM leak; the underlying failures are different." + "reason": "The second is an isolated reproduction of a separate issue, not the same VRAM leak." }, { "left": "issue:44263", "right": "issue:46172", "accept": false, - "reason": "A torch.split indexer issue and a DeepSeek tensor transpose bug are unrelated code-path problems." + "reason": "Different model families and different failure modes; no shared concrete bug." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "These concern different model implementations and different faults; no shared fix is plausible." + "reason": "DSA ReLU concern vs a DeepSeek-V4 dtype/transpose issue; unrelated." }, { "left": "issue:43012", "right": "issue:43408", "accept": false, - "reason": "Both mention warnings, but one is about bfloat16/PyTorch precision warnings and the other is a model-type mismatch warning." + "reason": "A PyTorch precision warning while compiling bfloat16 is not the same as a sam3 model-type warning." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "save_pretrained sharding defaults have nothing to do with a Qwen3.5 MoE loading failure." + "reason": "Save-pretrained shard size question is unrelated to Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both involve flash attention, but one is a Ministral-3 loading failure and the other is degenerate training with flash_attention_3; different bugs." + "reason": "Both mention flash attention, but they describe different concrete bugs: inference failure vs degenerate training." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 API settings and FineGrainedFP8 accuracy issues are both numerical, but they are not the same bug or fix." + "reason": "TF32 API guidance and FineGrainedFP8 accuracy regression are distinct topics." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "A SwitchTransformers sparse-layer config bug is unrelated to a PretrainedConfig typing regression." + "reason": "Sparse-layer construction bug and type-checking regression are unrelated." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Both are Qwen3-VL-related, but one is a weight-shape mismatch and the other is a Qwen3-VL-Moe compatibility bug." + "reason": "Both involve Qwen3-VL variants, but one is a weight-shape mismatch and the other is a broader v5/MoE loading issue." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "A removed config flag and a gguf_file loading breakage are different failure modes and code paths." + "reason": "A removed embedding-tie config warning does not match a gguf_file loading breakage." }, { "left": "issue:44933", "right": "issue:46231", "accept": false, - "reason": "A missing image_utils export is unrelated to a documentation mismatch in set_training()." + "reason": "Import error and documentation parameter-name mismatch are not the same bug." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Both are loading-related, but one is a tokenizer mapping None handling bug and the other is a specific model load failure for physical-intelligence/fast." + "reason": "TOKENIZER_MAPPING_NAMES returning None is a different failure from loading physical-intelligence/fast on >=5.1.0." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "These are separate GLM bugs affecting different components; not the same underlying issue." + "reason": "Missing image-processor method and torch.split return-value issue are unrelated." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "A missing torchvision dependency crash during AutoProcessor loading is unrelated to shard-size defaults." + "reason": "Missing torchvision crash in AutoProcessor is unrelated to max_shard_size behavior." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "These warnings stem from different model/config mismatches and do not share a concrete fix." + "reason": "These are both warnings, but they concern different models and different config mismatches." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version incompatibility and incorrect pipeline-parallel output are unrelated defects." + "reason": "Python version incompatibility and incorrect pipeline-parallel output are different problems." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "Different DeepSeek models and different bugs; one is implementation divergence, the other is a tensor operation issue." + "reason": "Different DeepSeek variants and different implementation details; not the same concrete defect." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "A v5 config removal is not the same as sinq quantization breaking across save/load." + "reason": "Embedding-tie removal warning is unrelated to quantized save/load corruption." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "The tie_embeddings regression is unrelated to a flash_attention integration crash on s_aux=None." + "reason": "Config warning about removed API is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "A wrong optimizer-state issue in Trainer/FSDP2/PEFT is not the same as Liger Kernel not being applied with model_init." + "reason": "Distributed optimizer-state corruption in FSDP2/PEFT is unrelated to Liger Kernel application with model_init." }, { "left": "issue:42673", "right": "issue:45200", "accept": false, - "reason": "VRAM leakage in dataloader threads and Gemma 4 mm_token_type_ids defaults are distinct bugs." + "reason": "VRAM leak in multithreaded loading and Gemma 4 mm_token_type_ids defaulting are unrelated." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "A tokenizer constructor argument collision is unrelated to TOKENIZER_MAPPING_NAMES returning None." + "reason": "Tokenizer AddedToken keyword collision is different from TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different symbols from different libraries; they are not the same missing export." + "reason": "Both are import errors, but for different missing symbols from different packages and code paths." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "These are both Qwen3-related, but one is a missing layer projection and the other is a config field being dropped by strict parsing." + "reason": "Missing projections in qwen3next is unrelated to a strict config field being dropped." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Different Qwen model families and different failures; no shared concrete code-path problem." + "reason": "One is a Qwen3-VL-Moe bug and the other is qwen3next layer-projection omission; not the same issue." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "A model weight mismatch and missing layer projections are not the same bug." + "reason": "Weight mismatch on Qwen3-VL-30B is a different failure mode from missing attn projections in qwen3next." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "local_files_only network leakage and a missing embed_tokens.weight are unrelated from_pretrained issues." + "reason": "The first is unintended network access with local_files_only, while the second is a missing embed_tokens weight on UMT5Encoder." } ] }, @@ -14062,8 +14197,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -14150,226 +14285,226 @@ "issue:45698|issue:46172", "issue:44038|issue:44230" ], - "prepared_review_unit_hash": "2a6852092a97bd43e988fe41f214422ce2ae60ad60ab173afb7cc0c55d51fe88", + "prepared_review_unit_hash": "83118e4eea221276b9093dff609782c7110b7a12886a8ddb27319078cd49d741", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36966, - "estimated_input_tokens": 9242, - "estimated_eval_tokens": 18740 + "serialized_chars": 36960, + "estimated_input_tokens": 9240, + "estimated_eval_tokens": 18736 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are unrelated model-specific bugs, API regressions, import failures, or test utility issues. I did not find any soft edge that looks like the same underlying defect.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue stands out because the issues do not share one concrete bug or code path.", + "summary": "This cluster is heterogeneous: import errors, processor/chat-template bugs, save/load regressions, compiler warnings, and model-specific failures. None of the soft pairs look like the same underlying bug or change, so no duplicates should be merged here. No PRs are present.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No best issue for deduplication: none of these items is clearly representative of the others.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45341", "right": "issue:45859", "accept": false, - "reason": "Different subsystems and failures: a testing_utils bug vs a CPU-offload rotary embedding incompatibility." + "reason": "Different subsystems: a testing_utils bug versus a Qwen3.5 RoPE CPU-offload issue." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and likely different fixes." + "reason": "Both are import errors, but they involve different missing symbols in different modules (`layer_type_validation` vs `HybridCache`)." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "A removed config flag regression and a quantization/offloading issue are not the same bug." + "reason": "One is a removed API warning/regression, the other is an offloading/quantization behavior issue; not the same bug." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Precision/epsilon configuration bug vs RoPE implementation issue; different code paths." + "reason": "MLA q_a_layernorm epsilon handling and GLM-5 RoPE implementation are unrelated code paths." }, { "left": "issue:44964", "right": "issue:45341", "accept": false, - "reason": "Model loading failure for Phi-4 multimodal is unrelated to a testing_utils bug." + "reason": "Model loading regression versus a testing_utils bug; no shared failure mode." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor kwargs forwarding and auto_docstring annotation handling are different areas and failures." + "reason": "AutoProcessor kwargs forwarding and auto_docstring annotation parsing are separate problems." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Missing torchvision in AutoProcessor video loading vs tokenizer mapping None in from_pretrained; different root causes." + "reason": "Different failures: missing torchvision for video model loading versus TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "Unrelated issues: 3d_parallel.py runtime failure vs a small testing_utils bug." + "reason": "A 3d_parallel.py runtime failure is unrelated to the testing_utils bug." }, { "left": "issue:43986", "right": "issue:45290", "accept": false, - "reason": "Video-model loading crash and chat-template tool-call crash are distinct code paths." + "reason": "Processor loading without torchvision and chat-template tool-call handling are different code paths." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same concrete bug." + "reason": "Wrong tokenizer decoder type in v5 and a reproduction of another issue are not the same concrete bug." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Missing torchvision crash is unrelated to the mistral_common ReasoningEffort import error." + "reason": "Video-model loading without torchvision is unrelated to the ReasoningEffort import crash in Gemma 4 loading." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs and Bamba torch.compile SDPA failure are different model-specific problems." + "reason": "xLSTM training bugs and a Bamba torch.compile/SDPA failure are distinct model-specific issues." }, { "left": "issue:45341", "right": "issue:45588", "accept": false, - "reason": "Testing_utils bug vs flash_attention s_aux=None crash; different modules and symptoms." + "reason": "Testing-utils bug versus flash_attention s_aux=None crash; different code paths and symptoms." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "A generic testing_utils bug is not the same as the xdist captured_info.txt race." + "reason": "Testing-utils bug versus pytest-xdist race on captured_info.txt; not the same issue." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens.weight on load vs save_pretrained shard-size question; different APIs and behavior." + "reason": "UMT5 from_pretrained missing weights and save_pretrained max_shard_size default are unrelated." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "AutoProcessor/video-model loading bug is unrelated to testing_utils." + "reason": "Video-model loading failure is unrelated to the testing_utils bug." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "PyTorch precision warning on bf16 compile is unrelated to a compile failure in Bamba SDPA." + "reason": "A precision warning during compile is not the same as a Bamba SDPA compile failure." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline task messaging and saving Pipeline objects are different issues." + "reason": "Pipeline task-support messaging and Pipeline save_pretrained failure are different defects." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Both are warnings, but one is PyTorch precision-related and the other is a tie_word_embeddings config warning." + "reason": "These are different warnings with different triggers: bf16 compile precision versus tie_word_embeddings in LoRA fine-tuning." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Saving dequantized models and saving Pipeline objects are not the same code-path bug." + "reason": "Saving dequantized models and saving Pipeline objects are different save_pretrained behaviors." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Dequantized save failure vs an infinite-loop image bug; unrelated." + "reason": "Model save failure and Qwen3.5 infinite-loop inference bug are unrelated." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and deepseek-v4 transpose/dtype bug are unrelated model failures." + "reason": "ESM2 breakage and deepseek-v4 tensor-op failure are separate model-specific bugs." }, { "left": "issue:43792", "right": "issue:46123", "accept": false, - "reason": "Whisper model runtime failure vs MaskGenerationPipeline final-batch handling; different behaviors." + "reason": "Whisper not running and MaskGenerationPipeline dropping final partial-batch results are different subsystems." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import/load failure is unrelated to the tensorflow_text backend mapping error." + "reason": "Python 3.14 load failure and a tensorflow_text backend mapping error have different causes." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs and BeitImageProcessorFast reduce_labels behavior are unrelated." + "reason": "xLSTM training bugs do not match the BeitImageProcessorFast reduce_labels bug." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "Chat/completions API regression and batched apply_chat_template crash are different problems." + "reason": "Both involve chat/completions behavior, but one is API acceptance and the other is batched apply_chat_template crashing; not the same fix." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained issue and Qwen3.5 MoE from_pretrained error are related only by model family, not the same bug." + "reason": "FP8 save_pretrained moe and Qwen3.5 moe from_pretrained errors are different load/save paths." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "A warning about bf16 precision is not the same as a torch>=2.9 compiled flex_attention failure." + "reason": "Both mention compilation, but one is a bf16 warning and the other is a flex_attention failure on torch>=2.9." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "Trailing-slash local model import warning and bf16 precision warning are unrelated." + "reason": "A trailing-slash local model import warning is unrelated to a bf16 compile warning." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Different missing-import failures: is_offline_mode from huggingface_hub vs Qwen2_5_VLForConditionalGeneration from transformers." + "reason": "Both are import errors, but they concern different missing names and different import surfaces." }, { "left": "issue:45698", "right": "issue:46172", "accept": false, - "reason": "Wrong custom module loading after save_pretrained is unrelated to the deepseek-v4 tensor operation bug." + "reason": "Wrong custom module after save_pretrained and a deepseek-v4 tensor operation bug are not the same underlying issue." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "A transformers 5.0 Qwen3-VL-Moe bug and fp8 inference support request are not the same concrete defect." + "reason": "Both mention Qwen/MoE/FP8, but one is a generic transformers 5.0 bug and the other is FP8 inference support; not one concrete duplicate." } ] }, @@ -14388,8 +14523,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -14479,250 +14614,250 @@ "issue:36683|issue:44117", "issue:43618|issue:44380" ], - "prepared_review_unit_hash": "58629337240049e1f657d8836b2b308214e53cd548ac34cc23a4082c82d36322", + "prepared_review_unit_hash": "8d3b813d737a6a08dede34cf795a7aff10880d55875552143f80453892b9427e", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 36, - "serialized_chars": 37674, - "estimated_input_tokens": 9419, - "estimated_eval_tokens": 19094 + "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 a heterogeneous mix of unrelated Transformers v5 regression reports and API breakages. None of the soft-edge pairs looks like the same underlying bug or change, so all proposed merges should be rejected.", - "confidence": 0.94, - "canonical_issue_reason": "No single issue cleanly serves as a canonical duplicate target; the items span unrelated models, APIs, and failure modes.", + "summary": "This cluster is mostly a grab-bag of distinct Transformers v5 regression reports (imports, config fields, model loading, attention backends, save/load paths, and a few model-specific bugs). The soft pairs are thematically similar in places, but none look like the same concrete underlying bug/change.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43611 is the broadest, clearest regression-framed report in the set and a reasonable anchor for the model-loading breakage theme, but the cluster is too heterogeneous for a strong duplicate canon.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43502 is the strongest standalone report: it has a clear user-facing repro and likely actionable fix scope, but it does not subsume the rest of the cluster.", + "best_issue_reason": "issue:43611 is the most representative single issue to keep as a triage anchor because it is concise, actionable, and closest to the common v5 loader-regression surface area.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "ImportError for a missing internal helper vs unexpected hub API calls under local_files_only; different modules and failure modes." + "reason": "Both are import-related, but the missing symbol (`layer_type_validation`) and the local-files-only network leak are different APIs and different failure modes." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "GenerationMixin cache/chunked prefill bug vs final partial batch bookkeeping in MaskGenerationPipeline; unrelated code paths." + "reason": "Both involve generation/pipeline behavior, but one is a missing cache during chunked prefill and the other is final partial-batch result dropping in MaskGenerationPipeline." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Chat/completions API request validation vs tokenizer chat template crash on tool-call messages; different layers and symptoms." + "reason": "Chat/completions request handling vs `apply_chat_template` crash on tool-call assistant messages are different code paths and bugs." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "CLIPOutput attentions field regression vs CLIPVisionModel hidden_states regression; both CLIP-related but not the same bug." + "reason": "Both are CLIP regressions, but one is missing `attentions` and the other is `hidden_states=None`; different outputs and likely different fixes." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2/PEFT optimizer-state issue vs GPT2 attention scaling ignored in SDPA/FlashAttention; unrelated training vs inference behavior." + "reason": "Trainer/FSDP/PEFT optimizer-state corruption is unrelated to GPT-2 attention scaling being ignored under SDPA/FlashAttention." }, { "left": "issue:43066", "right": "issue:46209", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 vs FP8 inference failure on B200; no shared underlying code path." + "reason": "Wrong tokenizer decoder type and FP8 inference failure on B200 are unrelated issues." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL MoE model loading break vs state_dict sorting load error; both loading-related but distinct causes." + "reason": "Both are loading failures, but one is Qwen3VL MoE model loading and the other is a generic sorted-state_dict load error; not the same bug." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Gemma3Config missing vocab_size vs SAM3 vision output missing fpn_position_embeddings; different models and missing attributes." + "reason": "Missing `vocab_size` in Gemma3Config vs missing `fpn_position_embeddings` in Sam3VisionEncoderOutput are distinct model-specific attribute errors." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Token embedding resize propagation bug vs bidirectional attention mask construction bug; different Gemma4 issues." + "reason": "Both are Gemma4-related, but resize-token-embeddings propagation and bidirectional-attention mask construction are different behaviors." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure vs Pipeline.save_pretrained serialization failure; different save paths and object types." + "reason": "Quantized model saving and pipeline object serialization both involve saving, but they are different failure paths with different objects and APIs." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Documentation mismatch for return_all_scores vs runtime pipeline save_pretrained failure; not the same bug." + "reason": "Documentation mismatch for `return_all_scores` is unrelated to pipeline `save_pretrained` failures." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Missing GenerationMixin import vs missing is_offline_mode import; separate import regressions in different packages." + "reason": "Separate import errors from different packages/symbols (`GenerationMixin` vs `is_offline_mode`)." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Qwen3.5 num_labels propagation vs Gemma4 attention-mask bug; unrelated model configs." + "reason": "Qwen3.5 `num_labels` propagation and Gemma4 attention-mask behavior are unrelated config/model bugs." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "base_model_prefix loading regression vs pipeline serialization regression; both save/load-adjacent but different code paths." + "reason": "Model loading with `base_model_prefix` and pipeline save serialization are different code paths and fixes." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "FlashAttention 2 unsupported for NemotronHForCausalLM vs GLM-5 GenerationConfig validation error; different models and problems." + "reason": "Nemotron FlashAttention 2 support gating and GLM-5 `GenerationConfig` validation are unrelated." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "State_dict ordering load failure vs incorrect Qwen3.5 checkpoint save; load-order bug and save-pretrained regression are distinct." + "reason": "Generic load error with sorted state_dict and Qwen3.5 checkpoint corruption from `save_pretrained` are distinct save/load bugs." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "Gemma4 attention-mask regression vs LlamaConfig head_dim validation; unrelated configuration issues." + "reason": "Gemma4 causal-mask bug and LlamaConfig `head_dim` validation are unrelated config/attention issues." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Local-files-only network leak vs GLM-5 RoPE implementation discussion; no shared bug." + "reason": "Local-files-only still making API requests is unrelated to GLM-5 RoPE implementation discussion." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Gemma3Config missing vocab_size vs Llama4Config missing pad_token_id; separate model config omissions." + "reason": "Both are missing-config-attribute reports, but they concern different models and different fields (`vocab_size` vs `pad_token_id`)." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions field regression vs Qwen3.5 checkpoint save regression; different models and failure modes." + "reason": "CLIP attention output regression and Qwen3.5 save_pretrained corruption are unrelated." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "GPT2 backend attention scaling vs AttentionInterface.register behavior change; different attention subsystems." + "reason": "Attention scaling in GPT-2 backends and `AttentionInterface.register` behavior changes are different problems." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "Missing Gemma3Config attribute vs MoE load-balancing loss omission; unrelated model/config bugs." + "reason": "Missing `vocab_size` and load-balancing loss omission are different subsystems and failure modes." }, { "left": "issue:36683", "right": "issue:43881", "accept": false, - "reason": "Gemma3Config missing vocab_size vs GLM-4V-9B loading failure; different models and symptoms." + "reason": "A missing config attribute in Gemma3 is not the same as a GLM-4V model loading failure." }, { "left": "issue:43906", "right": "issue:45753", "accept": false, - "reason": "Reproduction of an external issue vs Qwen3_5 infinite loop on a specific image; not the same problem." + "reason": "An isolated reproduction of an earlier issue and a Qwen3_5 infinite-loop-on-image report are not the same bug." }, { "left": "issue:43299", "right": "issue:45753", "accept": false, - "reason": "Qwen3VL MoE loading failure vs Qwen3_5 image-generation loop; distinct model families and bugs." + "reason": "Qwen3VL MoE loading failure and Qwen3_5 image infinite loop are different model/runtime issues." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "FlashAttention support error vs Python version compatibility break; unrelated issues." + "reason": "FlashAttention support issue and Python-version compatibility issue are unrelated." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "UMT5Encoder missing embed_tokens weight vs Qwen3.5 MoE from_pretrained error; different models and load failures." + "reason": "Missing `embed_tokens.weight` in UMT5Encoder loading is not the same as Qwen3.5 MoE `from_pretrained` failure." }, { "left": "issue:43441", "right": "issue:44380", "accept": false, - "reason": "Model-specific FlashAttention2 unsupported error vs GPT2 attention scaling config ignored; different attention backend bugs." + "reason": "Both mention attention/FlashAttention, but one is unsupported FA2 for Ministral-3 and the other is ignored GPT-2 attention scaling; different concrete bugs." }, { "left": "issue:43352", "right": "issue:45478", "accept": false, - "reason": "Nemotron FlashAttention2 incompatibility vs duplicate-titled Qwen3.5 MoE from_pretrained regression; unrelated." + "reason": "These are different issues entirely: Nemotron FlashAttention support vs Qwen3.5 MoE `from_pretrained` error." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "State_dict-sorting load error vs Gemma pipeline roundtrip failure on \u2581; different serialization/loading paths." + "reason": "Sorted-state_dict load failure and Gemma pipeline roundtrip on the underscore token are unrelated save/load behaviors." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "General compiled-file import errors vs Python 3.13 IndentationError from @torch.jit.script parsing; not the same root cause." + "reason": "Compiled Python-file errors and Python 3.13 `@torch.jit.script` parsing/indentation errors are different import-time problems." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention backend config ignored vs zero-loss classification with num_labels=1; unrelated inference vs training bug." + "reason": "Ignored attention scaling and degenerate zero loss with `num_labels=1` are unrelated model-training bugs." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "FlashAttention2 unsupported model vs GPT2 attention scaling ignored; same broad area only, not one fix." + "reason": "FlashAttention support gating is unrelated to GPT-2 scaling being ignored in other attention backends." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "State_dict ordering load failure vs missing deepstack_visual_indexes in Qwen3.5 vision config; different bugs." + "reason": "Load failure on sorted state_dict and missing `deepstack_visual_indexes` in Qwen3_5MoeVisionConfig are different issues." }, { "left": "issue:36683", "right": "issue:44117", "accept": false, - "reason": "Missing config attribute vs tokenizer mapping sometimes returning None; unrelated subsystems." + "reason": "Missing config field in Gemma3 and tokenizer-mapping returning `None` are unrelated failure modes." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output regression vs GPT2 attention scaling backend issue; both model-internal but distinct." + "reason": "CLIP attentions missing and GPT-2 attention scaling ignored are different regressions affecting different models and code paths." } ] }, @@ -14741,8 +14876,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -14829,20 +14964,20 @@ "issue:43723|issue:44556", "issue:43502|issue:44222" ], - "prepared_review_unit_hash": "3b3d8e077dbd25f9a3a2803a570cb7e83b593e6d8346a0fa2da65738f7a8e3a8", + "prepared_review_unit_hash": "7a1ebaf494bfc02a1f5fabd23bcca71e73106e8a8b0c5b8505535d5fe8f368b6", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36755, + "serialized_chars": 36754, "estimated_input_tokens": 9189, "estimated_eval_tokens": 18634 }, @@ -14851,217 +14986,211 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated product bugs, regressions, or docs/test issues. The only strong duplicate-like pair is the import-time slowdown reports; the rest are too semantically distant to merge confidently.", - "confidence": 0.71, - "canonical_issue_reason": "issue:41669 is the clearest root-cause report for the import slowdown, with a concrete symptom and likely fix path; issue:44246 reads like a broader duplicate symptom report.", + "summary": "This is not a duplicate cluster; the issues span unrelated import, loading, training, CI, docs, and model-specific regressions. None of the soft pairs looks like the same underlying bug/change.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the set is heterogeneous and does not form one duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the best representative issue because it is specific, actionable, and describes the underlying performance regression rather than just the user-facing slowdown.", + "best_issue_reason": "No strong best issue: there isn't a single report that plausibly subsumes the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41669", "right": "issue:44246", - "accept": true, - "reason": "Both describe transformers import being unexpectedly slow; 41669 gives the likely cause (import * in models) and 44246 reports the same user-visible slowdown." + "accept": false, + "reason": "Both mention import slowdown, but one is a concrete import-star cleanup proposal and the other is an intermittent performance complaint; not the same bug." }, { "left": "issue:43097", "right": "issue:43352", "accept": false, - "reason": "Different problems: one is about a removed tie_embeddings API indication, the other about Flash Attention support in a specific model." + "reason": "Different problems: removed config flag behavior vs Flash Attention support for NemotronH." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "One is flaky model integration tests from missing set_seed, the other is a pytest-xdist race in patched testing utils; different code paths." + "reason": "Both are test/CI failures, but one is missing set_seed flakiness and the other is an xdist file race." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "FSDP/PEFT optimizer-state corruption and a cache-strategy typo are unrelated." + "reason": "Unrelated: distributed optimizer state corruption vs a typo in cache strategy docs/code." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both mention Python version incompatibility, but the failing versions and underlying issues are different." + "reason": "Both are Python-version compatibility complaints, but they target different releases and failure modes." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "One is a load failure with sorted state_dicts; the other is incorrect save_pretrained keys for Qwen3.5 visual encoder state." + "reason": "Both involve loading/saving state, but one is a sorted-state_dict load error and the other is incorrect visual encoder keys on save_pretrained." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Different model families and different from_pretrained failures; too little evidence they share the same code path." + "reason": "Different model-specific load bugs; no clear shared code path." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel bugs are only broadly similar; one is a generic training bug and the other is an OOM when loading a specific large model." + "reason": "Both mention tensor parallelism, but one is a training bug and the other is an OOM on model load." }, { "left": "issue:43064", "right": "issue:44655", "accept": false, - "reason": "Optimizer-state corruption during training and inability to save Pipeline objects are unrelated." + "reason": "Optimizer-state mismatch in training is unrelated to Pipeline save_pretrained support." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "An ImportError for missing PyTorch is unrelated to a tokenizer loading regression." + "reason": "AutoModel missing PyTorch and tokenizer loading regression are unrelated components." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "CLIP attentions assignment and an AMP/CUDA loss crash are different failures." + "reason": "CLIPOutput attentions regression and an AMP/CUDA crash in LwDetrImageLoss are different bugs." }, { "left": "issue:44655", "right": "issue:44938", "accept": false, - "reason": "Saving Pipeline objects and Python 3.14 load failures are unrelated." + "reason": "Pipeline serialization and Python 3.14 import/load failures are unrelated." }, { "left": "issue:43618", "right": "issue:44977", "accept": false, - "reason": "CLIP attentions regression and Qwen3.5 flash-attention generation issues are different model paths." + "reason": "CLIPOutput attentions assignment is unrelated to Qwen3.5 flash-attention generation behavior." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Different features and failure modes; no shared underlying bug is evident." + "reason": "No shared bug: CLIPOutput output fields vs patchtsmixer post_init allowance." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "A missing ReLU in the DSA indexer is not the same as DeepSeekV3MoE divergence from the remote implementation." + "reason": "Both discuss model internals, but the DSA indexer ReLU issue and DeepSeekV3MoE divergence are not the same change." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "A flaky SAM-HQ test seed issue is unrelated to a model loading failure for Phi-4 multimodal." + "reason": "Test seeding flakiness is unrelated to a model loading failure." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Offline download behavior and AddedToken argument handling are separate bugs." + "reason": "Local-files-only network leak and AddedToken keyword collision are different bugs." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Test flakiness causes differ: missing seeds vs CI device override race." + "reason": "Both are CI/test issues, but the flaky seed problem and GraniteSpeech device-override failure are distinct." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "A test seeding issue and stale auto_mappings references are unrelated." + "reason": "Test flakiness and auto_mappings breaking CI due to removed configs are not the same underlying problem." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained behavior and missing tensor-parallel plan support are different code paths." + "reason": "Both concern Qwen3.5/MoE, but save_pretrained FP8 behavior and missing _tp_plan for tensor parallelism are different code paths." }, { "left": "issue:42890", "right": "issue:45341", "accept": false, - "reason": "A specific model test seed flake is not the same as a generic testing_utils bug." + "reason": "A generic testing-utils bug and SamHQ set_seed flakiness are not clearly the same defect." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "The transformers env CLI and AttentionInterface.register behavior are unrelated." + "reason": "Broken env command and changed AttentionInterface.register behavior are unrelated API issues." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols and likely different export regressions." + "reason": "Both are import errors, but they involve different missing symbols and likely different export paths." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and a slowdown after upgrading versions are not the same issue." + "reason": "Compiled-Python-file errors and post-upgrade filtering slowdown are unrelated." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Same model family, but one is a missing _tp_plan detail and the other is a broader from_pretrained error; not enough evidence for one fix." + "reason": "Same model family, but missing _tp_plan and generic from_pretrained error are not enough to treat as one concrete bug." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow reporting on failed training jobs and tokenizer loading regression are unrelated." + "reason": "MLflow job reporting on failure and tokenizer loading regression are unrelated." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 quantization memory regression and checkpoint reload failures are different problems." + "reason": "Int4 quantization memory growth and checkpoint reload compatibility are different regressions." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "One is tensor-parallel support, the other is a config field dropped by strict parsing; distinct bugs." + "reason": "A missing tensor-parallel plan and a dropped vision-config field are separate config/schema issues." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "A no_grad decoration request on cache updates is unrelated to import errors after upgrade." + "reason": "Cache update no_grad semantics and broad import errors after upgrade are not the same bug." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading stalls and forward-argument documentation confusion are unrelated." + "reason": "Sentence Transformers hanging on load and confusing forward-arg docs are unrelated." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Tokenizer loading regression and checkpoint reload incompatibility are related in theme but not clearly the same underlying bug." - }, - { - "left": "issue:44387", - "right": "issue:44556", - "accept": false, - "reason": "Both involve v5 regressions, but one is memory usage under int4 and the other is reload compatibility; different fixes." + "reason": "AutoTokenizer.from_pretrained failure and v4-to-v5 checkpoint reload failure are different loading problems." }, { - "left": "issue:44222", - "right": "issue:45375", + "left": "issue:43502", + "right": "issue:44222", "accept": false, - "reason": "FP8 save_pretrained issues and missing vision config fields are separate model serialization problems." + "reason": "Network calls despite local_files_only and FP8 save_pretrained on MoE are unrelated." } ] }, "evaluator_result": null, - "evaluator_used": true, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -15073,8 +15202,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -15165,250 +15294,250 @@ "issue:43525|issue:45850", "issue:44393|issue:45753" ], - "prepared_review_unit_hash": "5c468d85e1f85216a18a0b3bf73d2d532c1bfedf7a2ca3cc46263c4169f1ae24", + "prepared_review_unit_hash": "512ad6c1090d0d50a0c4038ee9c747f6d7f0ec2b85a75814d4ce233e20a3ecf3", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 37801, - "estimated_input_tokens": 9451, - "estimated_eval_tokens": 19158 + "serialized_chars": 37808, + "estimated_input_tokens": 9452, + "estimated_eval_tokens": 19160 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers issues around v5 regressions, import/loading failures, model-specific bugs, and docs/utility breakages. The soft-similarity links look superficial and should not be treated as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45850 is the broadest umbrella report in the set ('Since 5.0 version it breaks too many models'), so it is the closest cluster anchor, though it is not a precise duplicate target for the more specific bugs.", + "summary": "Mostly heterogeneous transformer v5 regressions, import/load failures, and docs-only issues; the soft links look like broad topical similarity rather than true duplicates. No PRs are present.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43576 is the cleanest concrete regression report in the set and the closest thing to an umbrella user-facing failure, though it still does not truly subsume the rest of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the best representative issue only because it captures the general v5 breakage theme; the rest are narrower, unrelated root causes and should stay separate.", + "best_issue_reason": "issue:43576 is the best single representative because it is specific and actionable; the other issues are either model-specific, documentation-only, or unrelated runtime regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "Reject: FP8 accuracy issues and a missing vision-processor method are different model-path bugs with different fixes." + "reason": "Different problems: FP8 accuracy vs a missing image-processor method/AttributeError." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Reject: docs referencing removed pipeline APIs is unrelated to a placeholder issue with no evidence of the same bug." + "reason": "Docs about pipeline removals are unrelated to the unrelated 'racoon' report." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Reject: local_files_only still triggering network access is a concrete hub/network bug; 'racoon' has no comparable evidence." + "reason": "local_files_only network leakage is a distinct bug from the unrelated 'racoon' issue." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Reject: no_grad decoration for cache update methods is a specific autograd fix, not the broad 'many models break' report." + "reason": "A no_grad decorator change for cache/layer updates is not the same bug as a broad 'v5 breaks many models' complaint." }, { "left": "issue:43010", "right": "issue:44368", "accept": false, - "reason": "Reject: cache/layer update autograd behavior is unrelated to tie_word_embeddings warning spam in fine-tuning." + "reason": "No_grad/update semantics are unrelated to the tie_word_embeddings warning during LoRA fine-tuning." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "Reject: MLflow reporting on failed training jobs is unrelated to tensor-parallel training behavior." + "reason": "MLflow reporting on failed jobs and tensor-parallel training are unrelated code paths." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Reject: compiled Python file loading errors and PretrainedConfig type-checking regressions are different code paths." + "reason": "Compiled-file execution errors do not match the separate PretrainedConfig type-checking regression." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Reject: Liger Kernel not applied via model_init and a DSA indexer ReLU complaint are unrelated." + "reason": "Liger Kernel application in model_init is unrelated to the DSA indexer/ReLU discussion." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Reject: compiled-file import/runtime errors are not the same as save_pretrained failing for Pipeline objects." + "reason": "Compiled Python file errors are not the same as pipeline save_pretrained failures." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Reject: tokenizer loading regression and Llama 3.2 Vision model loading breakage are different loader failures." + "reason": "Both involve loading, but one is tokenizer loading regression and the other is a specific Llama 3.2 Vision model-loading failure." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Reject: Qwen3VL MoE loading on v5.0 and checkpoint reload failures across v4.57->v5.2/5.3 are not the same defect." + "reason": "Both are load failures, but for different versions and model families; not the same concrete bug." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Reject: a broken env command and config type-checking regression are separate issues." + "reason": "CLI env command breakage and PretrainedConfig type checking are different failures." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Reject: missing `is_offline_mode` import from huggingface_hub is unrelated to the v5 env command breakage." + "reason": "An import error from huggingface_hub is unrelated to the transformers env command breakage." }, { "left": "issue:33453", "right": "issue:36296", "accept": false, - "reason": "Reject: tokenizer loading regression and tensor-parallel training bug affect different subsystems." + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training." }, { "left": "issue:43010", "right": "issue:45656", "accept": false, - "reason": "Reject: cache `update()` autograd decoration is unrelated to optimizer stepping twice under deepspeed." + "reason": "Cache/layer update decoration and double optimizer stepping under deepspeed are different bugs." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Reject: env command breakage and missing Qwen2.5-VL import are different v5 regressions." + "reason": "A broken env command and a missing Qwen2.5-VL export are separate issues." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Reject: FP8 save_pretrained for MoE and FP8 inference support for Qwen3-VL/Qwen3.5 MoE are distinct features/failures." + "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; different bugs." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Reject: unrelated placeholder issue versus Gemma4AudioRelPositionalEncoding bug." + "reason": "The 'racoon' placeholder issue is unrelated to Gemma4 audio positional encoding." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Reject: doc confusion for model.forward args is unrelated to an import error for `layer_type_validation`." + "reason": "Doc confusion on forward arguments is unrelated to the layer_type_validation import error." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Reject: AutoImageProcessor import failure and PreTrainedModel import failure are separate API export issues." + "reason": "Two different missing-symbol import errors for different APIs; not one underlying bug." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Reject: no_grad decoration issue is unrelated to Python 3.9/3.10 support breakage." + "reason": "Cache update decoration and Python-version compatibility are unrelated." }, { "left": "issue:43576", "right": "issue:44938", "accept": false, - "reason": "Reject: broken env command and Python 3.14 load failure are unrelated compatibility issues." + "reason": "The env command issue is not the same as Python 3.14 load failures." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Reject: inverse_sqrt scheduler kwargs bug is unrelated to the placeholder issue." + "reason": "Scheduler kwargs handling is unrelated to the 'racoon' report." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Reject: Liger Kernel application bug is unrelated to FP8 inference on B200 GPUs." + "reason": "Liger Kernel/model_init and FP8 inference on B200 are different code paths." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Reject: tensor-parallel training bug is unrelated to TextClassificationPipeline docs wording." + "reason": "Tensor-parallel training and text-classification docs are unrelated." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Reject: tensor parallel training bug and BeitImageProcessorFast reduce_labels bug are different components." + "reason": "Tensor-parallel training and BeitImageProcessorFast label reduction are different bugs." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Reject: missing files on processor.save_pretrained is a concrete bug; 'Bug report' provides no evidence it's the same issue." + "reason": "A missing-files save_pretrained bug is not the same as a vague placeholder bug report." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Reject: lazy loading behavior and TensorBoard/TensorFlow backend detection are unrelated." + "reason": "Lazy loading and a TensorFlow backend error from tensorboard-only installs are separate issues." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Reject: Qwen3-VL-Moe loading regression and missing `_tp_plan` for Qwen3_5Moe are different model integration problems." + "reason": "Qwen3-VL-Moe breakage and missing _tp_plan for Qwen3.5 MoE are related only by theme, not the same bug." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "Reject: FP8 inference support for Qwen models is unrelated to a request for a weight conversion script." + "reason": "FP8 inference support and a DeepSeek weight-conversion question are unrelated." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Reject: lazy loading regression and unsafe `sys.modules[]` access are distinct implementation bugs." + "reason": "Lazy loading malfunction is not the same as unsafe sys.modules access." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Reject: Qwen3-VL bounding-box output issue is unrelated to config type-checking regressions." + "reason": "Qwen3-VL bounding-box output problems are unrelated to PretrainedConfig type checking." }, { "left": "issue:44373", "right": "issue:44961", "accept": false, - "reason": "Reject: wrong docstring for `position_ids` is unrelated to the placeholder issue." + "reason": "A docstring fix for position_ids is unrelated to the 'racoon' issue." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Reject: env command breakage and checkpoint reload compatibility are different failure modes." + "reason": "Different version/regression reports with different symptoms and upgrade scenarios." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Reject: Llama4Config `pad_token_id` attribute error is a model config bug, not the broad v5 breakage umbrella." + "reason": "Missing pad_token_id in Llama4Config is a model-specific config bug, not the broad v5 breakage complaint." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Reject: 2D bounding-box output hallucination and an infinite-loop image case are different multimodal failures." + "reason": "Both involve Qwen3_5/Qwen3-VL, but one is bounding-box output and the other is an infinite loop on a specific image." } ] }, @@ -15427,8 +15556,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -15524,20 +15653,20 @@ "issue:41669|issue:43704", "issue:31795|issue:34689" ], - "prepared_review_unit_hash": "d1758d31ffe0c03b041d508512242a970377e04d86bdfa69ccdf9cf33025d4c9", + "prepared_review_unit_hash": "9d863233e03de2f7b1ae51ecec64557cc2168aa341fe224cbe4e45d1eb4ae9b9", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39265, + "serialized_chars": 39266, "estimated_input_tokens": 9817, "estimated_eval_tokens": 19890 }, @@ -15546,258 +15675,258 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of broad Transformers v5 regression/import-compatibility issues plus a handful of unrelated model-specific bugs. There are no strong duplicate PRs here, and most soft pairs are only loosely similar by subsystem or upgrade version.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45278 is the best umbrella candidate: it clearly captures the common v5 upgrade/regression theme with a concrete symptom (many import errors after moving from 4.57.0 to 5.5.0).", + "summary": "This is a loose grab-bag of Transformers issues, mostly v5 regressions and import/model-loading bugs, but the soft pairs generally describe different concrete code paths rather than duplicates.", + "confidence": 0.63, + "canonical_issue_reason": "issue:45850 is the broadest umbrella report about v5 breaking many models, so it best represents the cluster theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most actionable representative of the cluster; broader catch-all titles like 45850 are too vague to serve as a good canonical issue.", + "best_issue_reason": "issue:45850 is the best triage anchor because it captures the widest cross-model impact among the v5 regression reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Different problems: import-time bloat from `import *` vs missing cache state during chunked prefill." + "reason": "Both mention import/loading behavior, but one is about import * bloat and the other about chunked_prefill cache behavior." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are v5 loading regressions, but they describe different load paths and different failure modes." + "reason": "Both are v5 upgrade breakages, but they affect different model-loading paths and different failure modes." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific config regression vs broad complaint about many broken models; not the same bug." + "reason": "One is a StableLmConfig missing field regression; the other is a broad complaint about many models breaking." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "Model-specific chat crash, not the same underlying issue as a generic v5 breakage complaint." + "reason": "Qwen chat crash is a specific model/API failure, not the same bug as the broad v5 breakage report." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-performance regression vs a missing model class import; related area, different defect." + "reason": "Import-time slowdown from model imports is unrelated to a missing Qwen3_5 module import error." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Both mention import/lazy loading, but one is slow import and the other is lazy loading misbehavior; too vague to merge." + "reason": "Both concern imports/lazy loading, but one reports slow import timing and the other says lazy loading is broken." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Unrelated test flakiness vs scheduler kwargs handling." + "reason": "Test seeding flakiness and scheduler kwargs handling are unrelated." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Different model instantiation issues with different causes and code paths." + "reason": "Different subsystems: model-type mismatch warning versus Liger Kernel application in model_init." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs forwarding vs scheduler kwargs forwarding; same pattern, different APIs." + "reason": "AutoProcessor kwargs forwarding and inverse_sqrt scheduler kwargs are unrelated bugs." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save metadata bug vs a placeholder bug report with no clear shared root cause." + "reason": "Tokenizer_class save metadata mismatch is not the same as an unspecified bug report." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility issue vs Qwen3-VL bounding-box output bug." + "reason": "Python version compatibility and Qwen3-VL bounding box output are different problems." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Position_ids buffer access error is unrelated to the generic bug report." + "reason": "position_ids buffer access error is unrelated to the generic bug report." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention support error vs a multimodal output correctness issue." + "reason": "Flash Attention support warning and Qwen3-VL output errors are different code paths." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken env command vs scheduler kwargs handling; different subsystems." + "reason": "Broken env CLI and scheduler kwargs handling are unrelated." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler parameter plumbing vs PretrainedConfig type-checking regression." + "reason": "Scheduler kwargs omission and PretrainedConfig type checking are different issues." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import performance regression is not the same as a missing exported symbol." + "reason": "Import * cleanup/import overhead is not the same as a missing HybridCache export." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "No clear overlap: import slowdown vs a testing utility bug." + "reason": "Model import bloat and a testing_utils bug do not share the same underlying change." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Both are broad v5 breakage reports, but one is specifically import errors and the other is generic model breakage." + "reason": "Both are broad v5 regressions, but one is about import errors and the other about general model breakage." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Chunked-prefill cache regression vs a cache-strategy typo." + "reason": "Chunked_prefill cache missing and a cache-strategy typo are not the same bug." }, { "left": "issue:43874", "right": "issue:45850", "accept": false, - "reason": "Missing image-processor method for one model, not the same as a broad v5 breakage umbrella." + "reason": "Missing image-patch helper in GLM46V is a specific multimodal API gap, not the broad v5 regression." }, { "left": "issue:46143", "right": "issue:46144", "accept": false, - "reason": "Related RoFormer area, but one is kwargs forwarding and the other is attention-interface usage." + "reason": "One is missing kwargs passthrough, the other is attention implementation wiring; related area but different fixes." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python compatibility vs a specific missing Qwen3.5 class import." + "reason": "Python compatibility and a Qwen3_5 import failure are unrelated." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5/Qwen3-VL model bugs with distinct failure modes." + "reason": "Different model families and different failure reports." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "Optional `post_init` validation vs `_tied_weights_keys` type handling." + "reason": "Optional post_init validation and tied weight key AttributeError are unrelated." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Documentation mentioning removed `pipeline()` APIs vs scheduler kwargs behavior." + "reason": "Documentation still referencing pipeline() is unrelated to scheduler kwargs handling." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion vs tokenizer-loading regression; not the same underlying issue." + "reason": "Documentation confusion and tokenizer loading regression are different issues." }, { "left": "issue:41669", "right": "issue:44492", "accept": false, - "reason": "Import-time performance regression vs a typo in cache-strategy text." + "reason": "Import * cleanup and a typo in cache strategies are not duplicates." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger kernel application in `model_init` vs flash-attention-3 training degeneration." + "reason": "Liger Kernel application and flash_attention_3 training degeneration are different concrete bugs." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 settings migration vs cache-strategy typo." + "reason": "TF32 API guidance and a cache-strategy typo do not match." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken `transformers env` command vs lazy-loading behavior; similar theme, different issue." + "reason": "Broken env command and lazy loading failure are adjacent but distinct issues." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Pipeline save/export problem vs tied-weights key handling." + "reason": "Pipeline save_pretrained and tied weight key handling are unrelated." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA/ReLU discussion vs Gemma-4 audio positional encoding bug; unrelated." + "reason": "DSA indexer ReLU discussion and Gemma4 audio positional encoding are unrelated." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report vs Qwen3.5 MoE from_pretrained regression." + "reason": "Generic bug report versus a concrete Qwen3.5 MoE from_pretrained error." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading/import behavior vs scheduler kwargs handling." + "reason": "Lazy loading failure and inverse_sqrt scheduler kwargs are unrelated." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Broken CLI command vs missing model class import." + "reason": "Env CLI breakage and Qwen3_5 module import failure are different bugs." }, { "left": "issue:43827", "right": "issue:44492", "accept": false, - "reason": "Docs still referencing removed pipeline APIs vs a typo in cache strategies." + "reason": "Docs referencing pipeline() and a cache-strategy typo are not the same issue." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Different DeepSeek-related model bugs with different symptoms and code paths." + "reason": "Both are DeepSeek-V4 related, but one is about gating and the other about tensor transpose/casting." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp calculation vs GLM-5 RoPE implementation; different model math bugs." + "reason": "Qwen3VL timestamp math and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp calculation bug vs scheduler kwargs bug." + "reason": "Timestamp calculation and scheduler kwargs are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-performance regression vs VRAM leak in multi-threaded dataloader use." + "reason": "Import cleanup/performance and VRAM leak in multithreaded use are different failure modes." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation issue vs a model-loading breakage in a specific release." + "reason": "Documentation confusion and Llama 3.2 Vision model loading breakage are unrelated." } ] }, @@ -15816,8 +15945,8 @@ "review_unit_count": 50, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-433", "nodes": [ @@ -15908,286 +16037,286 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "faa97116d452c534f309484d9fa3122e60be3225c4b1b51890eb3976a6ec6bd5", + "prepared_review_unit_hash": "c810d020f658a421d92c689d57ce5b3691e8e6fe33e5f65186aea132e6efcdac", "component_budget": { "node_count": 433, "item_count": 433, "soft_pair_count": 1595, - "serialized_chars": 826484, - "estimated_input_tokens": 206621, - "estimated_eval_tokens": 413498 + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 }, "budget": { "node_count": 42, "item_count": 42, "soft_pair_count": 42, - "serialized_chars": 36084, - "estimated_input_tokens": 9021, - "estimated_eval_tokens": 18298 + "serialized_chars": 36087, + "estimated_input_tokens": 9022, + "estimated_eval_tokens": 18300 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated Transformer issues: v5 regressions/config/import/tokenizer changes, plus a few model-specific bugs. None of the soft pairs look like true duplicates or mergeable PR-equivalents; the broad v5-breakage report is the best umbrella issue.", - "confidence": 0.82, - "canonical_issue_reason": "issue:45850 is the broadest umbrella report in the set, capturing the recurring theme that Transformers 5.0 breaks many models and APIs.", + "summary": "This cluster is highly heterogeneous: the issues span unrelated model bugs, tokenizer regressions, config/import errors, docs, and performance complaints. None of the soft links look like the same underlying bug or change, so all are rejected.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 best represents the cluster because it is the only clearly general, cross-cutting issue rather than a narrow model-specific symptom.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint loading slowness and documentation confusion about forward args are unrelated." + "reason": "Different problems: slow checkpoint loading vs confusing forward() docs for input_ids/past_key_values." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "GLM-5 RoPE behavior and Gemma-4 audio positional encoding are different model implementations." + "reason": "Unrelated model internals: GLM-5 RoPE implementation vs Gemma-4 audio positional encoding." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are different change requests." + "reason": "Different topics: continuous batching support for audio models vs a DeepSeek-V4 weight conversion question." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leakage in Qwen3 is unrelated to GLM-5 RoPE implementation details." + "reason": "VRAM leak in multithreaded dataloading is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad v5 regression complaint is not the same underlying bug as a specific Qwen3_5 import failure." + "reason": "Generic v5 breakage report vs a specific Qwen3.5 import failure; not the same bug." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network leakage and Qwen3VL timestamp math are unrelated defects." + "reason": "Local-files-only network leakage vs a Qwen3VL timestamp calculation bug." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings guidance is unrelated to a Qwen3 VRAM leak." + "reason": "TF32 settings deprecation is unrelated to the VRAM leak report." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation in a processor and missing saved processor files are different code paths." + "reason": "Different processor bugs: timestamp math vs save_pretrained missing files." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Processor timestamp math and dataloader-thread VRAM leak are unrelated." + "reason": "Timestamp calculation bug and VRAM leak are unrelated code paths." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are not the same failure." + "reason": "CPU offload device mismatch is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split handling in a MoE indexer is unrelated to Gemma-4 audio positional encoding." + "reason": "torch.split return values in GlmMoeDsaIndexer is unrelated to Gemma-4 audio encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "RMS norm epsilon precision mismatch is a different bug from Gemma-4 audio positional encoding." + "reason": "Different model components: MLA q_a_layernorm epsilon handling vs Gemma-4 audio positional encoding." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "HybridCache import failure and Qwen3_5 module import failure are separate missing-symbol problems." + "reason": "Different missing symbols: HybridCache import vs Qwen3_5ForCausalLM import." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention support gating and an import error are unrelated issues." + "reason": "Flash Attention support gating and a module import failure are distinct issues." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "General tokenizer behavior change and wrong decoder type are too broad and not clearly the same bug." + "reason": "Both tokenizer-related, but one is a general v4-v5 behavior change and the other is a specific decoder-type mismatch." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs handling is unrelated to the unrelated 'Unemployment' issue." + "reason": "Scheduler kwargs bug is unrelated to the unrelated 'Unemployment' issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight-key crash and chat template tool-call crash happen in different subsystems." + "reason": "Tied-weight key handling and chat template tool-call crashes are separate failures." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in config is unrelated to tied weight key handling." + "reason": "Missing pad_token_id in Llama4Config is unrelated to _get_tied_weight_keys crashing." }, { "left": "issue:43704", "right": "issue:44998", "accept": false, - "reason": "A VRAM leak and an unrelated placeholder issue have no technical overlap." + "reason": "VRAM leak in dataloader threads is unrelated to the unrelated issue 44998." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "Missing StableLmConfig field and tied-weight-key crash are different config/model issues." + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key AttributeError." }, { "left": "issue:43066", "right": "issue:44998", "accept": false, - "reason": "Wrong tokenizer decoder type is unrelated to the non-technical 'Unemployment' issue." + "reason": "Tokenizer decoder type regression is unrelated to the unrelated issue 44998." }, { "left": "issue:41669", "right": "issue:44998", "accept": false, - "reason": "Import-star performance cleanup and the unrelated placeholder issue do not match." + "reason": "Import-star performance cleanup is unrelated to the unrelated issue 44998." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading and an ImportError for layer_type_validation are different failure modes." + "reason": "Checkpoint load slowness and missing layer_type_validation import are different bugs." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 support for Qwen3-VL/Qwen3.5 MoE is unrelated to processor save_pretrained file omission." + "reason": "FP8 inference support for Qwen3-VL is unrelated to processor.save_pretrained missing files." }, { "left": "issue:44998", "right": "issue:46218", "accept": false, - "reason": "The placeholder issue is unrelated to a Qwen3_5 import failure." + "reason": "The unrelated issue 44998 has nothing in common with a Qwen3.5 import error." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow job-reporting after failure and checkpoint loading speed are unrelated." + "reason": "MLflow reporting on failed training jobs is unrelated to slow from_pretrained checkpoint loading." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowness and tensor-parallel training bug do not describe the same problem." + "reason": "Checkpoint loading performance and tensor-parallel training bugs are different code paths." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 inference support and Qwen3.5 MoE from_pretrained errors are different requests/bugs." + "reason": "FP8 support request is not the same as a Qwen3.5 MoE from_pretrained regression." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Checkpoint loading performance and tokenizer loading regression are separate issues." + "reason": "Slow from_pretrained loading and tokenizer loading regression are distinct issues." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching for audio models and FP8 inference support are different features." + "reason": "Continuous batching support for audio models is unrelated to fp8 inference support." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights behavior is unrelated to RoFormer attention-interface usage." + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2Former ignore_value handling and DeepSeek-V4 expert gating are different model bugs." + "reason": "Mask2Former ignore_value usage and DeepSeek-V4 shared expert gating are different model bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Config type-hint correction and tokenizer regex warning are unrelated." + "reason": "ModernBERT norm_eps type hint and tokenizer regex warnings are unrelated." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value and loss kwargs passthrough are unrelated code paths." + "reason": "Unused ignore_value and kwargs not passed to loss calculation are different defects." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs handling is unrelated to RoFormer attention implementation." + "reason": "Loss kwargs propagation is unrelated to RoFormer attention implementation." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2Former ignore_value bug and tokenizer regex warning are not the same underlying defect." + "reason": "Mask2Former ignore_value bug does not match the tokenizer regex warning regression." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2Former ignore_value usage and RoFormer attention interface are different subsystems." + "reason": "Mask2Former ignore_value handling is unrelated to RoFormer attention implementation." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "A Mask2Former runtime bug is unrelated to a typo fix." + "reason": "Unused ignore_value and a typo in quantization_operation are unrelated." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs forwarding and a ModernBERT type hint issue are unrelated." + "reason": "Passing kwargs to loss and ModernBERT type hints are separate issues." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Type-hint correction and a typo in a string literal are not the same bug." + "reason": "Incorrect type hint and a typo in a conversion string are unrelated." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "A question about Llama4TextL2Norm semantics is unrelated to a ModernBERT type-hint bug." + "reason": "Llama4TextL2Norm normalization question is unrelated to ModernBERT norm_eps typing." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2Former ignore_value handling is unrelated to ModernBERT norm_eps typing." + "reason": "Mask2Former ignore_value usage is unrelated to a ModernBERT type-hint bug." } ] }, @@ -16206,8 +16335,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10519-3", "nodes": [ @@ -16220,52 +16349,52 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "542dc3b8a12a47fd5fb99f2ed264916110dc535ed7b64aed468c502307d8f99c", + "prepared_review_unit_hash": "174914c79743347044c48a7bcd83514e9ffd8ed6e3fab526224639a2284c3f4e", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2877, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2877, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three issues are distinct feature requests in different areas: truncation behavior, learning-rate scheduling, and Wav2Vec2 word timing offsets. None of the soft pairs looks like the same underlying problem.", + "summary": "These three items are unrelated feature requests: truncation behavior in tokenization, ReduceLROnPlateau-like scheduler functionality, and Wav2Vec2 word time offsets. The soft similarities are superficial, so none should be merged as duplicates.", "confidence": 0.95, - "canonical_issue_reason": "Issue 10519 is the clearest, most self-contained feature request with a specific API change request and well-bounded scope.", + "canonical_issue_reason": "issue:10519 is the clearest and most self-contained feature request, with a specific API change request and explicit scope for both truncation modes.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 10519 is the best representative issue because it states a concrete change request rather than a broad or ambiguous feature idea.", + "best_issue_reason": "issue:10519 is the best standalone representative because it is concrete, narrowly scoped, and describes an actionable enhancement rather than a broad question.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Both mention beginning/end wording, but one is about tokenizer truncation strategy and the other about extracting word offsets from Wav2Vec2; unrelated functionality." + "reason": "Different domains and problems: truncation strategy vs Wav2Vec2 word boundary timestamps. No shared code path or concrete duplicate relationship." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Tokenizer truncation options and ReduceLROnPlateau-like scheduler functionality are unrelated feature requests in different subsystems." + "reason": "One asks for tokenizer truncation-from-beginning behavior; the other asks for learning-rate scheduler functionality. These are unrelated features." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "A scheduler enhancement request and a speech-model timing-offset request do not describe the same bug or change." + "reason": "Scheduler feature request and Wav2Vec2 timestamp extraction are distinct tasks with no overlapping implementation target." } ] }, @@ -16284,8 +16413,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-13244-3", "nodes": [ @@ -16297,7 +16426,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "f5ca8884235bbe4b4cf0ee9e2cda2cd9a87d3f0e3bca9c0d541027e739f56f82", + "prepared_review_unit_hash": "eb758d59b360229cf07a5ccab9b6bb0bd7ba3f7e0815a712eec08c04a5715017", "component_budget": { "node_count": 3, "item_count": 3, @@ -16319,24 +16448,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues do not look like duplicates: one is about Tapas tokenization mismatch, one about a login/authentication error when downloading a private model, and one about failing to load bert-base-uncased. The soft similarities are too broad to treat as the same underlying bug.", - "confidence": 0.92, - "canonical_issue_reason": "No strong duplicate cluster emerges here; if forced to pick a representative, issue 16618 is the most concrete and user-visible model-loading failure.", + "summary": "No duplicate-quality match here. The only related pair is two model-loading complaints, but they describe different failure modes; the tokenization issue is unrelated.", + "confidence": 0.78, + "canonical_issue_reason": "Issue #16618 is the most general and representative of the shared model-loading theme, so it works best as the cluster\u2019s anchor despite not being a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 16618 is the best standalone representative because it describes a specific, reproducible loading problem with a common model name and broad user impact.", + "best_issue_reason": "Issue #16618 is the strongest standalone issue: it describes a concrete user-facing failure to load a named model, which is more central than the narrower auth/error-message report and far more relevant than the tokenization bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch is a tokenizer/behavior discrepancy, while 16618 is a model loading failure; they are different code paths and underlying problems." + "reason": "Different subsystems and symptoms: tokenization parity vs. model loading failure. No evidence they stem from the same bug." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "15980 is about an unauthenticated private-model download error message, not the same failure as being unable to load bert-base-uncased." + "reason": "Both are model-download/load related, but one is specifically a private-model login/error-message problem and the other is a generic failure loading bert-base-uncased; not the same underlying issue." } ] }, @@ -16355,8 +16484,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-26350-2", "nodes": [ @@ -16366,12 +16495,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "4fdee17ccfcf2e4ba70d02fcbd47b45d175bcb865189f7755979f87536f9bf74", + "prepared_review_unit_hash": "93d12371ab4d5ce713c3f024ddea9c008b91c986115303281b2fbc79149aa99b", "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 }, @@ -16379,7 +16508,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -16388,18 +16517,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two distinct community-contribution feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share the broad theme of adding architecture support, but they target different subsystems and concrete changes, so they should not be merged as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "Issue #26350 is the stronger canonical candidate because it has much higher discussion activity and is the earlier, more established request, but it is still a separate feature request from the GGUF item.", + "summary": "These are two distinct community-contribution feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share a similar framing but not the same underlying change or code path.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 26350 is the better cluster anchor because it is older and has more discussion activity, but it is not a duplicate of 33260.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #26350 is the best representative issue for this cluster because it is the most active and central item, even though the cluster is not a true duplicate set.", + "best_issue_reason": "Issue 26350 is the most suitable representative for the cluster only by age/activity; the requests themselves are unrelated features.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Both are support-expansion requests, but Flash Attention 2 and GGUF are different features with different code paths and implementation work; they are not the same underlying issue." + "reason": "Different feature requests in different subsystems: Flash Attention 2 support vs GGUF support. Similar contribution template, not the same bug/change." } ] }, @@ -16418,8 +16547,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-27088-8", "nodes": [ @@ -16459,188 +16588,191 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "652de6602a3b79daa0ca4a4e0dc498579f7841d196f01535cb43e394cfaa4384", + "prepared_review_unit_hash": "d2d9dea9ea57b8bc1e6293bfc4ed48f2a83022d4a60f235294aa023180637702", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 25, - "serialized_chars": 13679, - "estimated_input_tokens": 3420, - "estimated_eval_tokens": 7096 + "serialized_chars": 13675, + "estimated_input_tokens": 3419, + "estimated_eval_tokens": 7094 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 25, - "serialized_chars": 13679, - "estimated_input_tokens": 3420, - "estimated_eval_tokens": 7096 + "serialized_chars": 13675, + "estimated_input_tokens": 3419, + "estimated_eval_tokens": 7094 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Most items are separate i18n translation tracking issues for different languages and should not be merged. The only clear duplicate set is the generic template issues with the exact same placeholder title; the Thai-titled issue does not look like the same tracker.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 46222 is the best canonical representative of the exact generic template cluster: it matches the same placeholder title as 43834/44365/44370, is still open, and appears to be the most current active copy.", + "summary": "This cluster is mostly a set of i18n translation-request issues. The specific-language tickets (Turkish, Spanish, Romanian) are distinct from each other, while the placeholder-title i18n issues are near-identical and look like duplicate template instances.", + "confidence": 0.78, + "canonical_issue_reason": "issue:46222 is the best representative of the duplicate placeholder i18n template issues: it is current, open, and structurally matches the other generic translation-template tickets more closely than the language-specific requests.", "canonical_pr_reason": null, - "best_issue_reason": "46222 is the strongest global issue candidate because it is open, current, and exactly matches the duplicate template issues, while the language-specific issues are distinct translation requests.", + "best_issue_reason": "issue:46222 is the strongest cluster anchor because it represents the generic translation-template form and appears to be the latest active instance among the near-identical placeholder issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different target languages and different translation requests; same broad i18n subsystem but not the same issue." + "reason": "Different concrete translation requests; Turkish vs placeholder generic template are not the same issue." }, { "left": "issue:27088", "right": "issue:46222", "accept": false, - "reason": "Turkish-specific translation issue vs a generic placeholder-template translation tracker." + "reason": "Turkish translation request is distinct from the generic placeholder translation template." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Different language/request scope; not the same underlying issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Specific Turkish translation request does not duplicate the generic placeholder ticket." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Turkish and Romanian translation requests are separate language tasks." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Turkish and Spanish translation requests are separate issues, not duplicates." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation tracker is not the same as the generic placeholder issue." + "reason": "Spanish request vs generic placeholder template; different underlying ticket." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Spanish translation request is distinct from the generic placeholder issue." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Spanish and Romanian translation requests are different language-specific issues." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Spanish request does not match the generic placeholder translation template as the same issue." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation tracker is distinct from the generic placeholder issue." + "reason": "Romanian translation request vs placeholder template; not the same concrete issue." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different language-specific translation issue vs generic placeholder tracker." + "reason": "Romanian request is distinct from the generic placeholder translation ticket." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Romanian translation task is separate from the placeholder template issue." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Different language-specific translation issue." + "reason": "Different language-specific request versus generic placeholder template." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "No evidence this Thai-titled issue is the same translation request; titles do not match." + "reason": "Romanian translation request and the Thai-titled issue do not look like the same underlying ticket." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Exact same placeholder template title and same underlying generic i18n translation tracker." + "reason": "Both are near-identical placeholder i18n translation-template issues with the same title/body pattern." }, { "left": "issue:43834", "right": "issue:46222", "accept": true, - "reason": "Exact same placeholder template title; both appear to be the same tracker copy." + "reason": "These look like duplicate instances of the same generic translation-template issue." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Exact same placeholder template title and same underlying request." + "reason": "Same placeholder translation-template issue, no meaningful distinction visible." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Thai-titled issue does not clearly match the generic placeholder translation tracker." + "reason": "Thai-titled issue is not clearly the same generic placeholder translation template." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Exact same placeholder template title; same duplicate tracker." + "reason": "Near-identical placeholder i18n translation-template issues." }, { "left": "issue:44365", "right": "issue:46222", "accept": true, - "reason": "Exact same placeholder template title; same duplicate tracker." + "reason": "Same generic placeholder translation issue, likely duplicate." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "No clear evidence the Thai-titled issue is the same tracker." + "reason": "No clear evidence that the Thai-titled issue is the same underlying placeholder ticket." }, { "left": "issue:44370", "right": "issue:46222", "accept": true, - "reason": "Exact same placeholder template title; same underlying issue." + "reason": "These are effectively the same placeholder translation-template issue." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "No clear evidence the Thai-titled issue is the same tracker." + "reason": "The Thai-titled issue does not match the placeholder generic translation template closely enough to merge." }, { "left": "issue:46222", "right": "issue:46226", "accept": false, - "reason": "Different titles and no evidence they are the same translation request." + "reason": "Generic placeholder translation-template issue and the Thai-titled issue are not clearly the same ticket." } ] }, - "evaluator_result": null, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the generic template issues ([i18n-] Translating docs to ) are extremely close to each other, while the language-specific tickets are clearly different by title and are reasonably kept separate. The soft-edge verdicts are conservative overall, and the canonical issue choice (46222) is supported by its open state and recency among the near-identical template issues." + }, "evaluator_used": true, "retried": false, "accepted_nontrivial_soft_edge": true, @@ -16655,8 +16787,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-36817-16", "nodes": [ @@ -16715,262 +16847,250 @@ "issue:43837|issue:45177", "issue:45322|issue:46133" ], - "prepared_review_unit_hash": "8435704b1d7491c09b9b6cc3ebfb149a7f65f757c0231093b4977cebf2751e0f", + "prepared_review_unit_hash": "41501de06e65798bd4a694373a44742731df0e6039b194238226a014400d318c", "component_budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21656, - "estimated_input_tokens": 5414, - "estimated_eval_tokens": 11084 + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 }, "budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21656, - "estimated_input_tokens": 5414, - "estimated_eval_tokens": 11084 + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of broadly similar model-request issues, but they are not duplicates: each asks for a different architecture or modality support. The shared theme is only \u201cadd model support,\u201d not the same underlying change.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue: the items request different models/features (e.g. EuroBert, RF-DETR, NeoBERT, Qwen3-TTS/ASR, Voxtral, TIPSv2), so they should not be collapsed into one issue.", + "summary": "These are all separate model-request issues, not duplicates. The similarity is only at a broad topical level (\u201cadd support for a model\u201d), while the requested models and modalities differ materially.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue: each item requests a different model or feature (e.g., EuroBert, RF-DETR, Qwen3-TTS/ASR, ModernVBERT, Voxtral TTS, etc.), so they are not the same underlying request.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue for deduplication; none of these issues is a clear duplicate of the others.", + "best_issue_reason": "No global best issue for deduplication: none of these issues subsumes the others, and the overlap is only the generic 'add model support' theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different requested models (EuroBert vs DEIMv2); same broad category only." + "reason": "Different model requests: EuroBert config support vs DEIMv2. Same broad theme, but not the same underlying change." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config request is unrelated to Qwen3-TTS support." + "reason": "EuroBert config support and Qwen3-TTS support are unrelated model additions with different modalities." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert vs RF-DETR are distinct model additions." + "reason": "EuroBert vs RF-DETR are different models and different code paths; not duplicates." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are separate model architectures." + "reason": "RF-DETR and DEIMv2 are distinct model requests, not one concrete duplicate." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR and LFM2.5 Audio are different code paths and modalities." + "reason": "RF-DETR (vision detection) and LFM2.5 Audio 1.5B are different modalities and changes." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR and weight-sparse transformer requests are not the same change." + "reason": "RF-DETR and OpenAI weight-sparse transformer are separate model requests with no shared concrete implementation target." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR detection support is unrelated to Qwen3-ASR support." + "reason": "RF-DETR (vision) vs Qwen3-ASR (audio speech recognition) are different underlying features." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are different model integrations." + "reason": "RF-DETR and Qwen3-TTS are different models and different task domains." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT request is unrelated to Qwen3-ASR support." + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions; only the broad 'add support' theme matches." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT and circuit-sparsity transformer are distinct model changes." + "reason": "NeoBERT vs OpenAI weight-sparse transformer are different architectures and request targets." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT is not the same as LFM2.5 Audio support." + "reason": "NeoBERT is a text model; LFM2.5 Audio 1.5B is an audio model request." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are separate model requests." + "reason": "RF-DETR and NeoBERT are different model requests, not the same underlying bug/change." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different BERT-family models, but not the same concrete request." + "reason": "NeoBERT and ModernVBERT models are distinct requests despite both being BERT-like names." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support is unrelated to Qwen3-ASR support." + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different model features." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 and circuit-sparse transformer support are different model changes." + "reason": "MiniCPM3 and the weight-sparse transformer request are separate model additions." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are distinct architectures." + "reason": "MiniCPM3 architecture and ModernVBERT are different concrete model support requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS support are unrelated." + "reason": "MiniCPM3 and Voxtral-4B-TTS-2603 are different model families and modalities." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different model additions." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR support are not the same change." + "reason": "ModernVBERT models and Qwen3-ASR are different additions; not mergeable as one duplicate." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are distinct model requests." + "reason": "ModernVBERT and Qwen3-TTS are distinct model support requests." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT and BERT-like models with RoPE are related in theme only, not the same concrete request." + "reason": "ModernVBERT and BERT-like models with RoPE overlap only loosely on BERT naming, but request different implementations." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE support is unrelated to Voxtral TTS support." + "reason": "BERT-like models with RoPE and Voxtral TTS are unrelated features." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "TTS and ASR support are different tasks and likely different implementations." - }, - { - "left": "issue:43671", - "right": "issue:43909", - "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are separate model integrations." - }, - { - "left": "issue:43671", - "right": "issue:44121", - "accept": false, - "reason": "Qwen3-TTS support is not the same as adding a weight-sparse transformer." + "reason": "Qwen3-TTS vs Qwen3-ASR are related product names but different concrete modalities and code paths." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral-4B-TTS are different TTS models." + "reason": "Qwen3-TTS and Voxtral TTS are different TTS model requests, not the same model or change." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated modalities." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are different task domains (speech vs OCR)." }, { - "left": "issue:43837", - "right": "issue:43909", + "left": "issue:43671", + "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio requests are different model changes." + "reason": "Qwen3-TTS and the weight-sparse transformer request are unrelated model additions." }, { "left": "issue:43837", - "right": "issue:44121", + "right": "issue:45322", "accept": false, - "reason": "ASR support and weight-sparse transformer support are unrelated." + "reason": "Qwen3-ASR and EUPE are different model requests in different domains." }, { "left": "issue:43837", - "right": "issue:45322", + "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and EUPE are different model families and use cases." + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are both audio-related but not the same concrete model or implementation." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are separate model requests." + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model requests; no shared code-path bug/change." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are distinct audio-model support requests." + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are different audio model requests, not duplicates." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer support and EUPE are different proposed additions." + "reason": "OpenAI weight-sparse transformer and EUPE are distinct model requests with different targets." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "ASR support and circuit-sparse transformer support are not the same code-path change." + "reason": "Qwen3-ASR and OpenAI weight-sparse transformer are unrelated requests." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different model integrations." + "reason": "Qwen3-ASR and Voxtral-4B-TTS-2603 are different speech models and different support changes." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are unrelated requests." + "reason": "NeoBERT and Voxtral-4B-TTS-2603 are unrelated model additions." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are not the same model addition." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and implementation areas." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are different model requests from different projects." + "reason": "EUPE and TIPSv2 are different model requests from different authors/tasks; only broad vision-language similarity exists." } ] }, @@ -16989,8 +17109,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-41084-2", "nodes": [ @@ -17000,7 +17120,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "ee226305fe4a1ca2ab43d1733cc7bd5a686c9921810b6d828c20f03f8a88c9f0", + "prepared_review_unit_hash": "143e8a03e19b65b70ab0116efff78e12ac55d0fb431defe7b2fbf0a0402e669c", "component_budget": { "node_count": 2, "item_count": 2, @@ -17022,18 +17142,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are about different feature requests in different areas: one is about block decoding, the other about DeepSeek v3.2 support. They do not appear to describe the same underlying bug or change.", + "summary": "These two issues are not duplicates: one is a specific decoding feature request, the other is a model-support request for DeepSeek v3.2. The only similarity is very broad AI/model infrastructure overlap, which is not enough to merge them.", "confidence": 0.97, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue #41084 is the more concrete, narrowly scoped request and is easier to use as a representative anchor for this cluster, but it is not actually the same problem as #43037.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue #41084 is the better canonical-style representative because it is more specific and actionable than the broader model-support request in #43037.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Unrelated requests with no shared code-path or concrete symptom overlap; one asks for block decoding, the other for model/version support." + "reason": "Reject: the titles describe different underlying needs. 'Set Block Decoding' is a decoding behavior/feature request, while 'DeepSeek v3.2 support' is about adding support for a specific model. They do not appear to be the same bug or change." } ] }, @@ -17052,8 +17172,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43598-2", "nodes": [ @@ -17063,12 +17183,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "a416742cae6e73b849247cda8a3df086475de376602d953d7259ddf739049584", + "prepared_review_unit_hash": "a88e3b9538761bfd0a28116276a700b7863ccf192e03b3bc73ec5db9bec7c3d1", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17076,7 +17196,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17085,18 +17205,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are related only at a high level (custom Trainer behavior), but they target different functionality: unused-column removal vs FLOPs estimation. They should not be merged as duplicates.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43598 is the better anchor because it is a concrete Trainer customization problem and represents the broader API/behavior change more directly than the FLOPs-specific report.", + "summary": "The two issues are about different Trainer concerns: one asks to revisit `remove_unused_column` for customizability, while the other asks to account for custom trainers when estimating FLOPS. They share a broad Trainer/customization theme but not the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": "No canonical issue: the reported behaviors are different code paths and different requested fixes.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the strongest representative of the cluster since it focuses on a clear Trainer extensibility concern, while 43600 is a separate metric-accounting issue.", + "best_issue_reason": "Neither issue is a duplicate of the other; if one must be kept as the cluster representative, 43598 is the more concrete Trainer API/customization request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Related theme, but different code paths and outcomes: 43598 is about `remove_unused_column` behavior in `Trainer`, while 43600 is about FLOPs estimation for custom trainers. Not the same underlying bug or change." + "reason": "Different problems: `remove_unused_column` behavior versus FLOPS estimation for custom trainers. Similar subsystem, but not the same underlying issue." } ] }, @@ -17115,8 +17235,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43684-2", "nodes": [ @@ -17126,7 +17246,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "5c2dad1d956705e0f27579685a936eed830b8f9c23155718ebc58409539d2262", + "prepared_review_unit_hash": "78f0280bf96ec4989303cfd373eb4e7f3f3107ec015e52b0d259b08ea937d8b1", "component_budget": { "node_count": 2, "item_count": 2, @@ -17148,18 +17268,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two distinct feature requests for different Qwen model families and tasks, so they are not duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 43684 is the broader model-registration request and is the better representative of the cluster, but it is still separate from 44405.", + "summary": "These are not duplicates: one requests Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5/Qwen3_5Config.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43684 is broader and more directly about model registration support, making it the better representative of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "43684 is the more general support request (AutoModel and AutoModelForConditionalGeneration registration) and is the best single issue to anchor the cluster.", + "best_issue_reason": "Issue 43684 is the strongest standalone issue because it describes a concrete registration gap for a specific model family and auto-class pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different model targets and different capabilities: Qwen3-Omni registration vs. Qwen3.5 AutoModelForSequenceClassification support. They do not describe the same underlying bug or change." + "reason": "Different model families and different auto classes; they do not describe the same underlying bug or change." } ] }, @@ -17178,8 +17298,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43835-4", "nodes": [ @@ -17196,7 +17316,7 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "a3ef88ff0a416ef854b39cbefe9fbc3cdd3c3a1cb400ed56beaaf98918e300a1", + "prepared_review_unit_hash": "eb593e8fb5177a9792ad2fa3e98c968bdac801c3daf61f4157d4e17fc62ce838", "component_budget": { "node_count": 4, "item_count": 4, @@ -17218,58 +17338,55 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues share the exact title \"huggingface \" and look like duplicates; the \"Fett\" issue does not have enough evidence to connect it to that set. With no body detail or links, the safest cluster is the repeated \"huggingface \" reports.", - "confidence": 0.79, - "canonical_issue_reason": "Issue 43845 is the earliest of the three identical \"huggingface \" issues, so it is the best placeholder canonical issue for that duplicate set.", + "summary": "This cluster is mostly low-signal: three issues share the identical vague title \"huggingface \" and one issue is titled \"Fett\". With no bodies, targets, filenames, or references, there isn\u2019t enough evidence to confirm any pair as the same underlying bug, so no soft matches are accepted.", + "confidence": 0.36, + "canonical_issue_reason": "Issue 43845 is the earliest of the repeated \"huggingface \" issues and is the most representative placeholder entry in this cluster; 43835 has a different title and looks less central.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the strongest issue representative because it is the earliest and matches the other two nearly identical reports by title.", + "best_issue_reason": "43845 is the best single issue to anchor the cluster because it is one of the repeated identical titles and appears earliest among them; however, the issue data is too sparse to make a stronger choice.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Titles do not match and there is no body/link evidence showing the same underlying problem." + "reason": "Different titles and no descriptive content; soft similarity alone is not enough to conclude the same underlying issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "No concrete evidence ties \"Fett\" to the \"huggingface\" report cluster." + "reason": "Different titles and no supporting details such as bodies, targets, or references." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Insufficient evidence that this is the same issue; only weak generic similarity is present." + "reason": "Different titles and no evidence they describe the same bug or change." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same exact title, close creation times, and no contrary evidence; likely duplicate reports of the same issue." + "accept": false, + "reason": "Identical vague titles, but there is no body or other evidence showing they refer to the same underlying problem." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same exact title with no distinguishing details, so these look like the same underlying issue." + "accept": false, + "reason": "Same vague title, but insufficient concrete evidence to merge them as duplicates." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Same exact title and no evidence of a different problem, so they are likely duplicates." + "accept": false, + "reason": "Same vague title, but no substantive details to establish a shared underlying issue." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: three issues have the exact same title and 43845 is the earliest of that set. The rejection of 43835 is appropriately cautious given only weak similarity. The accepted soft edges are also conservative because they rely on exact title match with no contrary evidence." - }, - "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 }, @@ -17403,8 +17520,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45145-2", "nodes": [ @@ -17414,12 +17531,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "6099586d5cbff118430468159a04b0d2e1ef0234b4d55fa21a35a338f342656a", + "prepared_review_unit_hash": "5d3e530f4bba004dd2a5e30158177fff5349308e2c8d9feb71cac95f26d6ff93", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -17427,7 +17544,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -17436,28 +17553,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues are likely duplicates: they have identical titles and a soft-similarity signal, with #45145 slightly better suited as the canonical record because it has the only inbound reference.", - "confidence": 0.84, - "canonical_issue_reason": "#45145 is a reasonable canonical issue because it matches #45151\u2019s title exactly and has the only inbound reference, giving it a slight edge for keeping the primary record.", + "summary": "The two issues look highly similar and likely represent the same topic, but the evidence is not strong enough to call them a definitive duplicate. Issue 45145 remains the best canonical candidate because it is earlier and already has an inbound reference.", + "confidence": 0.72, + "canonical_issue_reason": "45145 is the earlier issue and has the only inbound reference, so it is the best canonical record if these turn out to be duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the best issue to keep as the primary record for this likely duplicate pair because it has the same title as #45151 and the only inbound reference.", + "best_issue_reason": "45145 is the strongest issue to keep as the primary record because it predates 45151 and has inbound reference activity.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": true, - "reason": "Identical titles plus the soft-similarity signal make these likely duplicates, though the packet does not include description text, so this should be treated as probable rather than certain." + "accept": false, + "reason": "The titles are identical and the bodies are similar, but the packet provides only soft-similarity evidence with no direct linkage or verbatim confirmation, so treating them as the same report is too strong." } ] }, "evaluator_result": { "accept": false, - "feedback": "The pair is plausibly duplicate based on the identical titles and soft-similarity signal, but the writeup overstates the evidence. The packet does not provide description text to support 'near-matching descriptions,' and the soft-pair verdict is not deterministic. Please keep the conclusion more conservative, e.g. 'likely duplicate' rather than 'clear duplicate pair,' and avoid claiming stronger canonical evidence than the inbound reference supports." + "feedback": "The canonical choice is reasonably grounded (45145 is earlier and has an inbound reference), but the duplicate/soft-edge claim is too strong for the packet. We only have identical titles, similar body lengths, and a soft-similarity signal; there is no direct evidence that the issues are 'verbatim' or definitively the same report. A more conservative verdict would be warranted." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -17522,9 +17639,9 @@ "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", - "reason": null, - "source": "cache", - "cache_hit": true, + "reason": null, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45672-2", "nodes": [ @@ -17534,40 +17651,40 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "0c99eb595c9401d391f8bca202f19a6591d1fd6e84069be760cdd001c7629bf0", + "prepared_review_unit_hash": "19a37480826bcbdc63fe8f18ced2749e9a1be302b5fb5c67addc51ecd45192a4", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both 4-bit inference issues, but they describe different failure modes and likely different fixes: one is a Gemma4/audio torch.finfo() TypeError on uint8 weights, the other is a BitsAndBytesConfig regression where llm_int8_skip_modules removes the default lm_head exclusion and triggers an AssertionError.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue: 45672 is a model/audio-module-specific uint8 torch.finfo() crash, while 45674 is a configuration-handling regression in BitsAndBytesConfig affecting lm_head exclusion. They do not appear to be the same underlying bug.", + "summary": "These are two separate 4-bit inference bugs in different code paths: one in Gemma4 audio modules hitting torch.finfo() on uint8 weights, and one in BitsAndBytesConfig module skipping that breaks the default lm_head exclusion. They should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If one issue must represent the cluster, 45674 is slightly better because it describes a concrete, reusable regression in a core config path; however it still does not cover 45672.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Shared 4-bit inference context only; the titles point to different code paths and symptoms (torch.finfo on uint8 audio weights vs llm_int8_skip_modules clearing lm_head exclusion). Not the same underlying bug." + "reason": "Both mention 4-bit inference, but the failures are different: one is a torch.finfo() TypeError on uint8 weights in Gemma4 audio modules, the other is an AssertionError caused by llm_int8_skip_modules clearing the default lm_head exclusion. Different triggers, code paths, and fixes." } ] }, @@ -17712,8 +17829,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45869-5", "nodes": [ @@ -17733,12 +17850,12 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "66376826eb3acfb4707444bb5c9566babeca1513face2f5fdbfa9f23ed6d8cb9", + "prepared_review_unit_hash": "0f018cdb21c1db38a5160dfabaee3c072f88390111d3c732395fd92343e0308a", "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 }, @@ -17746,7 +17863,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5726, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -17755,66 +17872,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster contains two clear exact-duplicate issue pairs and one thematically related but separate issue. Only the identical-title pairs should be merged; the dataset-sample issue is not the same underlying request.", - "confidence": 0.91, - "canonical_issue_reason": "No single canonical issue cleanly represents the whole cluster because it splits into two different duplicate pairs with different titles/topics.", + "summary": "This cluster splits into two exact duplicate pairs plus one unrelated dataset promo issue: 45870/45871 are the same example issue, 45872/45873 are the same DataForge scraping-service issue, and 45869 is a different financial data bundle request.", + "confidence": 0.96, + "canonical_issue_reason": "45870 is the oldest and cleanest representative of the exact duplicate pair for \"Example: Financial time series prediction with DataForge\"; 45872/45873 form a separate duplicate pair, and 45869 is unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "If one representative must be chosen, issue 45870 is the best fit: it is the earliest member of the exact duplicate pair 45870/45871 and has the most generic framing.", + "best_issue_reason": "45870 is a strong global representative because it is earliest among the duplicate example issues and has an exact-title twin; however, this cluster is really two separate issue groups rather than one unified topic.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different requests: a financial data bundle is not the same underlying issue as a financial time-series prediction example." + "reason": "Different subjects: a financial data bundle promo/request versus a DataForge example issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different requests: dataset/sample availability versus an example/prediction issue." + "reason": "Different subjects: a financial data bundle promo/request versus a DataForge example issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact same title and evidently the same issue; safe duplicate." + "reason": "Exact same title and apparent same underlying example issue; clear duplicate." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Related financial/DataForge theme, but the titles describe different deliverables and code paths (example prediction vs scraping service)." + "reason": "Both mention financial/time-series, but they describe different artifacts: an example issue versus a scraping-service integration." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Related topic, but not the same concrete request; likely separate issues." + "reason": "Different issue concepts: DataForge example versus DataForge scraping-service integration." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different concrete requests despite shared financial/DataForge wording." + "reason": "Different issue concepts despite shared financial/DataForge wording; not the same underlying request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different concrete requests despite shared financial/DataForge wording." + "reason": "Different issue concepts despite shared financial/DataForge wording; not the same underlying request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact same title and scope; clear duplicate pair." + "reason": "Exact same title and same underlying DataForge scraping-service issue; clear duplicate." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative: the identical-title pairs (45870/45871 and 45872/45873) are reasonable duplicate matches, while 45869 is appropriately kept separate. The representative and canonical-issue reasoning also matches the evidence." + "feedback": "The summary is grounded in the packet and stays conservative: the two exact-title pairs are reasonable duplicates, and 45869 is plausibly separate from the DataForge example/service issues. No soft-edge verdict appears overstated." }, "evaluator_used": true, "retried": false, @@ -17825,7 +17942,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17869,12 +17986,12 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "747f11d8a8b7ec2d2248e924215c064275d7217dff7f57b12f9b979cf0a07791", + "prepared_review_unit_hash": "4e7698c4ce846fd7e6f972378e3dd979f4b575c950d6100c8cd833e246be2358", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34776, + "serialized_chars": 34773, "estimated_input_tokens": 8694, "estimated_eval_tokens": 17644 }, @@ -17882,7 +17999,7 @@ "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34776, + "serialized_chars": 34773, "estimated_input_tokens": 8694, "estimated_eval_tokens": 17644 }, @@ -17891,126 +18008,144 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are largely unrelated; the only true duplicate pair is the two `fixed_cross_entropy` fixes for issue #43240. All other soft-similarity pairs touch different models, docs, tests, or subsystems and should stay separate.", - "confidence": 0.93, + "summary": "This cluster is mostly unrelated PRs that share broad themes like model onboarding, tokenizer fixes, or test updates. The only true near-duplicate is the pair of `fixed_cross_entropy` PRs, which both implement the same loss kwargs support for the same issue target.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43251 is the best canonical representative because it directly targets issue #43240, has the clearest bug-fix framing, and fixes the shared `loss_utils.py` code path; #43254 is essentially the same change with a shorter description.", + "canonical_pr_reason": "`pull_request:43251` is the best representative for the loss fix: it has the explicit issue target `43240`, touches the exact same code path as the companion PR, and presents the broader, more direct API change (`weight` and `label_smoothing`) in `fixed_cross_entropy`.", "best_issue_reason": null, - "best_pr_reason": "PR #43251 is the strongest overall PR in the cluster for duplicate-triage purposes: it is a concrete, narrowly scoped code fix with an explicit issue target, unlike the rest of the cluster which is a collection of unrelated PRs.", + "best_pr_reason": "`pull_request:43251` is the strongest global representative because it is the only PR in the cluster with an explicit issue target and it cleanly captures the concrete bug/fix pair around `fixed_cross_entropy`.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated PRs: VideoPrism model/docs addition vs an Executorch dynamic-shape test. Different features and code paths." + "reason": "Both are unrelated model/runtime additions; they share only a loose theme of Transformers feature work, not the same underlying bug or change." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated: Mistral tokenizer helper methods vs Executorch export tests. No shared underlying bug or change." + "reason": "Tokenizer method additions vs. executorch dynamic-shape testing are different changes in different code paths." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Unrelated: tokenizer API additions for Mistral vs adding a new MobileLLM model implementation." + "reason": "A Mistral tokenizer API gap and adding MobileLLM are unrelated model-specific changes." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Unrelated: processor auto-loading robustness vs a new model implementation. Shared broad framework code is not enough to merge." + "reason": "These are both processor/model-pretrained related only at a broad level; they address different models and different implementation problems." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Different changes: a loss-function kwargs fix vs adding MobileLLM model support." + "reason": "Loss-function kwargs support is unrelated to adding a new MobileLLM model implementation." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Different code paths: cross-entropy kwargs handling vs Mistral4 test/runtime fixes." + "reason": "The loss utility fix and Mistral4 test/model work do not target the same bug or code path." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Different subsystems: Executorch dynamic-shape export tests vs Mistral4 fixes." + "reason": "Executorch dynamic-shape export tests and Mistral4 test fixes are separate test concerns, not duplicate fixes." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Different changes: Executorch export test coverage vs new MobileLLM model support." + "reason": "They cover different subsystems and different failure modes; no shared concrete change." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "Different changes: adding a new model vs fixing config resolution order in core modeling utilities." + "reason": "Adding MobileLLM and fixing `cls.config` resolution are unrelated model vs. core inheritance behavior changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "Different changes: new MobileLLM model code vs Falcon output-interface refactor." + "reason": "A new model implementation and Falcon output-interface refactoring are different tasks." }, { - "left": "pull_request:43995", + "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "Different changes: Falcon modeling refactor vs a model-creation skill and unrelated scaffolding." + "reason": "MobileLLM addition and a model-creation skill/meta-PR are not the same underlying code change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "These are unrelated; one adds a model, the other changes class annotation/config resolution semantics." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Different model families and fixes; no same concrete bug or mergeable single change." + "reason": "Falcon output-collection refactor and Mistral4 fixes are different model-family changes with no shared concrete bug." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactoring is unrelated to the model-creation skill/meta automation PR." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Different changes: Mistral4 test/fix work vs a CLIP-like weight-renaming typo fix." + "reason": "Mistral4 test/model fixes and the `layrnorm` rename sweep are different bugs/changes, even if both touch model code." }, { "left": "pull_request:44827", - "right": "pull_request:46048", + "right": "pull_request:45213", "accept": false, - "reason": "Different changes: Mistral4 fixes vs adding a TDT loss kernel integration." + "reason": "A model-specific fix PR and a model-creation skill PR are not duplicate fixes." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Different code paths: model-specific test fixes vs core config-resolution ordering." + "reason": "Mistral4 fixes and `cls.config` resolution order are unrelated internal behaviors." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 fixes and adding the TDT loss kernel are different subsystems and different concrete changes." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Different changes: Mistral4 fixes vs a draft model-creation skill/scaffolding PR." + "reason": "The shared theme is only model development infrastructure; the concrete edits are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "Different changes: MobileLLM model addition vs model-creation skill infrastructure." + "reason": "A new model PR and a meta skill PR are not the same underlying change." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Same underlying bug fix in `loss_utils.py`: both add passthrough support for `cross_entropy` kwargs on the `fixed_cross_entropy` helper for issue #43240. They could plausibly be merged into one PR." + "reason": "Both PRs modify `src/transformers/loss/loss_utils.py` to make `fixed_cross_entropy` pass supported kwargs to `nn.functional.cross_entropy`; they target the same bug and could plausibly be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analyst output is well grounded in the packet. The only accepted duplicate pair, PR #43251 and PR #43254, is clearly mergeable into a single bug-fix PR: both modify `src/transformers/loss/loss_utils.py` to pass through `cross_entropy` kwargs for the same issue target (#43240). The other pairwise rejections are conservative and consistent with the diff previews, which show different models, subsystems, or scaffolding rather than a shared concrete bugfix. The summary does not overstate evidence." + "feedback": "The analysis is well grounded in the packet. The only accepted soft-edge pair, `pull_request:43251` vs `pull_request:43254`, is clearly mergeable as a single bugfix PR: both touch `src/transformers/loss/loss_utils.py`, both target issue `43240`, and both extend `fixed_cross_entropy` to pass supported kwargs to `nn.functional.cross_entropy`. The other pair verdicts are conservative and do not overclaim overlap beyond broad thematic similarity. The summary also stays within the evidence." }, "evaluator_used": true, "retried": false, @@ -18021,7 +18156,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18037,12 +18172,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "b431cd65b28f1c3230bf93126bc9b7911b8592a818fd7d0b806c50c854eabdc7", + "prepared_review_unit_hash": "deaa24b9e3f43f9b535319b7e8a7557041c8ca66d01759c6256e788b950f5afa", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7226, + "serialized_chars": 7225, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -18050,7 +18185,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7226, + "serialized_chars": 7225, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -18059,18 +18194,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related VibeVoice PRs, but they are not the same change: #40546 implements the base VibeVoice model, while #42781 adds a realtime variant plus an acoustic tokenizer and generation changes. They should not be deduplicated.", - "confidence": 0.91, + "summary": "Two related but distinct pull requests for the VibeVoice family: one adds the base model support, the other adds a realtime variant with additional tokenizer/feature-extractor/generation work. They overlap heavily in shared scaffolding, but they are not the same concrete change set.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "#40546 is the broader, earlier base VibeVoice implementation and is the more canonical representative of the core model integration.", + "canonical_pr_reason": "PR 42781 is the broader and more feature-complete VibeVoice-related change, but it is still a different deliverable from the base VibeVoice implementation in 40546 (realtime extension plus extra model components).", "best_issue_reason": null, - "best_pr_reason": "#40546 is the best single PR for the underlying VibeVoice addition because it focuses on the main model implementation without the extra realtime/tokenizer scope.", + "best_pr_reason": "PR 42781 is the best representative of the cluster because it covers the larger, newer VibeVoice integration surface and includes the realtime-specific additions beyond the base model support.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Different concrete changes: base VibeVoice integration vs a realtime extension with additional tokenizer/generation code. They share filenames because the second builds on the first, but they are not mergeable as one duplicate PR." + "reason": "They share the VibeVoice codebase and docs, but 40546 is the base implementation while 42781 adds a realtime variant with extra tokenizer/feature-extractor/generation changes. Related, but not the same underlying fix/change." } ] }, @@ -18084,7 +18219,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18100,12 +18235,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "c483fd01e4a132e4aef468e53b4911673ae937cd019768b62a0712cc76c65e8f", + "prepared_review_unit_hash": "9226e9a4c51f760b804ac0222b00df236dc24483bf11e45dd1b816c348815219", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4676, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18113,7 +18248,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4676, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18122,18 +18257,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are unrelated feature additions: TimesFM covariate forecasting vs. DINOv3 image classification support. They should not be clustered as duplicates.", - "confidence": 0.99, + "summary": "These are two unrelated feature PRs in different model areas: TimesFM covariate forecasting vs. DINOv3 image-classification support. They do not look like the same underlying change and should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR for this cluster: the PRs target different models and different code paths, so there is no duplicate relationship to normalize around.", + "canonical_pr_reason": "No single canonical PR fits: the items address different models, code paths, and user-facing features.", "best_issue_reason": null, - "best_pr_reason": "PR 40755 is the more substantial standalone change, but it is not a duplicate representative because the two PRs are unrelated.", + "best_pr_reason": "Neither PR is a good global representative for the cluster because they are not duplicates and solve unrelated problems.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models, different functionality, and different affected files; they do not appear to fix or add the same underlying code-path change." + "reason": "Different subsystems and goals: TimesFM forecasting with covariates vs. DINOv3 image classification. The filenames, APIs, and diffs indicate separate feature work, not one mergeable fix." } ] }, @@ -18147,7 +18282,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18210,7 +18345,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18228,7 +18363,7 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "0e01b8f84b0f791b9f495a575ace0344cd270a783499979faa346b9775b90150", + "prepared_review_unit_hash": "d9ad6072d6668006e9497862c02483ce96fda9a263dd89dc7644d0c1f62c487b", "component_budget": { "node_count": 3, "item_count": 3, @@ -18250,24 +18385,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three separate model-addition PRs with shared Transformers boilerplate, but each introduces a different model family (DeepSeek OCR, Isaac, Kimi2-6). They are superficially similar in structure yet not the same underlying change.", + "summary": "All three PRs are same-pattern model additions, but they target different models and code paths (DeepSeek OCR, Isaac, Kimi2-6), so they are not duplicates. Shared auto/docs plumbing is too generic to merge them together.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45186 (Isaac) is the most central of the three by cluster position: it shares boilerplate with both others and is a straightforward standalone model integration PR, making it a reasonable representative if one must be chosen.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45630 (Kimi2-6) is the strongest standalone candidate overall: it has the broadest model integration footprint and the highest review activity, suggesting the most complete and active implementation among the three.", + "best_pr_reason": "PR 45630 is the broadest and most complete single model-addition PR in this cluster, but it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Both are model-addition PRs and share auto-mapping/docs scaffolding, but they implement different models (DeepSeek OCR vs Isaac) with distinct code paths and files; not mergeable as one fix/change." + "reason": "Different models and implementations (DeepSeek OCR vs Isaac); shared docs/auto-registration changes are generic boilerplate, not the same underlying fix." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "These add different models (Isaac vs Kimi2-6). The overlap is only generic Transformers boilerplate and registration files, not the same concrete bug or model implementation." + "reason": "Different model families and code paths (Isaac vs Kimi2-6); they overlap only in standard integration files, so they should not be merged as one PR." } ] }, @@ -18281,7 +18416,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18297,12 +18432,12 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "8888027775f67853322e3ea3837b42badfcafd98d7ea802919d6aebcf7b95c52", + "prepared_review_unit_hash": "59615d1a5f05b4993bb991574c894c769c2d4cebcd871eee69ad460bc054b026", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5934, + "serialized_chars": 5935, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -18310,7 +18445,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5934, + "serialized_chars": 5935, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -18319,18 +18454,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related GitHub Actions maintenance PRs, but they are not the same change. One targets Node 24 compatibility across many workflows, while the other is a separate bump of workflow action versions (notably docker/build-push-action) and broader latest-version updates. The soft duplicate edge should be rejected.", - "confidence": 0.88, + "summary": "Two distinct workflow-maintenance PRs: one is a broad GitHub Actions upgrade for Node 24 compatibility, the other updates Docker build-related actions and a few workflow action versions. They overlap in CI files but do not appear to fix the same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better canonical PR because it is the broader, more central workflow upgrade focused on Node 24 compatibility across many GitHub Actions workflows.", + "canonical_pr_reason": "PR 42975 is the cleaner representative of the cluster because it has a single, explicit goal (Node 24 compatibility) across many workflows; PR 42976 is a separate action-version refresh centered on Docker/build workflows.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best representative of the cluster: it covers the main cross-workflow GitHub Actions compatibility upgrade, whereas 42976 is a separate dependency bump with different action targets.", + "best_pr_reason": "PR 42975 is the best single PR for this cluster because its scope is broader and more clearly defined, while PR 42976 is narrower and targets a different dependency set.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both touch workflow dependencies, but they address different changesets: Node 24 compatibility vs. action version bumps (including docker/build-push-action). Shared filenames alone are not enough to treat them as the same underlying fix." + "reason": "Reject: both touch GitHub Actions workflow files, but 42975 upgrades core action versions for Node 24 compatibility, while 42976 specifically bumps docker/build-push-action and related workflow actions. They are related maintenance changes, not the same concrete fix and would not naturally merge into one PR." } ] }, @@ -18344,7 +18479,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18369,76 +18504,76 @@ "pull_request:43451|pull_request:45490", "pull_request:45490|pull_request:46162" ], - "prepared_review_unit_hash": "cd00e5b96d6c5d4dda0f1b262b4c5098d25e43f151b5471813f8a75169fc375f", + "prepared_review_unit_hash": "d6d280af2cb9dbaee7905e6a902fe1815007c3adf05ef299dd5f488e1bd0a02f", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 7, - "serialized_chars": 16553, - "estimated_input_tokens": 4139, - "estimated_eval_tokens": 8534 + "serialized_chars": 16552, + "estimated_input_tokens": 4138, + "estimated_eval_tokens": 8532 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 7, - "serialized_chars": 16553, - "estimated_input_tokens": 4139, - "estimated_eval_tokens": 8534 + "serialized_chars": 16552, + "estimated_input_tokens": 4138, + "estimated_eval_tokens": 8532 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are superficially similar because they all touch Transformers boilerplate, but they are different model additions or a test-only tweak. No duplicate cluster emerges.", - "confidence": 0.97, + "summary": "No duplicate cluster here: the PRs are separate model additions (ViT NEPA, Molmo, Molmo2, CTSM) plus an unrelated XPU test-only tweak. The overlapping docs/auto-registration/test scaffolding files are generic and don\u2019t indicate the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the artifacts span distinct changes (ViT NEPA, Molmo, Molmo2, CTSM, and a test-only XPU annotation), with shared files reflecting integration boilerplate rather than one underlying fix.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "43451 is the most substantial standalone representative PR (Molmo2) with the broadest concrete model integration, but it is still not a duplicate of the others.", + "best_pr_reason": "No PR is a good canonical duplicate candidate because these are distinct model-addition or test-only changes. If forced to pick a representative, 43451 is the most substantial single PR, but it is still not duplicate-equivalent to the others.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions (ViT NEPA vs Molmo); shared auto/docs files are expected boilerplate, not the same change." + "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto files are standard registration plumbing, not the same bug/change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "ViT NEPA vs Molmo2 are unrelated model introductions with different codepaths and files." + "reason": "Different model additions: ViT NEPA vs Molmo2. The overlap is only generic model-registration/docs scaffolding." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model families and implementation targets; only generic Transformers registration/docs overlap." + "reason": "Different domains and models: ViT NEPA vs CTSM. Shared auto/docs edits are broad infrastructure, not one concrete fix." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are separate model implementations, not one concrete bug fix or identical feature change." + "reason": "Molmo and Molmo2 are separate model implementations with different files and behavior; they are not the same change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM are unrelated additions; overlap is limited to common registration/documentation files." + "reason": "Molmo vs CTSM are unrelated model additions; the common auto/docs files are incidental." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are different model integrations with no shared underlying bug or patch." + "reason": "Molmo2 and CTSM are distinct model additions with different code paths and artifacts; not mergeable as one duplicate PR." }, { "left": "pull_request:45490", "right": "pull_request:46162", "accept": false, - "reason": "CTSM model addition vs a test-only XPU expectation annotation; these are clearly different changes." + "reason": "CTSM model implementation vs a small XPU test-annotation tweak across existing tests; these are unrelated changes." } ] }, @@ -18452,7 +18587,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18471,7 +18606,7 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "15a636e8993525111d017f96434b2dfd34cb76f72e502b60a6d07dc8d80a560c", + "prepared_review_unit_hash": "ccc2e9ca3ff8c58e2bf3a436ba44baa03e33736283049f1c66d4f09479856ad0", "component_budget": { "node_count": 3, "item_count": 3, @@ -18493,30 +18628,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three pull requests appear unrelated as underlying changes. PR 44815 is the only substantive code fix; PR 43291 is a Whisper tokenizer/test fix; PR 43665 is test-file noise in CLIP/ViT and does not match the others. None of the soft edges should be merged.", + "summary": "These are three unrelated pull requests: one adjusts Whisper tokenization/tests, one is test-only noise for CLIP/ViT, and one is a substantive FP8/dequantization fix. The superficial similarity is just that they are all PRs touching tests or model code, but they do not target the same bug or change.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44815 is the strongest canonical PR because it contains a real multi-file implementation fix around dequantization/loading behavior, unlike the other two which are narrow test/tokenizer changes or test-only noise.", + "canonical_pr_reason": "PR #44815 is the strongest canonical PR because it contains the most concrete production fix: a dequant/loading-path change plus FP8 integration work across core code and a related model test. The other PRs are either test-only noise (#43665) or a narrower Whisper test/tokenization adjustment (#43291).", "best_issue_reason": null, - "best_pr_reason": "PR 44815 is the best representative of the cluster due to its concrete code-path fix and broader functional impact.", + "best_pr_reason": "PR #44815 is the best representative of the cluster because it is the most substantive and self-contained code change. It fixes a specific code path, unlike #43665 which is clearly not a real fix, and #43291 which is a narrower Whisper-specific test/tokenization tweak.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer decode logic vs. CLIP/ViT test modifications are different changes with no shared code path or bug." + "reason": "Whisper tokenization/test changes and dummy CLIP/ViT test edits are unrelated fixes; they do not address the same underlying bug or code path." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decoding/timestamp behavior is unrelated to FP8 dequantization and core model loading changes." + "reason": "Whisper decode/tokenization cleanup is unrelated to FP8 dequant/loading behavior. Different subsystems, different failure modes, no shared concrete fix." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Test-only edits in CLIP/ViT do not match the dequantization/loading fix in 44815 and are not the same underlying change." + "reason": "#43665 is test-only noise while #44815 is a real FP8/dequant fix. They do not plausibly merge into one PR for the same issue." } ] }, @@ -18530,7 +18665,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18553,73 +18688,73 @@ "pull_request:45932|pull_request:46100", "pull_request:46100|pull_request:46114" ], - "prepared_review_unit_hash": "309a3a956dda8cefff6e0f45e3e493ac5a1fddc4516b2c06fcc4cdf511f13fa2", + "prepared_review_unit_hash": "79d86350bcea9c8200a743a7c0dee81b871fe482fdd73260a00b54dc7aa204e2", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 5, - "serialized_chars": 11340, - "estimated_input_tokens": 2835, - "estimated_eval_tokens": 5926 + "serialized_chars": 11342, + "estimated_input_tokens": 2836, + "estimated_eval_tokens": 5928 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 5, - "serialized_chars": 11340, - "estimated_input_tokens": 2835, - "estimated_eval_tokens": 5926 + "serialized_chars": 11342, + "estimated_input_tokens": 2836, + "estimated_eval_tokens": 5928 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Only one true duplicate pair is present: PR #46100 and PR #46114, which have the same ViT test-file diff. The other soft links are false positives from shared test/workflow overlap or broader unrelated changes.", - "confidence": 0.93, + "summary": "Most PRs are unrelated repo-test / bot-playground changes that happen to touch the same test file or subsystem. Only the two identical ViT test-migration PRs look like true duplicates of the same change.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46100 is the clearest canonical representative of the duplicated ViT test change; PR #46114 carries the same diff but has a fork/test-specific title.", + "canonical_pr_reason": "PR #46100 is the cleaner canonical representative: it is earlier, has the generic test-migration title, and its diff matches #46114 exactly.", "best_issue_reason": null, - "best_pr_reason": "PR #46100 is the best overall PR candidate because it is the cleanest standalone instance of the identical ViT test-migration diff.", + "best_pr_reason": "PR #46100 is the best overall representative of the duplicate pair because it is the earliest and most neutral description of the identical change.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:46114", "accept": false, - "reason": "Both touch the ViT test file, but the actual changes are unrelated: setup/version and dependency tweaks versus a test-migration diff." + "reason": "Both touch tests/models/vit/test_modeling_vit.py, but the surrounding changes differ completely (#43488 also changes setup.py/dependency plumbing). Same file is not enough; not the same underlying PR." }, { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Shared Vit test file overlap is incidental; one is a repo-bot/setup check PR and the other is workflow/CI plumbing." + "reason": "Different edits and different intent: #43488 is a repo-bot/version/dependency tweak, while #43532 rewrites CI workflows and multiple model files. Not a duplicate change." }, { "left": "pull_request:43532", "right": "pull_request:46114", "accept": false, - "reason": "They both edit the same test file, but the surrounding changes are different and do not reflect the same underlying bug or fix." + "reason": "They only overlap on the ViT test file, but #43532 is a broad CI/workflow/model change and #46114 is a tiny test-migration stub. Not the same concrete change." }, { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "The packet does not support treating these as the same ViT test-file change; the overlap is too indirect and the filenames/diff context do not match." + "reason": "#45932 is a large multi-file grafana/OTEL and model-code change; #46100 is a tiny ViT test-migration edit. Shared test-file noise does not make them duplicates." }, { "left": "pull_request:46100", "right": "pull_request:46114", "accept": true, - "reason": "These are the same concrete diff in tests/models/vit/test_modeling_vit.py, adding the same failing assertions." + "reason": "These PRs have the same diff in the same file and appear to be forked/duplicate test-migration PRs. They are the same underlying change and could be merged as one." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Reject: the write-up overstates the evidence for the 45932\u219446100 pair. In the packet, 45932\u2019s listed filenames do not include the same ViT test file as 46100, so describing them as the \"same inserted failing ViT tests in the same file\" is not grounded. The summary also undercounts the relationship among the three ViT test PRs, since 46100 is linked to both 45932 and 46114 in the soft pairs. The other rejections look conservative, but the accepted duplicate claim is too strong for the packet." + "accept": true, + "feedback": "The assessment is grounded in the packet. The only accepted pair, 46100 and 46114, has the same single-file diff and is plausibly a forked duplicate test-migration PR, so it is mergeable as one bugfix. The rejected pairs are conservative: they share at most a filename or subsystem, but the surrounding changes are clearly unrelated in scope and intent." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -18627,7 +18762,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18643,7 +18778,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "822e2ab8d01d1b75363667e00cd24a3cdf413ae756f388265c997464a9b48f90", + "prepared_review_unit_hash": "a1f3f6d87e081d3bfebf3c418cb5637905430851d0df38d52789f18809979709", "component_budget": { "node_count": 2, "item_count": 2, @@ -18665,18 +18800,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate feature PRs for different segmentation pipelines. They share documentation/registration scaffolding, but they implement different tasks, different pipeline modules, and different model/test coverage, so they should not be clustered as duplicates.", - "confidence": 0.96, + "summary": "The two pull requests are closely related feature additions, but they target different segmentation pipelines and different model/code paths, so they should not be treated as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR as a duplicate target: PR 43612 adds Promptable Concept Segmentation, while PR 43613 adds Promptable Visual Segmentation with a different set of supported models and pipeline implementation.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR 43613 is the broader, more infrastructure-heavy pipeline addition, but it still represents a distinct feature rather than the same change as PR 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Both are new pipeline features, but for different concrete tasks: Promptable Concept Segmentation vs Promptable Visual Segmentation. Shared docs/auto-registration files are expected infrastructure overlap, not evidence of the same underlying code-path change, and the model/test files differ substantially." + "reason": "They share documentation/pipeline scaffolding, but the titles, filenames, and supported model sets show different underlying tasks and code paths, so they are not the same concrete change." } ] }, @@ -18690,7 +18825,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18706,7 +18841,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "172b0a67b9a9e8fa3c22fb05fd495d75dd3ec1c89dae6d482dc3508c84fc8fc6", + "prepared_review_unit_hash": "c36e7d070e991b31485a82fce4e83a47dd8d89d45d8dc4f49b6044c15820f24d", "component_budget": { "node_count": 2, "item_count": 2, @@ -18728,18 +18863,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two open PRs both touch `trainer.py`, but they address different Trainer extensibility points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering. They do not look like the same underlying change.", - "confidence": 0.97, + "summary": "The two PRs both touch `trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for overrideability. They are not the same underlying change and should not be merged as duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete bugfix-style change: it alters training loss scaling behavior for gradient accumulation. PR #43663 is a separate refactor around input column filtering and model signature inspection.", "best_issue_reason": null, - "best_pr_reason": "PR #43651 is the stronger representative of the cluster because it makes a concrete training-behavior change in `training_step` and adds a new control hook for loss scaling.", + "best_pr_reason": "PR #43651 is the best representative of the cluster because it targets a specific behavioral issue in the training loop, whereas #43663 is an unrelated extensibility refactor in the same file.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both modify `trainer.py`, but one is about gradient-accumulation loss normalization and the other is about signature-column inspection for filtering. Different code paths, different behavior, not the same fix." + "reason": "Different code paths and different goals: #43651 changes gradient-accumulation loss normalization in `training_step`, while #43663 adds an overrideable helper for signature-based column filtering. Shared file alone is not enough to treat them as the same fix." } ] }, @@ -18753,7 +18888,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18769,12 +18904,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "ef41ead1d054a3d6e71ba9404933d199f45f984eccfe6e3880a6a486dafcec10", + "prepared_review_unit_hash": "eb3175c3196b0d1fd677d3231e1e58a7f84be221c9ef1390ef3edd74f54d0cca", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -18782,7 +18917,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -18791,18 +18926,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs touch the same CLI file and both relate to TypeAdapter typing/import issues, but they fix different failure modes and should not be merged as duplicates.", - "confidence": 0.96, + "summary": "These PRs overlap on the same CLI file and the same TypeAdapter NameError theme, but they appear to fix different runtime paths with different techniques, so they should not be treated as duplicates.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the two PRs target different concrete bugs in the same module rather than the same change.", + "canonical_pr_reason": "PR 43842 is the more precise fix for the explicit TypeAdapter NameError when pydantic is unavailable, with a minimal targeted change in the affected code path.", "best_issue_reason": null, - "best_pr_reason": "PR #43842 is the more narrowly scoped, explicit fix for the pydantic-optional TypeAdapter NameError and has the clearest issue linkage.", + "best_pr_reason": "PR 43842 best captures the concrete bug report and applies the narrowest fix directly to the failing TypeAdapter annotation path.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Both edit src/transformers/cli/serve.py and involve TypeAdapter typing, but #43656 fixes annotation forward-reference/NameError behavior via __future__ annotations, while #43842 fixes missing pydantic TypeAdapter import when pydantic is absent. Different root causes and not mergeable as one PR." + "reason": "Both mention TypeAdapter NameError in src/transformers/cli/serve.py, but 43656 is a broader annotations/typing cleanup via __future__.annotations, while 43842 is a targeted fix for the pydantic-not-installed case. Different root causes and different fixes, so not a duplicate PR pair." } ] }, @@ -18816,7 +18951,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18882,7 +19017,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -18931,8 +19066,8 @@ "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:44066|pull_request:44068", "pull_request:44019|pull_request:44722", "pull_request:44086|pull_request:44722", "pull_request:44019|pull_request:44085", @@ -18961,9 +19096,9 @@ "pull_request:44073|pull_request:44101", "pull_request:44059|pull_request:44071", "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44068", "pull_request:44073|pull_request:44161", "pull_request:44056|pull_request:44101", - "pull_request:44059|pull_request:44068", "pull_request:44101|pull_request:44161", "pull_request:44017|pull_request:44018", "pull_request:43997|pull_request:44056", @@ -18994,268 +19129,316 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "f6bf089ff23b37144bb6a63340cba77bfea824f30affc2950e72b967781bb39a", + "prepared_review_unit_hash": "dfc36ac762f1cefe3d92f14dda99c02166e3ed523b56b3588bc8f8a8c2d91376", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84892, - "estimated_input_tokens": 21223, - "estimated_eval_tokens": 42702 + "serialized_chars": 84901, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75089, - "estimated_input_tokens": 18773, - "estimated_eval_tokens": 37802 + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a series of model-specific refactors from legacy output handling to `capture_outputs`/`can_return_tuple`. They are thematically related, but most touch different model code paths and should not be treated as duplicates; only same-model variants look mergeable.", - "confidence": 0.91, + "summary": "This cluster is mostly a set of model-specific output-tracing refactors under the same tracking issue. The only clear duplicates are the GPT-Neo pair and the GPT-J/CodeGen variants; the rest target different models and should stay separate.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44066 is the best representative: it explicitly targets #43979 and shows a concrete, substantive output-tracing migration on the GPT-J/CodeGen path.", + "canonical_pr_reason": "PR 44066 is the strongest canonical PR: it is explicitly tied to the tracking issue and covers the core GPT-J/CodeGen output-tracing refactor in the shared code path.", "best_issue_reason": null, - "best_pr_reason": "PR 44066 is the strongest canonical PR for the cluster because it is explicit about the tracking issue and reflects the core standardized output-tracing refactor most clearly.", + "best_pr_reason": "PR 44066 best represents the underlying change because it spans the shared GPT-J/CodeGen implementation and is directly linked to the issue driving this refactor wave.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; same refactor theme is not enough to treat them as duplicates." + "reason": "Different model families and code paths; same tracking issue alone is not enough to merge them." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "FNet/CVT vs DeBERTa are separate implementations; these are parallel refactors, not the same change." + "reason": "These refactor different models (CVT/FNet vs DeBERTa v2), so they are not the same change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same underlying model path and could plausibly be merged into one PR." + "reason": "Both touch the GPT-J output-tracing refactor in the same file/code path, so they look like duplicate variants of one change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "Different CNN families and files; they share a pattern but not the same bug or change." + "reason": "RegNet/ResNet/RT-DETR and EfficientNet are separate model implementations, not one duplicate bug fix." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths; same tracking theme only." + "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths; only the generic refactor theme overlaps." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen and EfficientNet are unrelated concrete implementations." + "reason": "GPT-J/CodeGen and EfficientNet are different model code paths, so this is not a duplicate." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are different subsystems with separate output-tracing rewrites." + "reason": "MobileNetV2 and DeBERTa v2 are unrelated implementations; same meta-pattern only." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen versus MGP-STR; same umbrella refactor, but not the same code path." + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are different model changes." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are different model implementations." + "reason": "GPT-J/CodeGen and MPT are separate model code paths, so this is not the same underlying change." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both target GPT-Neo output tracing on the same model file and appear to be overlapping revisions of the same change." + "reason": "Same GPT-Neo output-tracing refactor in the same model/file; 44068 is a later revision of the same change." }, { - "left": "pull_request:44066", - "right": "pull_request:44068", + "left": "pull_request:44018", + "right": "pull_request:44066", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are distinct model paths." + "reason": "GPT-Neo and GPT-J/CodeGen are different implementations, so they should not be merged." }, { - "left": "pull_request:44018", - "right": "pull_request:44066", + "left": "pull_request:44066", + "right": "pull_request:44068", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different models; only the refactor style matches." + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths, despite both being output-tracing refactors." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated concrete changes." + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated changes." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are separate model paths." + "reason": "MGP-STR and GPT-J/CodeGen touch different code paths and do not look like the same bug." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are unrelated model families." + "reason": "ResNet and GPT-J are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet do not share the same underlying code change." + "reason": "GPT-Neo and ResNet are separate model-specific refactors, not duplicates." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are separate implementations." + "reason": "ResNet and MPT are different model code paths; only the refactor style overlaps." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are different model code paths." + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are unrelated refactors." + "reason": "GPT-Neo and Flaubert/XLM are distinct code paths, so this is not a duplicate." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "These are both GPT-family refactors, but they target different models and files." + "reason": "GPT-Neo and GPT-J/CodeGen are different model refactors." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are separate implementations." + "reason": "MPT and GPT-J/CodeGen are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different model paths." + "reason": "GPT-Neo and GPT-J are different code paths; same broad output-tracing theme is not enough." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated code paths." + "reason": "GPT-Neo and MGP-STR are unrelated model changes." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are different model implementations." + "reason": "GPT-Neo and MGP-STR target different models and different forward paths." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 and MGP-STR are separate model-specific changes." + "reason": "GPT-2 and MGP-STR are separate model-specific refactors." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are unrelated code paths." + "reason": "GPT-J and MGP-STR are unrelated model code paths." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and Flaubert/XLM are different model families." + "reason": "ViLT and Flaubert/XLM are different models; they are only thematically similar." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT-2 and GPT-J/CodeGen are distinct model implementations." + "reason": "GPT-2 and GPT-J/CodeGen are different model code paths." }, { - "left": "pull_request:44018", + "left": "pull_request:44059", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are unrelated concrete changes." + "reason": "GPT-2 and MPT are unrelated model implementations." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are separate code paths." + "reason": "SegFormer and GPT-J/CodeGen are separate model refactors." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are different models; same refactor family only." + "reason": "SegFormer and LongT5 are different model code paths." }, { "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "SegFormer and VisualBert do not share the same underlying fix." + "reason": "SegFormer and VisualBert are unrelated model implementations." }, { "left": "pull_request:44068", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are different model implementations." + "reason": "GPT-Neo and MPT are different model code paths." }, { "left": "pull_request:44071", "right": "pull_request:44086", "accept": false, - "reason": "MPT and MGP-STR are unrelated code paths." + "reason": "MPT and MGP-STR are unrelated models." }, { "left": "pull_request:43997", "right": "pull_request:44073", "accept": false, - "reason": "RegNet and VisualBert are different subsystems." + "reason": "RegNet and VisualBert are different model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are separate model-specific changes." }, { "left": "pull_request:44073", "right": "pull_request:44101", "accept": false, - "reason": "VisualBert and Flaubert/XLM are separate model code paths." + "reason": "VisualBert and Flaubert/XLM do not share the same concrete code path." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are different models, so this is not a duplicate." }, { "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are unrelated concrete changes." + "reason": "MPNet and LongT5 are unrelated model code paths." }, { "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are separate model-specific refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44059", + "accept": false, + "reason": "GPT-Neo and GPT-2 are different models and forward paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-2 and GPT-Neo are unrelated model code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are distinct model implementations." + }, + { + "left": "pull_request:44017", "right": "pull_request:44018", "accept": false, - "reason": "RegNet and GPT-Neo are different models." + "reason": "SegFormer and GPT-Neo are different model code paths." }, { - "left": "pull_request:44056", + "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "MPNet and SegFormer are unrelated implementations." + "reason": "CVT/FNet and SegFormer are separate model-specific changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are unrelated model implementations." }, { "left": "pull_request:44010", @@ -19264,64 +19447,76 @@ "reason": "SqueezeBert and MPT are different code paths." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are separate model refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", "accept": false, - "reason": "TextNet and Flaubert/XLM are unrelated model-specific refactors." + "reason": "RegNet and TextNet are separate model-specific changes." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert and DeBERTa are distinct implementations." + "reason": "SqueezeBert and DeBERTa v2 are unrelated model code paths." }, { "left": "pull_request:44044", - "right": "pull_request:44074", + "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and TextNet are different model paths." + "reason": "DeBERTa v2 and SpeechT5 are different model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing migration on overlapping files; these look like duplicate or overlapping revisions." + "reason": "They modify the same GPT-J/CodeGen output-tracing code paths and look like duplicate iterations of the same refactor." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same underlying model path." + "reason": "Both touch the GPT-J output-tracing implementation in modeling_gptj.py, so they look like the same change." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated changes despite similar refactor language." + "reason": "GPT-J and SpeechT5 are unrelated model code paths." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "FNet/CVT and RegNet/ResNet are separate model families with different concrete edits." + "reason": "CVT/FNet and RegNet/ResNet/RT-DETR are different model-specific refactors." }, { - "left": "pull_request:44044", + "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are unrelated model implementations." + "reason": "CVT/FNet and SpeechT5 are unrelated changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and VITS touch different code paths and different output semantics." + "reason": "DeBERTa v2 and ViTS are different model code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: it distinguishes shared refactor theme from true overlap, and the accepted soft-edge pairs are limited to same-model/same-file GPT-J or GPT-Neo variants that plausibly belong in one merged bugfix PR. No overstatement stands out." + "feedback": "The summary stays close to the packet: the canonical PR choice is supported by the shared GPT-J/CodeGen file scope and explicit issue link, and the accept/reject calls are conservative for the cross-model pairs. No pair is overstated as mergeable beyond what the diffs and titles support." }, "evaluator_used": true, "retried": false, @@ -19332,7 +19527,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 28, + "component_count": 29, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -19359,70 +19554,70 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "a9f3530067de162ccb201beed8f578720edf7a41e19a0dbaf96d185a798d6991", + "prepared_review_unit_hash": "ddea8061b1c372284d2f4bc2ffc7c51f71c0076a5eb03e35272260ef41217949", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84892, - "estimated_input_tokens": 21223, - "estimated_eval_tokens": 42702 + "serialized_chars": 84901, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17124, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are thematically similar standardized-output/refactor changes, but they target different model implementations and code paths, so they are not duplicates and should not be merged together as one change.", - "confidence": 0.95, + "summary": "These are model-specific refactors to standardized output tracing/capture_outputs across different transformer architectures. They share the same broad maintenance theme, but each PR targets a different model code path, so they should not be deduplicated together.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical duplicate PR: each item is a separate model-specific output-tracing refactor.", + "canonical_pr_reason": "PR 44129 is the most representative standalone candidate because it has the clearest end-to-end refactor shape and an explicit issue target, but it is still model-specific rather than a cluster-wide duplicate.", "best_issue_reason": null, - "best_pr_reason": "44129 is the best representative because it is narrowly scoped, has an explicit linked issue, and clearly implements a SpeechT5-specific output-capture refactor.", + "best_pr_reason": "PR 44129 is the best single PR to anchor the cluster because it includes an explicit linked issue and a concrete output-tracing refactor, making it the strongest exemplar of this batch.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Different models and code paths (FocalNet vs ImageGPT); same refactor theme, but not the same bug/change." + "reason": "Different model implementations (FocalNet vs ImageGPT). They follow the same refactor pattern, but they do not fix the same concrete code-path problem and would not plausibly merge into one PR." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet have separate model implementations; these are parallel refactors, not one concrete shared fix." + "reason": "VisualBert and TextNet are separate model code paths; this is the same kind of standardized-output migration, not the same underlying bug or change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet vs VisualBert: same output-tracing theme, but distinct model-specific behavior and files." + "reason": "MPNet and VisualBert are unrelated model internals. Shared output-capture refactoring alone is too broad to treat as a duplicate." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "MPNet and TextNet are unrelated code paths; similarity is only at the refactor pattern level." + "reason": "These are different model-specific migrations, not one shared fix to a single code path." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are different architectures with different forward/output mechanics, so these are not duplicates." + "reason": "RegNet and SqueezeBert are separate implementations; the PRs are analogous refactors, but not the same bug/change." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 touch different model families and different output-capture implementations; not the same underlying change." + "reason": "SqueezeBert and SpeechT5 target different forward paths and output signatures, so they should remain separate PRs." } ] }, @@ -19436,7 +19631,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19444,7 +19639,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-43998-14", + "cluster_id": "cluster-43998-15", "nodes": [ "pull_request:43998", "pull_request:43999", @@ -19459,7 +19654,8 @@ "pull_request:44028", "pull_request:44029", "pull_request:44030", - "pull_request:46106" + "pull_request:46106", + "pull_request:46255" ], "soft_pairs": [ "pull_request:43998|pull_request:44002", @@ -19501,304 +19697,332 @@ "pull_request:44028|pull_request:44029", "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030", + "pull_request:46106|pull_request:46255", "pull_request:44027|pull_request:46106", "pull_request:44026|pull_request:46106", "pull_request:44001|pull_request:46106", "pull_request:44002|pull_request:46106", - "pull_request:43998|pull_request:46106" + "pull_request:44027|pull_request:46255", + "pull_request:44026|pull_request:46255", + "pull_request:44001|pull_request:46255", + "pull_request:44002|pull_request:46255" ], - "prepared_review_unit_hash": "5ad44f739a266a8e045228bdadbdea5cd9221f8312e1f75ea38753a94742b421", + "prepared_review_unit_hash": "2c32db8e887026383eaa55812bb6031337e2119d81a09ee2a9b56b61fe772805", "component_budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 44, - "serialized_chars": 43388, - "estimated_input_tokens": 10847, - "estimated_eval_tokens": 21950 + "node_count": 15, + "item_count": 15, + "soft_pair_count": 48, + "serialized_chars": 46843, + "estimated_input_tokens": 11711, + "estimated_eval_tokens": 23678 }, "budget": { - "node_count": 14, - "item_count": 14, - "soft_pair_count": 44, - "serialized_chars": 43388, - "estimated_input_tokens": 10847, - "estimated_eval_tokens": 21950 + "node_count": 15, + "item_count": 15, + "soft_pair_count": 48, + "serialized_chars": 46843, + "estimated_input_tokens": 11711, + "estimated_eval_tokens": 23678 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs share a generic 'refactor output tracing' theme, but they target different models/files and usually different forward signatures or tracing mechanisms. I do not see any pair that is the same concrete change, so I rejected all soft edges.", - "confidence": 0.96, + "summary": "The set is not a duplicate cluster. Most items are separate model-specific output-tracing refactors, and the last pair is an unrelated docs/kernel rename. None of the soft edges should be merged as same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the best representative of the cluster because it is a substantial, two-file output-tracing refactor in the Mamba/FalconMamba path; however, the cluster is not a duplicate set, just a group of similar model-specific refactors.", + "canonical_pr_reason": "No single canonical PR fits: the artifacts split into distinct model-family refactors plus an unrelated docs/kernel change.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest single representative: it spans two related model implementations and shows the clearest concrete output-tracing change among otherwise separate per-model refactors.", + "best_pr_reason": "No PR is a good global duplicate target; 44025 is the broadest of the output-tracing refactors, but it still covers only one model family and does not unify the rest.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different models and code paths (timm_backbone vs upernet); same output-tracing pattern is too broad to be a duplicate." + "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 (timm_backbone vs vision_text_dual_encoder); similar refactor style, not the same underlying change." + "reason": "Different architectures; same refactor theme but not the same bug or change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different models (timm_backbone vs univnet) and distinct forward semantics; not one concrete bug/change." + "reason": "Separate model implementation, so this is not a duplicate change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Separate model implementations (timm_backbone vs mamba/falcon_mamba); this is a shared refactor theme, not a duplicate fix." + "reason": "Falcon/Mamba vs Timm backbone are unrelated code paths." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "timm_backbone and rwkv are unrelated code paths; no evidence of the same underlying bug or change." + "reason": "Different model families; not mergeable as one concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated models; both touch output handling, but not the same fix." + "reason": "MobileNetV1 and UnivNet are unrelated model code paths." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths (mobilenet_v1 vs upernet); same tracing refactor pattern is insufficient." + "reason": "Different models; shared tracing refactor wording is too broad." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "MobileNetV1 vs DepthAnything are different models and different forward APIs; not a duplicate." + "reason": "Different model families and different forward paths." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "MobileNetV1 and VisionEncoderDecoder have unrelated implementations; similarity is only at a high-level refactor pattern." + "reason": "No shared concrete bug; only similar output-tracing cleanup." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 vs SpeechEncoderDecoder are different models and code paths; not the same concrete change." + "reason": "Distinct encoder-decoder model wrappers, not the same underlying issue." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "vision_text_dual_encoder and univnet are unrelated model families; both are output-tracing refactors, but not duplicates." + "reason": "Vision-text dual encoder and UnivNet are unrelated code paths." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different models (vision_text_dual_encoder vs upernet) and different interfaces; no shared concrete bug." + "reason": "Different model implementations; not the same change." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different model implementations and different output-capturing logic; too broad to merge as one PR." + "reason": "Unrelated model families; cannot plausibly be one PR." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "vision_text_dual_encoder and depth_anything are separate model paths; no indication of the same change." + "reason": "Different models and different output signatures." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated models; same refactor style does not make them duplicates." + "reason": "UnivNet and UperNet are unrelated architectures." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and different tracing/capture implementations; not the same underlying fix." + "reason": "Separate model families; only the refactor style matches." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "UnivNet vs DepthAnything are unrelated code paths; only the general output-tracing theme matches." + "reason": "Different models and different tracing/capture changes." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are unrelated model implementations; no same concrete bug/change." + "reason": "UperNet and RWKV are unrelated code paths." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different subsystems and model code paths; these are parallel refactors, not duplicate fixes." + "reason": "Different model stacks; no single concrete fix." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/FalconMamba output tracing and CodeGen output tracing are separate model-specific changes." + "reason": "Falcon/Mamba and CodeGen are different model implementations." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "timm_backbone and codegen are unrelated models; same high-level refactor theme is not enough." + "reason": "Timm backbone and CodeGen are unrelated; only both touch model outputs." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen have different forward paths and output behavior; not the same change." + "reason": "UnivNet and CodeGen do not share the same underlying bug." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are unrelated models; this is only a shared 'output tracing' refactor pattern." + "reason": "UperNet and CodeGen are different code paths." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "vision_text_dual_encoder vs codegen are different code paths with no shared concrete bug." + "reason": "Different model families; too broad to treat as duplicates." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "DepthAnything and VisionEncoderDecoder are distinct model families; similar output-tracing work but not the same PR." + "reason": "Depth Anything and Vision-Encoder-Decoder are different concrete model wrappers." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "DepthAnything and SpeechEncoderDecoder are unrelated implementations; no evidence they fix the same issue." + "reason": "Depth Anything and Speech-Encoder-Decoder are unrelated." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "DepthAnything vs SuperPoint are different tasks and code paths; not duplicates." + "reason": "Different models; same refactor theme is insufficient." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "DepthAnything and RWKV are unrelated models; same refactor language only." + "reason": "Depth Anything and RWKV are unrelated code paths." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "DepthAnything and DPR are different models and forward contracts; not the same concrete change." + "reason": "Depth Anything and DPR are different model families." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44027", "accept": false, - "reason": "VisionEncoderDecoder and RWKV are unrelated code paths; no shared underlying bug." + "reason": "Vision-Encoder-Decoder and Speech-Encoder-Decoder are separate wrappers, not one fix." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44028", "accept": false, - "reason": "VisionEncoderDecoder and SpeechEncoderDecoder are different implementations; the overlap is only in output handling refactoring." + "reason": "Different models; no shared concrete defect." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44029", "accept": false, - "reason": "VisionEncoderDecoder vs SuperPoint are unrelated models; not a duplicate fix." + "reason": "Different architectures and code paths." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "VisionEncoderDecoder and DPR touch different model families and different forward/output behavior." + "reason": "Vision-Encoder-Decoder and DPR are unrelated changes." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "SpeechEncoderDecoder and SuperPoint are unrelated model code paths; shared theme is too broad." + "reason": "Speech-Encoder-Decoder and SuperPoint are unrelated models." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "SpeechEncoderDecoder and RWKV are different architectures; no same concrete code-path problem." + "reason": "Different code paths; only the tracing refactor is common." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "SpeechEncoderDecoder and DPR are separate model families with different output semantics." + "reason": "Speech-Encoder-Decoder and DPR do not fix the same bug." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated models; these are not mergeable as one bug fix." + "reason": "SuperPoint and RWKV are unrelated implementations." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint and DPR do not share the same underlying change; only a generic tracing refactor." + "reason": "Different model families and output behavior." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are different model paths and output contracts; not the same concrete bug." + "reason": "RWKV and DPR are different concrete code paths." + }, + { + "left": "pull_request:46106", + "right": "pull_request:46255", + "accept": false, + "reason": "Docs toctree reshaping and kernel-package renaming are unrelated changes." }, { "left": "pull_request:44027", "right": "pull_request:46106", "accept": false, - "reason": "SpeechEncoderDecoder code refactor and a docs toctree change are unrelated." + "reason": "Model code vs docs navigation; not the same underlying change." }, { "left": "pull_request:44026", "right": "pull_request:46106", "accept": false, - "reason": "VisionEncoderDecoder model code and a documentation navigation update are unrelated." + "reason": "Different domains: model implementation versus documentation." }, { "left": "pull_request:44001", "right": "pull_request:46106", "accept": false, - "reason": "UnivNet model code and a docs toctree change are unrelated artifacts." + "reason": "UnivNet output tracing and docs toctree edits are unrelated." }, { "left": "pull_request:44002", "right": "pull_request:46106", "accept": false, - "reason": "UperNet model code and a documentation toctree update are not the same change." + "reason": "Model refactor versus docs restructuring; no duplicate basis." }, { - "left": "pull_request:43998", - "right": "pull_request:46106", + "left": "pull_request:44027", + "right": "pull_request:46255", + "accept": false, + "reason": "Speech-encoder-decoder refactor is unrelated to kernel package naming." + }, + { + "left": "pull_request:44026", + "right": "pull_request:46255", + "accept": false, + "reason": "Vision-encoder-decoder code and kernel docs/imports are different changes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:46255", + "accept": false, + "reason": "UnivNet refactor and liger-kernels rename do not target the same bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:46255", "accept": false, - "reason": "timm_backbone model code and a docs navigation change are unrelated." + "reason": "UperNet tracing cleanup is unrelated to the kernel package rename." } ] }, @@ -19812,7 +20036,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19831,52 +20055,52 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "f3df71bd3d3e7b1424d49f0cd06e94d2511af4c13706f6b655ba7f650bb36de6", + "prepared_review_unit_hash": "3de8a058c425e710849e633676e7c95395a5ba7e0ebe70bd81275bfa47d80e8b", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10913, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10911, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10913, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10911, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated model-addition PRs. They only overlap in standard Transformers boilerplate (docs, auto-registry, model init files), not in the same concrete model or code-path change.", - "confidence": 0.99, + "summary": "All three pull requests are separate model additions, not duplicates: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS. They only overlap on standard Transformers boilerplate such as auto registries, docs TOC, and test scaffolding.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS are distinct model implementations with different files, configs, and tests.", + "canonical_pr_reason": "No canonical PR: the three PRs implement different models and do not appear to fix or add the same concrete code path.", "best_issue_reason": null, - "best_pr_reason": "No single PR is a duplicate target; if one must be named, 45075 is the most self-contained implementation PR, but it is still unrelated to the others.", + "best_pr_reason": "No single PR is a good duplicate representative for the cluster because the changes are model-specific and not mergeable into one shared fix.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and modalities (PenguinVL vs Voxtral TTS). Shared registry/docs boilerplate is not enough to treat them as the same change." + "reason": "Different model implementations (PenguinVL vs Voxtral TTS); shared auto-registry/docs files are generic and do not indicate the same bug or change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "Both add model support, but for different models with different implementation files and conversion paths; not the same underlying bug or feature." + "reason": "PenguinVL and DeepSeek-OCR-2 are distinct model adds with different files, conversion code, and tests; overlap is only framework boilerplate." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "DeepSeek-OCR-2 and Voxtral TTS are separate model integrations with distinct code paths; overlap is only generic auto-registration/doc updates." + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model additions; they do not target the same concrete code-path problem and would not plausibly merge into one PR." } ] }, @@ -19890,7 +20114,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19906,12 +20130,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "000475f2050e9d2677e6ab6ed6003586967976c70bb4ea739c73d0d608477a6b", + "prepared_review_unit_hash": "fd00323e0e43efe37e212de4e6ba94e6ee7db76c9e5122ea658c21b2012b3d9e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4407, + "serialized_chars": 4408, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -19919,7 +20143,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4407, + "serialized_chars": 4408, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -19928,18 +20152,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related feature PRs, but not duplicates: one adds the generic module fusion API, the other adds a specific n-to-1 KernelConfig-based fusion integration on top of that idea.", - "confidence": 0.86, + "summary": "Two open PRs about fusion are related at a high level but cover different layers: one adds a generic Module Fusion API, the other wires n-to-1 kernel fusion through KernelConfig and model loading. They are not the same concrete change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "If forced to pick one representative, PR #44979 is the better canonical because it introduces the broader foundational API; PR #45363 is a narrower integration of a specific fusion use case.", + "canonical_pr_reason": "PR #45363 is the more user-facing, end-to-end change: it threads KernelConfig through model loading, adds integration hooks, and includes an example. It better represents the practical fusion feature in this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #44979 is the best single representative of the cluster because it is the core API change rather than one concrete application of it.", + "best_pr_reason": "PR #45363 is the strongest standalone PR in the cluster because it delivers the broader functional behavior and has the clearest application path. PR #44979 is a foundational API piece, but not the best single representative of the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related feature work, but not the same concrete change: #44979 adds the generic module_fusion API, while #45363 implements n-to-1 fusion via KernelConfig and integration plumbing. They affect different code paths and could be merged independently." + "reason": "Both are about fusion, but they target different code paths and abstractions: #44979 introduces a generic Module Fusion API in src/transformers/module_fusion.py, while #45363 implements n-to-1 kernel fusion via KernelConfig and related integration changes. They could coexist in separate PRs and are not the same underlying change." } ] }, @@ -19953,7 +20177,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19969,12 +20193,12 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "526dfbafaaad8b41abe36b4891e85e509d1e0c75bb801f7dfff6cc99b2aede44", + "prepared_review_unit_hash": "41422635a6e8ec1b2ea1d3f382e1aca7693f43ca2b17a936a771c9b506b7c076", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4414, + "serialized_chars": 4413, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -19982,7 +20206,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4414, + "serialized_chars": 4413, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -19991,18 +20215,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two near-duplicate PRs fixing the same Mistral regex patch AttributeError in the same code path. The later PR is the better representative because it includes the same fix plus tests.", - "confidence": 0.98, + "summary": "Two pull requests address the same Mistral regex patch bug in the same code path; the later PR adds tests and is the more complete candidate.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45317 is the most complete representative: it implements the same fix as #45086 and adds test coverage, making it the better canonical PR for this bug.", + "canonical_pr_reason": "PR 45317 is the better canonical PR: it fixes the same AttributeError in _patch_mistral_regex and also adds test coverage, making it the more complete version of the change.", "best_issue_reason": null, - "best_pr_reason": "PR #45317 is the best overall PR because it addresses the same concrete AttributeError in _patch_mistral_regex and includes validation via tests.", + "best_pr_reason": "PR 45317 is the strongest overall representative because it contains the core fix plus tests, while PR 45086 is a smaller earlier version of the same patch.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug and same code-path fix in src/transformers/tokenization_utils_tokenizers.py; #45317 is effectively a superset with tests, so these can plausibly be merged into one change." + "reason": "Accept: both PRs target the same AttributeError in _patch_mistral_regex, touch the same file, and the later PR is effectively an expanded version of the earlier fix with tests." } ] }, @@ -20016,7 +20240,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20032,7 +20256,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "b1d7142d17ce644d335faf857cd6905f7b726279701006b93fa0ebbcf627b088", + "prepared_review_unit_hash": "335d4b7688821f8d4d1305ff75bbd378d5e1402e66d989b43a168d5f51ee1e95", "component_budget": { "node_count": 2, "item_count": 2, @@ -20054,18 +20278,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about NemotronH, but they address different change sets: one is an inheritance/conversion refactor, the other is a broader split into NemotronHDense/NemotronHSparse with docs, auto-mapping, and tests.", - "confidence": 0.93, + "summary": "These are related NemotronH pull requests, but they address different changes: one is a conversion-mapping/refactor for inheriting GraniteMoeHybrid-style weights, while the other splits NemotronH into dedicated dense/sparse model implementations and updates docs/auto-mapping.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45569 is the better canonical PR because it represents the broader, more user-visible architectural change and includes the new model split, docs, and tests.", + "canonical_pr_reason": "PR #45569 is the better canonical representative because it is the broader, more complete architecture-level change and includes the new dense/sparse model split plus the associated docs and auto-mapping updates.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45569 is the best representative of the cluster since it covers the larger end-state NemotronH redesign and is more complete than the narrower refactor PR.", + "best_pr_reason": "PR #45569 is the strongest standalone candidate for the cluster because it covers the main user-facing model reorganization, while PR #45115 is a narrower conversion-mapping refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Shared NemotronH files are not enough here: 45115 is a converter/inheritance refactor, while 45569 adds separate dense/sparse model implementations and docs/tests. They do not look like the same concrete fix and would not plausibly merge into one PR." + "reason": "They touch overlapping NemotronH code, but they do not fix the same concrete bug or implement the same change. #45115 is about checkpoint conversion key renames and inheritance mapping; #45569 introduces separate NemotronHDense/NemotronHSparse implementations and documentation changes. These are distinct PRs and not mergeable as one duplicate change." } ] }, @@ -20079,7 +20303,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20095,12 +20319,12 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "5f7e2ee6ac8a441baf930c88b6166d9a8fd7a055a691d4dc4efc42017d37bebe", + "prepared_review_unit_hash": "10b259a884d9b144ed5fc7fea945fba8f636e079edacef12cdd5ff11cd8210e2", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5397, + "serialized_chars": 5398, "estimated_input_tokens": 1350, "estimated_eval_tokens": 2956 }, @@ -20108,7 +20332,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5397, + "serialized_chars": 5398, "estimated_input_tokens": 1350, "estimated_eval_tokens": 2956 }, @@ -20117,18 +20341,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs: one introduces heterogeneous per-layer config plumbing, the other adds heterogeneous modeling/runtime support. They share helper files because the model PR depends on the config work, but they are not the same change.", - "confidence": 0.92, + "summary": "These are related feature PRs, but not duplicates: one adds heterogeneous per-layer config support, the other adds heterogeneous modeling/runtime support on top of that foundation.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader umbrella change: it introduces heterogeneous model support across cache, masking, modeling, and GPT-OSS codepaths, so it best represents the cluster theme.", + "canonical_pr_reason": "PR 45332 is the broader, more representative change: it covers model/runtime integration for heterogeneous layers and incorporates the config machinery it needs. PR 45333 is the narrower config-only foundation.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the best representative of the cluster because it covers the end-to-end heterogeneous model functionality rather than just the configuration layer.", + "best_pr_reason": "PR 45332 is the best single representative for the cluster because it reflects the end-user-facing feature and the larger code-path change, while 45333 is an enabling subset.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "These are related prerequisite/feature PRs, but not the same concrete change. #45333 adds per-layer config support; #45332 adds model/runtime handling built on top of that. They would not plausibly be merged as one identical fix." + "reason": "Do not merge as duplicates: 45333 is config-layer support only, while 45332 adds heterogeneous cache/masking/modeling behavior. They share infrastructure files, but they address different layers of the feature and 45332 appears to build on 45333 rather than fix the same concrete problem." } ] }, @@ -20142,7 +20366,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20158,7 +20382,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "7ede852ea9c11a3795231f6277fa37bda433281467348a19d3540d1468e852bb", + "prepared_review_unit_hash": "cebfe99123b860b8e3e20ee7b00c33fcae340e95db3a1048a17e5dc32694c1f5", "component_budget": { "node_count": 2, "item_count": 2, @@ -20180,18 +20404,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both touch typing-related cleanup in Transformers, but they address different changes: one broadly enables type checking across many `src/transformers/*.py` files, while the other adds shared typing protocols and aliases for specific model/PEFT utilities.", + "summary": "Both PRs are typing-related and overlap on a few files, but they are not the same change: one is a broad repository-wide type-checking sweep, while the other adds specific typing protocols/aliases and related model utilities support. The overlap looks incidental rather than a single duplicated fix.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45425 is the better canonical representative because it makes a focused change to shared typing infrastructure (`_typing.py` and related utility code) rather than a broad repository-wide type-checking sweep.", + "canonical_pr_reason": "PR 45425 is the more focused and coherent typing change; PR 45415 is a much broader type-checking sweep across many files and config.", "best_issue_reason": null, - "best_pr_reason": "PR #45425 is the best single PR to anchor the cluster if needed: it has a narrower, more concrete purpose and is less likely to be a catch-all of unrelated typing edits than #45415.", + "best_pr_reason": "PR 45425 is the best representative of the cluster because it is narrower and centers on concrete typing infrastructure additions rather than a large cross-file type-checking pass.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-related, but they are not the same underlying change. #45415 is a broad type-checking enablement across many modules plus config tweaks, while #45425 adds specific typing aliases/protocols for modeling_utils and PEFT integration. They would not plausibly be merged as one PR." + "reason": "Related typing work, but not the same underlying change: 45415 is a broad type-checking rollout, while 45425 adds specific typing aliases/protocols and model utility support. They would not plausibly be merged into one PR as a single fix." } ] }, @@ -20205,7 +20429,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20213,13 +20437,12 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-45487-9", + "cluster_id": "cluster-45487-8", "nodes": [ "pull_request:45487", "pull_request:46069", "pull_request:46073", "pull_request:46109", - "pull_request:46135", "pull_request:46151", "pull_request:46154", "pull_request:46196", @@ -20234,111 +20457,104 @@ "pull_request:46073|pull_request:46154", "pull_request:46073|pull_request:46196", "pull_request:46109|pull_request:46151", - "pull_request:46135|pull_request:46151", "pull_request:46151|pull_request:46230", "pull_request:46154|pull_request:46196", "pull_request:46154|pull_request:46230" ], - "prepared_review_unit_hash": "1c31928ffc1f3db130d2d724b99fbd56bce8a629fe70c801974cb2c0eff33cf3", + "prepared_review_unit_hash": "18c21345f4c0271e28bef60a51b1ee5ad32cb2592d17896aceb6b74884e44790", "component_budget": { - "node_count": 9, - "item_count": 9, - "soft_pair_count": 12, - "serialized_chars": 23819, - "estimated_input_tokens": 5955, - "estimated_eval_tokens": 12166 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 11, + "serialized_chars": 21558, + "estimated_input_tokens": 5390, + "estimated_eval_tokens": 11036 }, "budget": { - "node_count": 9, - "item_count": 9, - "soft_pair_count": 12, - "serialized_chars": 23819, - "estimated_input_tokens": 5955, - "estimated_eval_tokens": 12166 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 11, + "serialized_chars": 21558, + "estimated_input_tokens": 5390, + "estimated_eval_tokens": 11036 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a mix of loosely related PRs, not true duplicates: a small SDPA/attn_implementation support subcluster across different models, a couple of distributed-test expectation updates, and separate model-parallel/device-placement fixes. None of the soft edges looks like the same concrete bug/change.", + "summary": "This cluster is mostly a set of unrelated, model-specific PRs that share broad themes like SDPA/dispatch support, model-parallel fixes, or test expectation updates. They are not duplicates of the same concrete bug/change, so all soft edges should be rejected.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the items split into distinct model-specific feature PRs and test/fix PRs, so there is no clear duplicate to collapse onto one PR.", + "canonical_pr_reason": "PR #46154 is the clearest representative of the common SDPA-dispatch theme: it makes concrete code changes and updates tests, while the other PRs are narrower model-specific variants or unrelated trainer/test adjustments.", "best_issue_reason": null, - "best_pr_reason": "If one PR must stand in for the cluster, 46154 is the clearest representative of the SDPA-dispatch subcluster, but it still does not duplicate the other items.", + "best_pr_reason": "PR #46154 is the best single representative of the cluster\u2019s recurring attention-dispatch work, with a substantial implementation plus test coverage; it is more central than the purely model-specific or test-only PRs.", "soft_edge_verdicts": [ { "left": "pull_request:45487", "right": "pull_request:46230", "accept": false, - "reason": "Both are cross-device/model-parallel fixes, but they address different code paths in different models: token-type buffering in AltCLIP/ChineseCLIP vs pad_token_id realignment during generation." + "reason": "Different bugs: #45487 fixes model-parallel token type/device handling in AltCLIP/ChineseCLIP, while #46230 fixes pad_token_id device placement during generation for multimodal model-parallel setups." }, { "left": "pull_request:45487", "right": "pull_request:46151", "accept": false, - "reason": "AltCLIP/ChineseCLIP model-parallel inference fix is unrelated to token-classification pipeline test expectations." + "reason": "One is a model-parallel runtime fix; the other is a test expectation update for distributed trainer device strings." }, { "left": "pull_request:46069", "right": "pull_request:46073", "accept": false, - "reason": "Same broad feature theme (attention dispatch), but different model implementations and code paths; not the same underlying change." + "reason": "Same broad feature area, but different models and different code paths: ShieldGemma2 vs CTRL SDPA dispatch." }, { "left": "pull_request:46069", "right": "pull_request:46154", "accept": false, - "reason": "Different model-specific SDPA dispatch PRs, not a single concrete bug or mergeable duplicate." + "reason": "Both add attn dispatch support, but for different architectures (ShieldGemma2 vs T5/MusicGen); not the same concrete fix." }, { "left": "pull_request:46069", "right": "pull_request:46196", "accept": false, - "reason": "Different model-specific attention-dispatch implementations; related feature area, but not duplicate fixes." + "reason": "Different model families and implementation details; this is not one shared bug or mergeable fix." }, { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both add SDPA dispatch support, but for different models and separate code paths." + "reason": "Both are SDPA dispatch PRs, but they target different model implementations and are not the same code-path problem." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "Same feature class, different model-specific implementations; not duplicate PRs." + "reason": "Different model architectures and separate dispatch implementations; too broad to treat as duplicates." }, { "left": "pull_request:46109", "right": "pull_request:46151", "accept": false, - "reason": "Both tweak test expectations, but they target different suites and behaviors (distributed trainer device naming vs token-classification pipeline outputs)." - }, - { - "left": "pull_request:46135", - "right": "pull_request:46151", - "accept": false, - "reason": "Kosmos2.5 XPU integration expectations and token-classification pipeline expectations are different test changes, not the same bug." + "reason": "Both are test-related, but they cover different test suites and different assertions; not the same underlying change." }, { "left": "pull_request:46151", "right": "pull_request:46230", "accept": false, - "reason": "Pipeline test expectation changes are unrelated to the generation-time pad_token_id device fix." + "reason": "#46151 is a pipeline test expectation update; #46230 is a generation/model-parallel runtime fix." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "Both are SDPA dispatch support, but they modify different model code paths and are not mergeable as one duplicate PR." + "reason": "Different model-specific SDPA/attention dispatch implementations; same theme, but not the same bug or code path." }, { "left": "pull_request:46154", "right": "pull_request:46230", "accept": false, - "reason": "Attention-dispatch support in T5 is unrelated to the model-parallel generation device fix in the multimodal/generation stack." + "reason": "T5 SDPA dispatch and multimodal generation device realignment are unrelated fixes and would not plausibly collapse into one PR." } ] }, @@ -20352,7 +20568,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20360,73 +20576,62 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-45513-3", + "cluster_id": "cluster-45513-2", "nodes": [ "pull_request:45513", - "pull_request:46084", - "pull_request:46098" + "pull_request:46084" ], "soft_pairs": [ - "pull_request:45513|pull_request:46084", - "pull_request:46084|pull_request:46098" + "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "ea5838a6e8dd1495f4b59fc5bab06c7816c758ce960807e79e47d26230cf0266", + "prepared_review_unit_hash": "ed06a07864666b1fdf51998a87b0249dc43168248f7d35806792a30cb89aec0c", "component_budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 7428, - "estimated_input_tokens": 1857, - "estimated_eval_tokens": 3970 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4859, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 }, "budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 7428, - "estimated_input_tokens": 1857, - "estimated_eval_tokens": 3970 + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4859, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two clusters are actually independent: the Qwen3.5 linear-attention cache fix is unrelated to the Mamba2 cache-length bug, while the two Mamba2 PRs address the same underlying use_cache seq_len>1 problem, with one being a broader superset of the other.", - "confidence": 0.89, + "summary": "Two open PRs both address cache-aware forward passes with seq_len > 1, but they target different model families and code paths, so they are not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46098 is the best canonical PR because it directly targets the tracked issue, covers both CPU and GPU paths, and includes tests/docs, while #46084 appears to be a narrower partial fix of the same Mamba2 bug.", + "canonical_pr_reason": "PR 45513 is the broader representative change in the cluster: it spans multiple Qwen/Olmo hybrid model files and fixes the multi-token cached-forward path rather than a single model-specific kernel.", "best_issue_reason": null, - "best_pr_reason": "#46098 is the strongest representative artifact: it is the most complete fix for the Mamba2 use_cache seq_len>1 regression and is easiest to understand as the primary change in the cluster.", + "best_pr_reason": "PR 46084 is the strongest standalone PR: it has an explicit linked issue, a focused Mamba2-specific fix, and a concrete seq_len > 1 cache bug in one code path.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different models and bugs: #45513 fixes Qwen3.5 GDN linear-attention cached forward behavior, while #46084 fixes Mamba2 cache handling for seq_len>1. Same broad theme of caching is not enough to treat them as duplicates." - }, - { - "left": "pull_request:46084", - "right": "pull_request:46098", - "accept": true, - "reason": "Both are about the same Mamba2 use_cache seq_len>1 incorrect-results bug. #46084 fixes a narrower CUDA/cache path, while #46098 addresses the same underlying issue more broadly across CPU and GPU paths, so they look like the same change family and could be merged into one PR." + "reason": "Both involve cache handling for seq_len > 1, but 45513 fixes linear-attention/GDN cached forward across Qwen/Olmo variants while 46084 fixes Mamba2 CUDA-kernel cache logic. Different model families and different underlying code paths, so they should not be merged as one PR." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: #45513 concerns Qwen3.5/Qwen3-style cached linear-attention behavior, while #46084 and #46098 both target Mamba2 use_cache seq_len>1 and even share issue target 46032 and files. The false verdict on 45513 vs 46084 is conservative, and the true verdict on 46084 vs 46098 is supported by the packet as the same bugfix family with one broader than the other." - }, - "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 }, { "label": "pull_request", "component_index": 24, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20442,7 +20647,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "022d2859df89309c2b9cfd49286d7add284036560614f04e0ffb261931f75dc1", + "prepared_review_unit_hash": "ada951a51f0fcb95f66c530f02f157678afb44b63e116f0ed72590ed261ec538", "component_budget": { "node_count": 2, "item_count": 2, @@ -20464,18 +20669,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about expert-parallel loading behavior, but they fix different concrete bugs: one is about preserving distributed_config / EP flags during serialization and accelerate launch, while the other fixes expert parameters being overwritten during FSDP2/rank-0 broadcast loading. They are related but not duplicates.", - "confidence": 0.93, + "summary": "Both PRs concern expert-parallel loading, but they fix different bugs: one is about serializing/loading `distributed_config` and adding `has_ep`, while the other prevents expert weights from being overwritten during FSDP2/rank-0 broadcast. They are related by theme, not duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45662 is the stronger representative of the cluster because it targets a specific runtime loading failure with a clear code-path fix in model loading and related EP integrations.", + "canonical_pr_reason": "PR #45662 is the stronger canonical PR: it targets a specific, concrete loading corruption path (expert parameters overwritten during distributed loading) and the diff directly implements the fix across the affected code paths.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the best standalone PR in this cluster since it addresses a concrete correctness bug in expert parameter loading rather than a narrower config-serialization/flag-preservation issue.", + "best_pr_reason": "PR #45662 is the best fit for the cluster because it addresses the clearest underlying runtime bug in expert-parallel loading, whereas PR #45548 is about a different accelerate/ZeRO-3 serialization/loading compatibility issue.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Different underlying bugs: #45548 preserves EP/distributed_config state for loading via accelerate/ZeRO-3, while #45662 fixes expert parameters being overwritten by rank-0 broadcast during FSDP2 loading. Related subsystem, but not the same code-path problem." + "reason": "Different failure modes and code paths: #45548 handles config serialization / `has_ep` for accelerate+ZeRO-3 loading, while #45662 fixes expert parameters being overwritten during FSDP2 rank-0 broadcast." } ] }, @@ -20489,7 +20694,7 @@ { "label": "pull_request", "component_index": 25, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20505,40 +20710,40 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "9b9f7bc737c6fd6ea6dc6501e08acd69c579d7c593665164184e03616ee69b6e", + "prepared_review_unit_hash": "9d16a74d3b4e377fe3b4367fbee6d9d95c538a2488919fa75d36f3e453252413", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4590, - "estimated_input_tokens": 1148, - "estimated_eval_tokens": 2552 + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4590, - "estimated_input_tokens": 1148, - "estimated_eval_tokens": 2552 + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same xdist/captured_info artifact collision problem under issue 45561; PR 45645 is the more complete superset because it updates the producer and the CI consumer of the logs.", - "confidence": 0.88, + "summary": "Both PRs address the same xdist/captured_info problem in testing_utils, with 45645 extending the same fix to CI log collection. They are closely related, but 45645 is the more complete change set.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the better canonical PR because it contains the xdist-safe captured_info fix plus the workflow-side handling needed to read multiple worker logs, making it the more complete representation of the change.", + "canonical_pr_reason": "PR 45645 is the broader, more complete version of the xdist captured_info fix: it includes the core testing_utils change plus the workflow/log-consumption updates needed to preserve the debug artifacts.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 best represents the final intended fix: it addresses the same captured_info collision bug as 45639 and also preserves CI debug log collection across multiple files.", + "best_pr_reason": "PR 45645 is the best representative of the cluster because it covers the underlying bug fix and the downstream CI handling changes in one place.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Accept: both PRs fix the same underlying xdist collision in captured_info artifact generation; 45645 appears to extend 45639 with CI log-consumption updates, so they are close enough to the same change to triage together." + "reason": "They share the same core xdist-safe captured_info fix in testing_utils and same issue target; 45645 appears to be a superset that adds CI artifact handling, so they look like the same underlying change." } ] }, @@ -20552,7 +20757,7 @@ { "label": "pull_request", "component_index": 26, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20568,40 +20773,40 @@ "soft_pairs": [ "pull_request:45670|pull_request:46229" ], - "prepared_review_unit_hash": "028828b996455e8b83d5c3de4a29370f13341f38ba3a53c41f1af86dc59c0370", + "prepared_review_unit_hash": "1ab5092bc1a21f62285d327f6bf6bea7dbca77134db8af4cc8b438922ddd6653", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4247, - "estimated_input_tokens": 1062, - "estimated_eval_tokens": 2380 + "serialized_chars": 4355, + "estimated_input_tokens": 1089, + "estimated_eval_tokens": 2434 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4247, - "estimated_input_tokens": 1062, - "estimated_eval_tokens": 2380 + "serialized_chars": 4355, + "estimated_input_tokens": 1089, + "estimated_eval_tokens": 2434 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one adds a model-class mapping entry for GLM-ASR in AutoModel, while the other implements a new MTP candidate generator path. They should not be clustered as duplicates.", - "confidence": 0.96, + "summary": "These PRs are unrelated: one is a small AutoModel registration tweak for GLM ASR, while the other adds a substantial MTP generation feature across multiple files. They should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46229 is the more substantive and representative change in this cluster: it adds a new MTP generation implementation across generation code, whereas #45670 is a narrow AutoModel registration nit.", + "canonical_pr_reason": "No duplicate/canonical relationship is evident between the two PRs; if one must be treated as the representative change, #46229 is the more substantial and self-contained feature PR.", "best_issue_reason": null, - "best_pr_reason": "PR #46229 is the best representative PR because it targets a concrete code-path change with broader functional impact and clearer standalone purpose.", + "best_pr_reason": "#46229 is the stronger standalone PR because it implements a concrete feature across generation code paths, whereas #45670 is a narrow model-registration nit.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46229", "accept": false, - "reason": "Different underlying changes: #45670 only updates model-to-AutoModel mappings, while #46229 adds MTP candidate generation support. They do not fix the same bug and would not be merged as one PR." + "reason": "Different code paths and goals: AutoModel registry placement vs. MTP candidate generation. They do not fix the same underlying bug or change." } ] }, @@ -20615,7 +20820,7 @@ { "label": "pull_request", "component_index": 27, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20681,7 +20886,7 @@ { "label": "pull_request", "component_index": 28, - "component_count": 28, + "component_count": 29, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20697,7 +20902,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "41224e315b33ef16f35006842fb02d673191b1bc2c8be643b0723009d28db8fe", + "prepared_review_unit_hash": "4876ba5941fb0f6e6a627d9383c2f7941eb6ea298962746101710455246313a5", "component_budget": { "node_count": 2, "item_count": 2, @@ -20719,18 +20924,81 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests address file-source introspection using safer APIs, but they fix different code paths in different modules: one adds traversable resolution in import_utils, the other replaces open()-based source inspection in modeling_utils. They are related in theme, not the same underlying change.", - "confidence": 0.91, + "summary": "These two PRs are related only at a thematic level: both replace file-open source reading with safer source inspection, but they touch different code paths and fix different mechanisms in different modules. They should not be treated as duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46207 is the cleaner representative: it targets one concrete behavior in _can_set_attn_implementation, updates the implementation and tests, and has a narrow, well-defined scope.", + "canonical_pr_reason": "PR #46207 is the better representative of the cluster because it is a concrete, narrowly scoped fix in `modeling_utils.py` with tests, addressing a specific source-reading heuristic in `_can_set_attn_implementation`.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the best standalone PR in this cluster because it addresses a specific bug/heuristic in a single code path and includes test coverage, whereas PR 46152 is a separate utility refactor in a different subsystem.", + "best_pr_reason": "PR #46207 is the strongest standalone candidate: it has the clearest problem statement, the most explicit behavioral fix, and includes validation via tests. PR #46152 is a separate import-utils change and not a better cluster anchor.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Both avoid opening source files directly, but they operate in different modules with different mechanics and targets; they are not the same concrete fix and would not plausibly be merged into one PR." + "reason": "Both PRs improve source-file access, but they operate on different modules and different call sites (`import_utils.py` vs `modeling_utils.py`). The fixes are not the same underlying bug and would not plausibly be merged into one PR without broadening into unrelated code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 29, + "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-46259-2", + "nodes": [ + "pull_request:46259", + "pull_request:46260" + ], + "soft_pairs": [ + "pull_request:46259|pull_request:46260" + ], + "prepared_review_unit_hash": "461374eb8d8bb06e43b25ff24b2739e5167316a1edcc198e726d32afafb1f680", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate modeling-rule PRs affecting different model families and different code paths: one targets image/video processor do_* flag consumption, the other targets audio/speech model rule coverage. They look superficially similar because they both add docs/rules infrastructure, but they are not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the two pull requests implement different modeling rules for different subsystems, so they should remain separate.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for the cluster; each PR is the right representative for its own distinct rule addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46259", + "right": "pull_request:46260", + "accept": false, + "reason": "Different underlying rule additions and model families; they share documentation/rules plumbing but not the same concrete bug or change." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index b081da68fc8fddde0ddb2d93d63a202c84a8c7d6..7972a28c14abc5bbe7da6f9b2c00dc9a604e2ea9 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260528t060028z", + "analysis_id": "hybrid-model-20260528t120029z", "archived_artifacts": { - "hybrid": "snapshots/20260528T060028Z/analysis-runs/hybrid-model-20260528t060028z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260528T060028Z/analysis-runs/hybrid-model-20260528t060028z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-05-28T06:05:25Z", + "published_at": "2026-05-28T12:12:32Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260528T060028Z", + "snapshot_id": "20260528T120029Z", "variant": "hybrid" } diff --git a/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..f262cbad1c48a8995074e07308f6880f8463464d --- /dev/null +++ b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json @@ -0,0 +1,2053 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260528T120029Z", + "generated_at": "2026-05-28T12:12:30Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.797, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.831, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 730, + 737 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 720, + 726 + ], + [ + 1331, + 1337 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.532, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.543, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3529, + 3536 + ], + [ + 3753, + 3787 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46100-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46100, + "issue_numbers": [], + "pr_numbers": [ + 46100, + 46114 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46100, + "right_pr_number": 46114, + "code_similarity": 1.0, + "size_similarity": 1.0, + "file_overlap": 1.0, + "area_overlap": 1.0, + "patch_similarity": 1.0, + "shared_filenames": [ + "tests/models/vit/test_modeling_vit.py" + ], + "shared_file_areas": [ + { + "filename": "tests/models/vit/test_modeling_vit.py", + "left_ranges": [ + [ + 216, + 226 + ] + ], + "right_ranges": [ + [ + 216, + 226 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46100-2", + "canonical_pr_number": 46100, + "duplicate_pr_numbers": [ + 46114 + ], + "target_issue_number": null, + "reason": "PRs in cluster-46100-2 are treated as duplicates because they share soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 13 discussion comments plus 8 inbound references.", + "score": 164.0 + }, + "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/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..4dd09f72f8cc54330f81320023ba7a06b65f3465 --- /dev/null +++ b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,21013 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260528T120029Z", + "generated_at": "2026-05-28T12:12:30Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "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:43572", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43866", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:45920", + "issue:46015", + "issue:46165" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44805|issue:45823", + "issue:44918|issue:46165", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:44805|issue:44918", + "issue:43572|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43975|issue:44457", + "issue:43039|issue:45161", + "issue:44292|issue:44457", + "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:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:43541|issue:43716", + "issue:43716|issue:44292", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805", + "issue:39290|issue:41720", + "issue:41762|issue:43596", + "issue:45237|issue:45901", + "issue:43901|issue:44512", + "issue:44841|issue:45823", + "issue:43039|issue:44560", + "issue:43975|issue:46015", + "issue:45812|issue:45920", + "issue:44863|issue:46015", + "issue:43866|issue:44863", + "issue:44292|issue:46015" + ], + "prepared_review_unit_hash": "f270fcaf8003e02597617bfd45fb5a51b4e0d409170c3bafa404b2b6be4686b6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 69, + "serialized_chars": 48409, + "estimated_input_tokens": 12103, + "estimated_eval_tokens": 24462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous and not a single duplicate set. The only clear duplicate families are the AutoTokenizer wrong-token-ID v5 regression and the near-identical DeepSpeed ZeRO-3 BERT index-out-of-bounds reports.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 45920 is the broadest and most representative active report of the AutoTokenizer token-ID regression affecting multiple model families.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45920 is the best global issue anchor because it captures the widest version of the recurring tokenizer regression and is still open for follow-up.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 streaming/API failure vs config attribute error; not the same bug." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different model and hardware paths: AMD GPU failure vs MPS histogram/int error." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different multimodal dtype mismatch vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Both are dtype-related at a high level, but different models and code paths." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model failures; not the same concrete method/code-path bug." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different Qwen NVFP4 runtime issue vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Export-to-ONNX issue vs FP8 config failure; unrelated bugs." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export and CUDA assert during auto device map are unrelated." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash vs outdated RAG docs; different problem types." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup for v5 removals, but they refer to different removed items and not one concrete duplicate change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export bug vs GenerationMixin import breakage; unrelated." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: mask shape mismatch vs Gemma4 device mismatch." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding unpacking vs resume/trust-remote-code training issue; unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are runtime IndexErrors, but different models and failure mechanisms." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different zero3/BERT init failure vs mask shape mismatch." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "BERT zero3 init bug vs training-resume/trust-remote-code issue; unrelated." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video StopIteration in Qwen3-vl embedding vs training resume issue; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "StableLmConfig padding regression vs training resume issue; unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor preset bug vs ModernBERT ONNX export bug; unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but one is collective mismatch and the other is zero3 index error; different bugs." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "FSDP checkpointing recompute mismatch vs mask shape mismatch; unrelated." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same model family, but CPU-offload incompatibility and uninitialized-memory materialization are different code-path problems." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger cross_entropy call vs GPT-OSS tensor-parallel issue; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Different model/config errors: Gemma3TextConfig attribute vs Mistral tokenizer patch crash." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Both Gemma4-related, but device mismatch and expensive reverse-embedding lookup are different issues." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility vs Qwen NVFP4 runtime error; unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Different dtype/autocast issue vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL embedding unpacking vs GPT-OSS TP bug; unrelated." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Both are runtime failures, but not the same model or code path." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Different multimodal dtype issue vs MPS histogram int issue." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different models and failure modes; not duplicates." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different model families and errors." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method vs MPS histogram int error; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs TRL embedding unpacking; unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Both are DeepSpeed ZeRO-3 related, but the models and concrete failures differ." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Different docs gaps about different removed v5 features." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different model-specific issues; not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Dtype/autocast issue vs MPS histogram int issue; unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL embedding unpacking vs MPS histogram int issue; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different subsystems and failures; not a duplicate." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch vs trainer signature change; unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP mismatch vs GenerationMixin import error; unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU model failure vs docs example drift; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger kernel cross_entropy call vs tensor mask shape mismatch; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different model/runtime errors; not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 loading issues, but Gemma3 and Bert failures are different concrete bugs." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "GPT-OSS AMD issue vs table QA crash; unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs problems, but they cover different pipeline/API removals." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure for Voxtral vs Gemma4 device mismatch; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs Qwen3-vl video StopIteration; unrelated." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization regression vs docs example drift; unrelated." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same AutoTokenizer wrong-token-ID v5 regression across model families; 45920 is the broader report." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading bug vs outdated examples docs issue; unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint report vs model implementation load bug; not the same issue." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen NVFP4 runtime error vs docs issue; unrelated." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed/BERT zero3 failure vs training resume/trust-remote-code issue; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 + Bert index-0 out-of-bounds failure, just described with slightly different model wording." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific failures and different code paths." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs GPT-OSS TP issue; unrelated." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "BERT zero3 init bug vs GPT-OSS tensor-parallel issue; unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific failures and different code paths." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Different subsystems: Liger kernel loss call vs TRL embedding unpacking." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Different ZeRO-3 failure modes and different models." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Different doc gaps; not a single duplicate change." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different multimodal issues; not duplicates." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different runtime errors and different models." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "Different failures: TRL SFT embedding unpacking vs MPS histogram issue." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different subsystems and symptoms; not a duplicate." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch vs trainer signature break; unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP checkpointing mismatch vs GenerationMixin import error; unrelated." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "Model/runtime failure vs docs example drift; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger kernel cross_entropy issue vs shape mismatch error; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different model/runtime issues; not the same underlying bug." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading issue vs BERT ZeRO-3 loading issue; similar area, different bug." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "GPT-OSS AMD failure vs table QA crash; unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Different documentation updates for different removed APIs." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure vs Gemma4 device mismatch; unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger kernel issue vs Qwen3-vl video StopIteration; unrelated." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization regression vs docs drift; unrelated." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "Checkpoint loading bug vs docs issue; unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint report vs model load bug; not the same issue." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Model runtime error vs docs issue; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most items are clearly heterogeneous, and the two accepted duplicate families are well supported by the titles. The soft-edge rejections stay conservative, and the accepted pairs (45812/45920 and 43596/43638) look justified from the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 2, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36010", + "issue:39186", + "issue:39290", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43645", + "issue:43716", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44589", + "issue:44841", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45542", + "issue:45632", + "issue:45685", + "issue:45698", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45901", + "issue:45902" + ], + "soft_pairs": [ + "issue:39290|issue:43531", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:41093|issue:43638", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43404|issue:43866", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:36010|issue:39290", + "issue:45800|issue:45823", + "issue:41093|issue:41762", + "issue:43572|issue:45070", + "issue:43296|issue:43366", + "issue:43541|issue:45161", + "issue:44387|issue:45005", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:45020|issue:45800", + "issue:43531|issue:45070", + "issue:43541|issue:44589", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:45542|issue:45800", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:40990|issue:44960" + ], + "prepared_review_unit_hash": "d83af649a88d0d52de6bb2587079d09172d244da442a6ae3b6ec33f4cbb98687", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 39, + "serialized_chars": 39127, + "estimated_input_tokens": 9782, + "estimated_eval_tokens": 19820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated closed issues. Several pairs are superficially similar around model loading, zero-3, tied weights, and custom-module handling, but they do not look like the same concrete bug/change, so I would not merge any of the soft edges.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45020 is the broadest and most umbrella-like report in the set: it captures a general regression class (\u2018recent transformers versions break models using remote_code\u2019), which makes it the best candidate to anchor this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the most reusable triage target because it is broad, regression-oriented, and likely to subsume multiple adjacent compatibility complaints better than the model-specific tickets.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both are model-specific config issues, but they involve different models and different attributes; too broad to be the same bug." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper runtime failure vs table-question-answering crash; different feature areas and code paths." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "One is a GLM-4.7-Flash load/test issue, the other a corrupted Ovis2 checkpoint; unrelated failures." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both mention deepspeed zero-3 and out-of-bounds errors, but one is a mask-shape mismatch while the other is a Bert init/indexing failure." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch-version incompatibility is not the same bug as the pydantic PretrainedConfig regression." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Both are compatibility/runtime complaints, but one is a zero-3 BertModel init problem and the other is a torch/transformers version mismatch." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Incorrect detokenization and table QA crashing are different symptoms and code paths." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure does not match the PretrainedConfig field regression." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming API failure; not the same concrete bug." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Different models and unrelated failure modes." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Different model families and different underlying issues (config attribute vs device mismatch)." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming chat API failure and an MPS histogram kernel issue are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Weight-tying bug vs corrupted checkpoint; not the same problem." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Grouped-mm tracing failure and an MPS int kernel limitation are different code paths." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both mention dtype/precision, but they are different models and different execution contexts." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor loading failure and template-node compilation error are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding-window config errors." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "A package-version incompatibility is not the same as a Gemma4 device-map mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are zero-3 related index errors, but they affect different models and appear to stem from different failure conditions." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx after v5 update is a different regression than the PretrainedConfig pydantic issue." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are distinct requests/issues." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Torch-dynamo grouped_mm tracing failure is not the same as GPT-OSS tensor-parallel support." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 OOM regression and tied-weights translation-model bug are different code paths." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch and streaming chat API failure are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "init_empty_weights argument handling is separate from the PretrainedConfig field regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GPT-OSS GGUF support and Qwen3.5 streaming chat failure are unrelated." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Broad remote_code breakage and a specific torch/transformers version mismatch are different issue classes." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is not the same as the pydantic PretrainedConfig issue." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Torch-dynamo grouped_mm tracing and missing Float8 storage are unrelated runtime issues." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Different models and different zero-3 error shapes; not one underlying bug." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both concern siglip2 quality/behavior, but one is an embedding-quality complaint and the other is an AutoModel/pipeline result bug." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Composite-config unwrapping and uninitialized inv_freq memory are distinct bugs." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard-without-TF backend error is unrelated to torch/transformers incompatibility." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Gemma/Mistral image-text dtype mismatch and Qwen3.5 TRL embedding unpacking are different failures." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight-shape error and Qwen3.5 embedding unpacking are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback gating are separate problems." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both touch custom-module loading, but one is a leaf-directory cache collision and the other is wrong-module selection after save_pretrained; not the same concrete bug." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but they concern different model classes and likely different fixes; too broad to merge." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and a GLM5 issue are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 3, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41628", + "issue:41762", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43828", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:45005", + "issue:45020", + "issue:45237", + "issue:45399", + "issue:45507", + "issue:45663", + "issue:45685", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45902" + ], + "soft_pairs": [ + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44863|issue:45005", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:44492|issue:44512", + "issue:44292|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:43828|issue:44928", + "issue:36331|issue:41762", + "issue:41628|issue:45800", + "issue:43950|issue:45902", + "issue:43872|issue:44589", + "issue:43866|issue:44403", + "issue:41553|issue:43883", + "issue:38175|issue:43054", + "issue:45663|issue:45902", + "issue:44910|issue:45859", + "issue:43054|issue:45701", + "issue:41553|issue:43881", + "issue:42947|issue:44387", + "issue:44403|issue:44451", + "issue:43975|issue:45237", + "issue:43054|issue:43493", + "issue:44403|issue:44488", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:43881|issue:43883", + "issue:45685|issue:45823" + ], + "prepared_review_unit_hash": "27b865c65caf95e645b0382ae9b8d7e29d5a98bc5acc177f067003a59d29a26b", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37581, + "estimated_input_tokens": 9396, + "estimated_eval_tokens": 19048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. The only clear duplicate pair is the Transformers v5 non-persistent buffer regression; the rest are related-at-a-distance model/loading/quantization reports but not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest, most detailed report of the non-persistent-buffer regression, and issue:44534 is essentially the same bug described more generally.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 best represents the one true duplicate pair in this cluster and has the strongest problem description and reproduction context.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Different problems: FP8 config failure in Qwen3MoE vs flash-attn2 fallback being blocked by checks." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache crash on Qwen3.5 is unrelated to tied-weights handling in translation models." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior is a different code path from v5 tied-weights issues." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer ReLU discussion is unrelated to docs mentioning a removed CLI command." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is not the same as translation-model tied weights." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3.5 embedding unpacking in TRL is a different failure than GraniteMoEHybrid calling an invalid method." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Incorrect detokenization and a Qwen-3 NVFP4 runtime error are distinct bugs." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo is unrelated to a removed docs command." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "NVFP4 execution failure and GraniteMoEHybrid invalid method are separate issues." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Both report the same v5 regression where non-persistent buffers registered with persistent=False get corrupted/junk values on load." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Docs example errors for SigLIP2 are not the same as a broader text-embedding quality complaint." + }, + { + "left": "issue:43643", + "right": "issue:43950", + "accept": false, + "reason": "Remote-code config field loss is a different bug from non-persistent buffer corruption." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a dtype mismatch under autocast and the other is an RLHF NaN/SDPA fallback issue." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "CustomTrainer compute_loss API mismatch is unrelated to Gemma3 loading with ZeRO-3." + }, + { + "left": "issue:41628", + "right": "issue:45800", + "accept": false, + "reason": "Missing AutoImageProcessor import and torch/transformers version incompatibility are not the same bug." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Both involve buffers/materialization, but one is a from_pretrained regression and the other is a Qwen3_5Moe rotary embedding uninitialized-memory bug." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "bitsandbytes Int8Params argument incompatibility is different from a Float8 storage lookup failure." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "A corrupted Ovis2 checkpoint is not the same as a generic loading-noise complaint." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to missing all_tied_weights_keys on Molmo." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Zero probabilities in SigLIP2 and worse text embeddings are related model-quality reports but not the same concrete defect." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states rebuild and Qwen3_5Moe inv_freq uninitialized memory are separate model-specific bugs." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Flash-Attention packed-sequence misinterpretation and CPU-offload incompatibility in rotary embedding are different failures." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality regression is not the same as a general tokenization-version change report." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Voxtral tokenizer error handling and glm-4v-9b loading failure are unrelated." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "LoRA gradient checkpointing ineffectiveness is a different issue from int4 quantization memory growth/OOM." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise is not the same as inability to load a specific model repo." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "DeepSeek detokenization regression and GPT-OSS-20B AMD GPU failure are unrelated." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 text embedding quality and HF-vs-JAX implementation discrepancy are not the same report." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Generic loading noise is not the same as a specific sleng-bert load failure." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Missing fields in AutoConfig with trust_remote_code is too specific to merge with a broader remote_code breakage report." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Gradient checkpointing with PEFT LoRA and inefficient MoE memory usage are different problems." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading and Qwen3MoE FP8 config failure are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "glm-4v-9b loading failure and Molmo tied-weights attribute error are different model-specific bugs." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram support in MoE is unrelated to Gemma4 PLE device mismatch under device_map=auto." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43950/issue:44534, is well supported by the matching v5 non-persistent-buffer regression titles, and the rejected soft pairs are framed as distinct bugs without overstating certainty. No PR mergeability concerns apply here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 4, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43404", + "issue:43638", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45538", + "issue:45588", + "issue:45823", + "issue:45923" + ], + "soft_pairs": [ + "issue:44991|issue:45399", + "issue:42491|issue:44368", + "issue:43824|issue:43872", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:45081|issue:45399", + "issue:43782|issue:44960", + "issue:43742|issue:45399", + "issue:45020|issue:45356", + "issue:43404|issue:43950", + "issue:43638|issue:45823", + "issue:42915|issue:45020", + "issue:43827|issue:44509", + "issue:43792|issue:44220", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43257|issue:43866", + "issue:41553|issue:44991", + "issue:42947|issue:45923", + "issue:36331|issue:41093", + "issue:44928|issue:45588", + "issue:43940|issue:45313", + "issue:43257|issue:45305", + "issue:43854|issue:44863", + "issue:43872|issue:45538", + "issue:41553|issue:44488", + "issue:44479|issue:44560", + "issue:41093|issue:44805", + "issue:24643|issue:30064", + "issue:44530|issue:44863", + "issue:43366|issue:43828" + ], + "prepared_review_unit_hash": "331e76db7a687030b5bd96fbc24fcfd54c12fb283e2eb0d7318c55b889d87155", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 37049, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issues, but there are a few true duplicates: the v5 pipeline-docs regression, the `_is_hf_initialized` TypeError during model loading, and the mask/tensor shape-mismatch IndexError. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 41093 has the clearest concrete bug report for the shape-mismatch IndexError, and 44805 is just the generalized/placeholder version of the same failure.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 41093 is the strongest representative runtime bug in this set: specific, reproducible, and duplicated by the generalized 44805 report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Different subsystems and failures: tokenizer loading vs flash-attn fallback gating." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Related Qwen/LoRA wording, but one is a model-compatibility failure and the other is a warning about `tie_word_embeddings`; not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Import error for a missing class vs bitsandbytes kwarg incompatibility; different failure modes." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Broad remote-code regression vs a Gemma4 device-mismatch bug; not the same underlying issue." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer-mapping-name edge case is unrelated to the remote_code regression." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different code paths: autocast dtype mismatch vs device placement mismatch." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch crash is unrelated to flash-attn2 fallback checks." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL `weight_only` load error and GLM5 loading are different issues." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM key error is unrelated to flash-attn fallback logic." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Both mention recent-version regressions, but one is broad remote-code breakage and the other is a specific Kimi-K2.5 tokenizer regression." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Tied `lm_head` weights and non-persistent buffer corruption are different bugs." + }, + { + "left": "issue:43638", + "right": "issue:45823", + "accept": false, + "reason": "ZeRO-3 index error with a non-pretrained Bert model is unrelated to the Gemma4 device mismatch." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "Qwen3 MoE FP8 failure is unrelated to the remote_code tokenizer/model-loading regression." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression: v5 removed pipeline tasks, but the docs still reference them." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper runtime failure and `_torch_extract_fbank_features()` are too vague to confirm the same underlying bug." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same unexpected `_is_hf_initialized` TypeError during loading; one reports the bitsandbytes `Int8Params` path and the other the `init_empty_weights` path." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Different model/loading problems: Qwen3 MoE FP8 vs bitsandbytes init kwarg incompatibility." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Qwen3 MoE weight-loading issue is unrelated to the corrupted Ovis2 checkpoint report." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer error-message quality issue vs a tokenizer loading failure for a different model." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Gradient-checkpointing effectiveness and missing support flag are related in theme but not the same concrete bug." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Unexpected `compute_loss` kwarg and mask-shape IndexError are unrelated training failures." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Catastrophic gradient explosion and flash-attn `s_aux=None` crash are different problems." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are DeepSpeed weight-loading issues, but they target different Qwen variants and symptoms; not enough to treat as one bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Model-loading failure for GLM-4.7-Flash vs NemotronH checkpoint incompatibility are unrelated." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "Bitsandbytes kwarg error is unrelated to CLIPTokenizer's `model_max_length` default." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Bad error message for Voxtral tokenizer loading is unrelated to `cjvt/sleng-bert` not loading." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-input regressions, but the affected models and failure signatures differ; not the same concrete bug." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask/tensor shape-mismatch IndexError; 44805 is the genericized duplicate of the exact error in 41093." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training weight-shape runtime error and segmentation-map processing are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache group-type crash and NemotronH checkpoint loading are different failures." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support and autocast dtype mismatch are unrelated issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted duplicates are well supported by the titles/body patterns in the listed issues, and the rejected pairs avoid overstating weak similarity. No PR pairs are present, so there is no mergeability concern." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 5, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "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:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45276", + "issue:45313", + "issue:45325", + "issue:45357", + "issue:45584", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834", + "issue:45902" + ], + "soft_pairs": [ + "issue:45325|issue:45902", + "issue:45663|issue:45823", + "issue:43366|issue:43531", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:43550|issue:44910", + "issue:43645|issue:43646", + "issue:43366|issue:44589", + "issue:44560|issue:44918", + "issue:42915|issue:43957", + "issue:42617|issue:43366", + "issue:43824|issue:43957", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:43632|issue:45834", + "issue:44560|issue:44805", + "issue:38175|issue:43994", + "issue:44155|issue:45584", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:44877|issue:45030", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:45313|issue:45357", + "issue:43644|issue:44534", + "issue:42491|issue:45276", + "issue:42915|issue:44756" + ], + "prepared_review_unit_hash": "e1f931eabae8f43bc620cf9a1b1104d5ee597f14d6408426a53da1d74c5a1082", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36681, + "estimated_input_tokens": 9171, + "estimated_eval_tokens": 18598 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated Transformers regressions across loading, config, and model-specific inference bugs. Only a few pairs look like true duplicates: custom model initialization, non-persistent buffer corruption, SigLIP2 bad outputs, and the BERT loading regression affecting multiple models.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43646 is the broadest, most representative issue in the cluster: it frames the shared Transformers v5 custom-model initialization regression in a generic way that can absorb the more specific notebook repro.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best overall anchor because it is the clearest umbrella for the custom-init regression and is less model- or environment-specific than the nearby duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different bugs: Qwen2.5-VL temporal position scaling vs Qwen3_5Moe rotary embedding reading uninitialized memory after meta materialization." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but different failures: FSDP2 shared_kv_states KeyError vs device_map='auto' PLE device mismatch." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Unrelated issues: GGUF support for gpt-oss vs a Qwen3-MoE sliding_window problem." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Different code paths and symptoms: non-persistent buffer corruption vs Kosmos2.5 long-OCR index error." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Different model/runtime bugs: Qwen3-VL video StopIteration vs MPS histogram not implemented for Int." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different concrete failures: Bamba torch.compile+SDPA crash vs Qwen3.5 flash-attention illegal memory access from 3D position_ids." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying Transformers 5.0 custom-model initialization regression; the notebook wording is just a more specific reproduction." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated: gpt-oss GGUF support request vs Float8 storage lookup error." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different bugs: Qwen3-VL video StopIteration vs Qwen3.5 embedding unpacking failure in TRL SFT trainer." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Different regressions: Qwen3-MoE FineGrainedFP8Config failure vs generic meta-device loading issues." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "No shared underlying bug: 3d_parallel.py failure vs GGUF gpt-oss support." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Different problems: missing export for Qwen2.5-VL class vs model loading with torch.device('meta')." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Completely different: GenerationMixin import error vs Qwen3-MoE FP8 config failure." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Different areas: tokenizer mapping/add-new-model-like bug vs Gemma4 PLE forward device mismatch." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "Different regressions: _is_hf_initialized flag breakage vs Kosmos2.5 OCR index error." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different symptoms and likely different causes: video StopIteration vs mask/tensor shape mismatch." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 model regression: zero probabilities and nonsensical outputs point to the same bad inference/loading behavior." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different model-specific bugs: AudioFlamingo3 batched leakage vs Whisper empty-transcription failure." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Different issues: ZeRO-3 BertModel index error vs add-new-model-like/tokenizer-mapping failure." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different bugs: zero3 BertModel initialization index error vs tokenizer mapping validation issue." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Unrelated failures: tokenizer mapping logic vs mask shape mismatch at indexing time." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both mention config validation, but the affected models and failures are different; not enough to be the same bug." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Different concerns: gpu+cpu offload dequantized load failure vs disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different code paths: qwen3_moe LoRA incompatibility across versions vs BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Different runtime failures: auto device mapping cudaErrorAssert vs Qwen3-MoE FineGrainedFP8Config." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Unrelated model bugs: GLM5 issue vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Same loading regression class: both report current/latest Transformers failing to load specific BERT models." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Exact same bug: both report Transformers v5 filling non-persistent buffers with junk." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 issues: ZeRO-3 load failure for language_model vs incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:44910", + "right": "issue:44918", + "accept": false, + "reason": "Different Qwen3.5 problems: flash-attention misreading 3D position_ids vs TRL SFT embedding unpacking failure." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet, and the accepted soft-edge pairs are conservative. The rejected pairs are also phrased cautiously and do not overclaim duplicate status." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 6, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43335", + "issue:43381", + "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:44292", + "issue:44387", + "issue:44402", + "issue:44488", + "issue:44756", + "issue:44910", + "issue:44912", + "issue:44960", + "issue:44991", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:46050" + ], + "soft_pairs": [ + "issue:43708|issue:44075", + "issue:39692|issue:43994", + "issue:45488|issue:45812", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:43856|issue:43873", + "issue:42491|issue:43931", + "issue:43644|issue:43645", + "issue:43856|issue:44387", + "issue:42491|issue:45292", + "issue:44488|issue:44991", + "issue:44910|issue:45588", + "issue:45362|issue:45406", + "issue:44164|issue:44291", + "issue:45276|issue:45335", + "issue:36010|issue:39404", + "issue:43425|issue:43957", + "issue:43404|issue:43454", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:45663|issue:45759", + "issue:45663|issue:45859", + "issue:43873|issue:44387", + "issue:44155|issue:45823", + "issue:44291|issue:44756", + "issue:43335|issue:45663", + "issue:44292|issue:44912", + "issue:44960|issue:46050", + "issue:44387|issue:44756", + "issue:42947|issue:43381", + "issue:45276|issue:45292", + "issue:43854|issue:45216" + ], + "prepared_review_unit_hash": "813553c9573c56a14ea9c10c98698b7aa6e8c4b1168ce41eb9cdc8167fd1961f", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37734, + "estimated_input_tokens": 9434, + "estimated_eval_tokens": 19124 + }, + "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. The only pair that looks like a true duplicate is the resize_token_embeddings regression affecting output embeddings in Gemma4/generic form (45276 and 45292).", + "confidence": 0.71, + "canonical_issue_reason": "issue:45276 is the most representative issue because it describes the broad resize_token_embeddings failure and includes the concrete affected model plus the shared output-embedding symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45276 is the best single issue to keep: it is the clearest, most actionable representative of the resize_token_embeddings regression and has the strongest duplicate overlap with another item.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Different training bugs: checkpoint resumption step \u0a97\u0aa3culation vs optimizer arguments being ignored." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a broken documentation example while the other is a model/pipeline output regression." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families and different failure modes." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model-loading failures for different models; no clear shared underlying bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to GLM-4.7-Flash loading." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention memory/quantization, but one is training-memory inefficiency and the other is offloading not working under quantization." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "LoRA compatibility across hf versions is not the same issue as a Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Both are Transformers 5 regressions, but one is non-persistent buffers and the other is notebook custom-model initialization." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Related to memory, but the concrete problems differ: Qwen3 MoE training memory vs int4 quantization CUDA reservation/OOM." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Different problems: LoRA model compatibility versus token-embedding resize propagation." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/load failures, but the affected models and likely root causes differ." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Different flash-attention crashes with different root causes and code paths." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different runtime failures: chat invocation crash vs serve processor attribute error." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Serialization extra_state handling is unrelated to init_empty_weights/_is_hf_initialized initialization." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Same API area, but the affected models/fields differ enough that this looks model-specific rather than one concrete duplicate bug." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin and Whisper return_language pipeline regression are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "General Torch 2.10 incompatibility is not the same as meta-device model-loading breakage." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both concern untied lm_head weights, but they appear to be separate model-specific bugs rather than one concrete duplicate." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Documentation/example errors are unrelated to the GLM5 issue." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both mention tokenizer/vocab loading, but they are different models and likely different causes." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA and an apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "FSDP2 KeyError in Gemma4 attention is unrelated to Qwen multimodal text_config unwrapping." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different model families and different offload/path failures." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Quantization/offloading issues, but not the same bug or change." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 batched inference leak is unrelated to Gemma4 device-map mismatch." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Empty-weights initialization and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Sparse-layer creation in SwitchTransformers is unrelated to Gemma4 FSDP2 attention state rebuilding." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and different load failures; not the same underlying bug." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 is unrelated to sinq save/load breakage." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Different OOM causes and different environments." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing and the other is eval-mode restriction." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "These describe the same resize_token_embeddings regression: output embeddings are not updated, with 45276 being the model-specific variant and 45292 the generic one." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different model-loading/save-pretrained issues affecting different models and versions." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The 45276/45292 duplicate call is well supported by the packet, and the other soft-edge rejections stay conservative without overstating evidence. The summary is mostly accurate; at most, 45276/45335 is a borderline similarity, but not enough in the packet to overturn the analyst\u2019s conservative split." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 7, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:41720", + "issue:43257", + "issue:43381", + "issue:43425", + "issue:43452", + "issue:43596", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43901", + "issue:43906", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44155", + "issue:44164", + "issue:44265", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44928", + "issue:45084", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45823" + ], + "soft_pairs": [ + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:44479|issue:44877", + "issue:45588|issue:45663", + "issue:41720|issue:44155", + "issue:45684|issue:45698", + "issue:43957|issue:44756", + "issue:43646|issue:43708", + "issue:43645|issue:44164", + "issue:45292|issue:45335", + "issue:43828|issue:43957", + "issue:43425|issue:43828", + "issue:44402|issue:44451", + "issue:44589|issue:44756", + "issue:44683|issue:45823", + "issue:43257|issue:44863", + "issue:39404|issue:45701", + "issue:43425|issue:45823", + "issue:43611|issue:43645", + "issue:43425|issue:43716", + "issue:44530|issue:45084", + "issue:43906|issue:44530", + "issue:39404|issue:44912", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:43425|issue:44918", + "issue:43901|issue:44509", + "issue:44589|issue:45084", + "issue:44164|issue:45759", + "issue:36032|issue:43452", + "issue:43596|issue:44683", + "issue:43994|issue:44079" + ], + "prepared_review_unit_hash": "dae3557c8203ff772db92fee53f59acae40748a6b95242ee6697011b169175a6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36883, + "estimated_input_tokens": 9221, + "estimated_eval_tokens": 18698 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous Transformers regression reports with one clear duplicate pair: `resize_token_embeddings` not updating output embeddings (`issue:45292` / `issue:45335`). The other soft-similarity pairs are related at a high level but describe different failures or code paths, so they should stay separate.", + "confidence": 0.76, + "canonical_issue_reason": "`issue:45292` is the clearest canonical for the only true duplicate pair: it states the generic `resize_token_embeddings` output-embedding mismatch, while `issue:45335` is a model-specific instance (`t5gemma`/decoder.embed_tokens).", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:43611` is the strongest single representative issue overall: it is a broad, well-scoped Transformers 5.0.0 loading regression affecting model initialization/loading, which matches the cluster\u2019s main theme better than the narrower model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Both mention training/runtime instability, but one is about gradient checkpointing in eval mode and the other about Qwen3.5 RLHF NaNs from 3D position ids / SDPA fallback; different bug paths." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "One is about `extra_state` save/load handling, the other about disabling mmap on Strix Halo to avoid OOM; unrelated mechanisms." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models vs strict config blocking `granite_speech`; different model families and failure modes." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention crash on `s_aux=None` for sink-less models is a different attention-path bug than Gemma4 FSDP2 `shared_kv_states` rebuilding." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Auto device mapping CUDA assert on Qwen3 vs AudioFlamingo3 batched inference token/embedding leakage; not the same underlying defect." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Read-only permission propagation in `save_pretrained` is a packaging/filesystem issue; wrong custom module loaded after save is an auto-class/module resolution issue." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading breakage on some models vs mmap OOM on Strix Halo; shared theme is loading, but the root causes differ." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization regression vs Trainer checkpoint-resume `max_steps` miscalculation; separate trainer/loading code paths." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook-specific custom model init breakage vs `save/from_pretrained` failing to handle `extra_state`; different subsystems." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same underlying regression: `resize_token_embeddings` updates one embedding path but not the output/decoder embedding; `t5gemma` is a model-specific manifestation of the same bug." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Both involve dtype/device behavior during loading or inference, but one is an autocast mismatch in Phi-tiny-MoE and the other is meta-device loading failures across models." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility is a broad environment issue; the other is a specific autocast dtype mismatch on a particular model." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab-size mismatch for one model vs inability to load a different model at latest version; too little evidence they share one code path." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage deserialization error is unrelated to mmap/OOM on Strix Halo." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 vs Gemma4 PLE device mismatch under `device_map=auto`; different layers and bugs." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed is distinct from NemotronH checkpoint loading failure." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` regression and tokenization changing across transformers versions are different features with different symptoms." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 compatibility is a general dependency issue, while Gemma4 PLE device mismatch is a model-forward placement bug." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both concern Transformers 5 custom model handling, but one is about `base_model_prefix` loading and the other about notebook redefinition/init behavior; not the same concrete bug." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 incompatibility vs Mistral-3 image preprocessor/model dtype mismatch; different scope and root cause." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache crash on Qwen3.5 linear_attention vs compile-time template-node error; unrelated failures." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "An isolated reproduction of an older issue vs a specific Qwen3.5 cache crash; no evidence they are the same code-path bug." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline `return_language` and MXFP4 quantization fallback are unrelated model/pipeline behaviors." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "`torch.export`/`torch_compilable_check` failure vs Perceiver non-default resolution interpolation failure; different APIs." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression in several Qwen VL models vs tensor mask shape mismatch; not the same failure mode." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility vs Qwen3.5 embedding unpacking with TRL SFT trainer; different problem domains." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Docs drift about `return_all_scores` vs docs mentioning removed pipeline tasks; documentation-only issues, but not the same missing/changed content." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure vs compile template node error; unrelated symptoms." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "`extra_state` save/load handling vs `AutoModelForCausalLM.from_config` not unwrapping `text_config`; separate model-construction paths." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5Tokenizer special-token method conflict vs `gguf_file` breaking `from_pretrained`; no shared bug mechanism." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed ZeRO-3 init index error in BertModel vs compiled flex_attention breaking on torch>=2.9; unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 nonsensical outputs with AutoModel/pipeline vs `ModelOutput` key assignment bug; different layers of the stack." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge calls. The lone duplicate pair (`issue:45292` / `issue:45335`) is a reasonable canonical merge, and the other pairs are rejected without overclaiming shared root cause. No evidence of overstated similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 8, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:32090", + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42915", + "issue:43299", + "issue:43317", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43856", + "issue:43883", + "issue:43931", + "issue:44164", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44464", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45584", + "issue:45859", + "issue:45987" + ], + "soft_pairs": [ + "issue:43299|issue:43931", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:44737|issue:45859", + "issue:44164|issue:44479", + "issue:43582|issue:43632", + "issue:44488|issue:45081", + "issue:44661|issue:44683", + "issue:43611|issue:43644", + "issue:32090|issue:39404", + "issue:40990|issue:43421", + "issue:43716|issue:45084", + "issue:44488|issue:44987", + "issue:36032|issue:42915", + "issue:43643|issue:44530", + "issue:34567|issue:40444", + "issue:43883|issue:44488", + "issue:44811|issue:45584", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:42915|issue:44530", + "issue:43644|issue:43646", + "issue:44464|issue:45987", + "issue:44206|issue:44220", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:36032|issue:43645", + "issue:44849|issue:45325" + ], + "prepared_review_unit_hash": "ab75c416dd885b5860c0d286f9cecb61bb033a71a62b9ebddfe4363de662ad4c", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36003, + "estimated_input_tokens": 9001, + "estimated_eval_tokens": 18258 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose semantic grouping of Transformers v5 regressions across many unrelated models/components. I don\u2019t see any soft pair that clearly describes the same underlying bug, so I\u2019d keep the issues separate and use one broad model-init regression as the representative issue.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 43646 is the broadest, most central representative of the shared v5 regression theme around model initialization/loading; the other items are mostly model-specific or unrelated subsystems.", + "canonical_pr_reason": null, + "best_issue_reason": "43646 is the best single issue to anchor this cluster because it is general enough to represent the common Transformers v5 breakage pattern without being tied to one model or backend.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading failures, but they involve different model families and different failure modes (FineGrainedFP8Config vs weight-shape mismatch)." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both affect tokenizer loading, but one is an Mistral regex patch crash and the other is a specific tokenizer regression for EMBEDDIA/est-roberta; not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Both mention memory/OOM, but one is Qwen3 MoE training memory usage and the other is Strix Halo mmap OOM; different code paths and fixes." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction corruption is not the same issue as tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM prediction bug are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both are device/offload related, but one is XLNet CPU allocation and the other is Qwen3.5 MoE rotary embedding CPU-offload compatibility; different bugs." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Save/from_pretrained extra_state handling and video-input regression are distinct problems." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized flag regression." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Both are loading failures, but the models and root causes differ; one is a specific repository loading issue and the other is a Mistral tokenizer patch crash." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizers mapping/model registration failure is unrelated to compiled flex_attention failing on torch >= 2.9." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Base_model_prefix loading regression and non-persistent buffer junk initialization are separate v5 regressions." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error is unrelated to Whisper return_language behavior." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and runtime post-processor updating for special tokens are not the same bug." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor dtype mismatch and template compilation error are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are model-loading regressions, but they target different repositories and likely different code paths." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer special-tokens conflict is unrelated to Qwen3Moe FineGrainedFP8Config failure." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss and PagedAttentionCache invalid group type on Qwen3.5 models are different issues." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counter update and Qwen2.5-VL iterable dataset multi-image failure are unrelated." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is not the same as a tokenizer/model loading failure for cjvt/sleng-bert." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both are Whisper-related, but one is batch_decode skip_special_tokens handling and the other is empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Two separate model-loading regressions for different repositories; no evidence they share the same underlying fix." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing failure for Mixtral is unrelated to the template node compilation TypeError." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache linear_attention crash are different model/backend problems." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Non-persistent buffer junk and custom model initialization regressions are both v5 issues, but not the same defect." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward and StaticCache seq-length return type mismatch are distinct cache/compilation issues." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "The generic _torch_extract_fbank_features issue and the LasrFeatureExtractor center-arg regression are related in area but not clearly the same bug." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both mention GGUF loading regressions, but one is a broad from_pretrained breakage and the other is a specific Qwen2.5-GGUF failure; not enough to merge." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "device_map/offload loading failure and GGUF loading failure are different code paths." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method-name conflict and Jupyter custom-model initialization regression are unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug and Qwen2.5-VL temporal position-id scaling bug are separate model-specific regressions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 9, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36032", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43089", + "issue:43262", + "issue:43296", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43540", + "issue:43611", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44062", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45507", + "issue:45987", + "issue:46178" + ], + "soft_pairs": [ + "issue:42175|issue:46178", + "issue:45084|issue:45507", + "issue:43883|issue:44991", + "issue:43262|issue:43377", + "issue:44987|issue:44991", + "issue:43089|issue:45987", + "issue:43646|issue:44534", + "issue:43322|issue:43540", + "issue:44403|issue:45081", + "issue:43404|issue:44466", + "issue:43994|issue:44610", + "issue:42915|issue:45084", + "issue:43645|issue:44534", + "issue:43377|issue:43819", + "issue:44265|issue:44610", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:44220|issue:44683", + "issue:45216|issue:45357", + "issue:36032|issue:43742", + "issue:42915|issue:44291", + "issue:43632|issue:44756", + "issue:42915|issue:44589", + "issue:44062|issue:44987", + "issue:43296|issue:43531", + "issue:43404|issue:45127", + "issue:43531|issue:44877", + "issue:43611|issue:44534", + "issue:36032|issue:41720", + "issue:43425|issue:44403", + "issue:44479|issue:45325", + "issue:40990|issue:44811" + ], + "prepared_review_unit_hash": "14e890e2ced348dd75dc27e37439f4088c9734087c9913acabc798a0ddb26b2f", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37300, + "estimated_input_tokens": 9325, + "estimated_eval_tokens": 18906 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated and should stay separate. The only solid duplicate-like match is the Qwen3.5 `save_pretrained` regression pair (`issue:45216` / `issue:45357`). The Qwen-VL video/rope-index pair (`issue:44479` / `issue:45325`) is too weakly supported from titles alone and should not be accepted as the same bug.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 45216 is the best canonical issue anchor because it describes the broader Qwen3.5 `save_pretrained` regression and has the clearest duplicate-style follow-up in 45357.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45216 is the strongest overall issue representative because it has the cleanest confirmed overlap and a broader regression framing than its duplicate follow-up.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Different backend-mapping failures: missing TensorFlow backend in a torch install vs an offending tensorflow_text backend entry." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Unrelated failures: template-node compilation error vs an invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Different model-loading bugs with different symptoms and causes; not the same concrete regression." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Both are audio-related, but sampling-rate defaults and MIMI batching/padding-mask support are separate code paths." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer/model loading regressions affecting different models and mechanisms." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead is unrelated to the StaticCache return-type contract issue." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization in notebooks and non-persistent buffer junk are separate v5 regressions." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both are multimodal loading issues, but Llava Next segfault and Qwen3OmniMoe video ValueError are different bugs." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Noise while loading a transformer is not the same concrete failure as the Mistral regex patch crash." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both involve tied weights, but one is missing tying in a model class and the other is serialization inconsistency." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Different subsystems: SigLIP2 inference quality vs OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FP8 config loading failure and non-template-node compile failure are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Notebook custom-model initialization breakage is not the same as non-persistent buffer corruption." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Different audio/model numeric bugs affecting different code paths." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export compatibility with torch_compilable_check is unrelated to the OmDet processor resolution mismatch." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer-loading failures for different models and different causes." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Loading physical-intelligence/fast is unrelated to the Mistral regex patch crash." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "fbank feature extraction and compiled flex_attention on torch>=2.9 are unrelated." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same underlying Qwen3.5 `save_pretrained` regression; 45357 is a specific manifestation of the broader issue in 45216." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and MobileLLM key error are different loading failures." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "FineGrainedFP8Config failure and unexpected _is_hf_initialized argument are separate v5 issues." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "_is_hf_initialized flag breakage and mmap OOM avoidance are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "FP8 config failure is not the same as missing Float8 storage object resolution." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken special-argument conflict and model loading failure for physical-intelligence/fast are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window behavior are different bugs." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both mention embeddings/tied weights, but missing tie in Mistral3ForConditionalGeneration is not the same as LoRA merge collapse with extended vocab." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Qwen-VL-related titles alone are not enough to show the same underlying bug: one is a broad video-input regression, the other a still-image temporal position-id scaling issue." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on GPT-OSS is unrelated to Whisper batch_decode ignoring skip_special_tokens." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The review is mostly conservative, but one accepted soft-edge verdict overstates the evidence. `issue:44479` and `issue:45325` are both Qwen-VL-related, yet the packet only shows titles: one is a broad video-input regression across several Qwen VL variants, while the other is a still-image temporal-position-id scaling bug in Qwen2.5-VL. That is enough to say they may be related, but not enough to confidently call them the same underlying regression or a duplicate symptom. The Qwen3.5 `save_pretrained` pair (`issue:45216` / `issue:45357`) looks better grounded. Because the summary and canonical-anchor reasoning lean on the stronger claim for `44479` / `45325`, I would not accept as-is." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 10, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:33666", + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43381", + "issue:43421", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43646", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44829", + "issue:44898", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45834", + "issue:45923" + ], + "soft_pairs": [ + "issue:43792|issue:44589", + "issue:33666|issue:40444", + "issue:41553|issue:43582", + "issue:43632|issue:44534", + "issue:44683|issue:45020", + "issue:44403|issue:44991", + "issue:43421|issue:44568", + "issue:44079|issue:45663", + "issue:43381|issue:45923", + "issue:40444|issue:45325", + "issue:44291|issue:44589", + "issue:44462|issue:45812", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:43317|issue:43856", + "issue:42175|issue:45542", + "issue:43482|issue:43525", + "issue:43818|issue:45357", + "issue:41720|issue:45081", + "issue:43377|issue:45491", + "issue:43531|issue:44291", + "issue:44265|issue:45072", + "issue:42175|issue:43296", + "issue:43611|issue:43646", + "issue:43632|issue:45072", + "issue:43792|issue:43957", + "issue:45325|issue:45834", + "issue:43994|issue:44898", + "issue:44079|issue:44534" + ], + "prepared_review_unit_hash": "b4274158f2e1b3a16eee3976b2ca3ce96bbc07e76aa1c4d2cf55160d5d4d0368", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36040, + "estimated_input_tokens": 9010, + "estimated_eval_tokens": 18276 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of separate model- or backend-specific regressions that only share superficial themes like \u201cv5 breakage,\u201d \u201cloading fails,\u201d or \u201ctokenizer/model init issues.\u201d I did not find any soft pair that looks like the same underlying bug/change.", + "confidence": 0.69, + "canonical_issue_reason": "issue:45020 is the best umbrella issue here: it\u2019s the broadest, most central report about recent Transformers versions breaking remote-code models, which matches the cluster\u2019s overall pattern of v5 regressions better than the narrower model-specific tickets.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best representative issue for this mixed set because it captures the widest class of affected failures and is the most plausible anchor for related downstream regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Both are model-loading failures, but one is Whisper-specific and the other is a float8 storage/type error; they are different concrete bugs and code paths." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both involve Qwen VL training, but one is generic multi-GPU training and the other is an IterableDataset/multi-image prompt failure; not the same bug." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "These are unrelated: Voxtral tokenizer error messaging versus an Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is about _is_hf_initialized and the other about non-persistent buffers; different mechanisms." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Torch 2.9 compiled flex_attention failure is unrelated to remote_code model breakage." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading-related, but one is generic loading noise and the other is a specific tokenizer-loading regression for one model family." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both touch special-token behavior, but one is runtime post-processor updates and the other is add_special_tokens not adding BOS/EOS; not the same concrete issue." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "Unrelated bugs: ModelOutput key assignment versus Gemma-4 FSDP2 shared_kv_states rebuilding." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing in eval mode is a different issue from a missing supports_gradient_checkpointing flag on a trust_remote_code model." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL issues, but one is multi-image IterableDataset training and the other is rope-index scaling for still images; different code paths." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Both involve loading/initialization regressions, but one is an unexpected _is_hf_initialized argument and the other is a float8 storage lookup failure." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Both concern AutoTokenizer behavior, but one ignores tokenizer.json while the other produces wrong token IDs for Granite models; not the same bug." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure and CLIPTokenizer max-length behavior are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both cause bad training dynamics, but one is flash_attention_3 degeneration and the other is single-label classification with num_labels=1; different root causes." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Device-map/offload loading failure is unrelated to Qwen3 MoE memory inefficiency during training." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both are packaging/backend issues around TensorFlow, but one is missing TF in a torch extra install and the other is tensorboard installed without TF causing backend errors; not clearly the same bug." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Different model-loading problems: Qwen2.5-GGUF failure versus Llama4Config missing pad_token_id." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Completely different model families and failures: Video-LLaVA tower issues versus Qwen3.5 save_pretrained key regression." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "An A800 cudaErrorAssert in Qwen3 device mapping is unrelated to a Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both mention padding/masking-related batching issues, but they affect different models and different failure modes; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the _is_hf_initialized init_empty_weights regression." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Torch export failures with torch_compilable_check and dtype mismatches in bfloat16 inference are different code paths and bugs." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend dependency issues versus a vLLM model-loading failure for PaddleOCR-VL are not the same problem." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are Transformers v5 regressions in model initialization/loading, but the reported failures are different enough that they do not look like one concrete bug." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized regressions and bfloat16 dtype mismatches are unrelated." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper loading failure and torch.device(\"meta\") loading breakage are separate model-init issues with different symptoms." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL temporal position scaling and Kosmos2.5 long OCR index errors are different multimodal bugs." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 incorrect outputs with AutoModel/pipeline and Perceiver interpolation failures are unrelated model-specific issues." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer filling are different internal regressions with no clear 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": 17, + "review_unit_index": 11, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43644", + "issue:43701", + "issue:43746", + "issue:43761", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:44079", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44568", + "issue:44610", + "issue:44617", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45325", + "issue:45381", + "issue:45584", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45920", + "issue:45923" + ], + "soft_pairs": [ + "issue:43454|issue:44466", + "issue:40444|issue:45381", + "issue:42915|issue:43482", + "issue:45325|issue:45381", + "issue:43644|issue:44079", + "issue:43296|issue:45800", + "issue:43761|issue:44079", + "issue:44479|issue:45823", + "issue:35707|issue:40444", + "issue:43606|issue:44265", + "issue:44291|issue:44877", + "issue:44387|issue:44617", + "issue:43881|issue:44488", + "issue:43701|issue:45325", + "issue:44568|issue:45920", + "issue:43606|issue:43819", + "issue:41720|issue:45823", + "issue:44186|issue:44265", + "issue:43454|issue:43550", + "issue:43262|issue:43746", + "issue:39692|issue:40444", + "issue:40990|issue:45584", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:43572|issue:45759", + "issue:43257|issue:45923", + "issue:40990|issue:42947", + "issue:43377|issue:44898", + "issue:43317|issue:43873", + "issue:43873|issue:45701" + ], + "prepared_review_unit_hash": "1ac53a031010810160be5b850d2ee3fb59724744ccfdfe89b9866a65a448a4e5", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated model- or subsystem-specific regressions that only share broad vocabulary (tokenization, quantization, device_map, Qwen/VL, etc.). I found no pair that looks like the same concrete bug/change, so the safe triage result is to keep them separate.", + "confidence": 0.96, + "canonical_issue_reason": "issue:43873 is the broadest concrete regression in the set: it is open, has inbound references, and clearly describes an offloading/quantization failure path, making it the most plausible representative issue for this heterogeneous cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best global representative because it is reproducible, reasonably specific, still open, and has the strongest signal of being a user-facing regression among the listed issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Related theme (tied lm_head weights), but one is a config/tie-weights generation bug and the other is a serialization inconsistency; not the same concrete failure." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Different modalities and failure modes: iterable-dataset multi-image finetuning vs video position-id regression." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Completely different models and bugs: FP8 config failure vs GGUF loading failure." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both touch Qwen2.5-VL position handling, but one is still-image temporal scaling and the other is video vision_position_ids; too different to merge as one bug." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Unrelated: non-persistent buffers filled with junk vs ModelOutput key assignment behavior." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "Different layers and symptoms: model loading in vLLM vs torch/transformers version incompatibility." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "Hidden_states being dropped in CLIPVisionModel is unrelated to ModelOutput key assignment semantics." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Different regressions: video-input handling in Qwen VL models vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Both are generation/training-related, but one is progressive generation with inputs_embeds/past_key_values and the other is multimodal finetuning with multiple images." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch is not the same bug as torch.export failures around torch_compilable_check." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Different bugs: unexpected _is_hf_initialized argument vs strict config blocking granite_speech loading." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is reserved-memory growth under int4 quantization and the other is a model-specific CUDA OOM." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are model loading failures, but they involve different models and likely different root causes." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL rope/position-id handling." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizer/model families and appear to be separate defects." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Device offload mismatch is unrelated to DAC.from_latents not matching forward pass." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both involve device_map, but the concrete failure modes and models differ too much to treat as one bug." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer crashes on NER/batching and torch.export with torch_compilable_check are unrelated issues." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Different problems: tied weights / garbage generation vs torch.compile+SDPA failure." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting is unrelated to PEFT checkpoint loading for GraniteSpeech." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 are unrelated to Qwen2.5-VL multi-image finetuning failures." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Different domains and root causes: perplexity regression on GPT-OSS vs Whisper empty-transcription generation failure." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Missing padding-mask support in MIMI encoder is unrelated to OmDet-Turbo processor/model input-size mismatch." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Both involve model behavior under certain setups, but one is tied weights and the other is dtype mismatch in inference." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Unrelated: missing pad_token_idx in StableLmConfig vs composite multimodal config unwrapping." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is a different failure than missing supports_gradient_checkpointing on a trust_remote_code model." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "Perplexity regression and gradient-checkpointing/LoRA ineffectiveness are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Different models and bugs: MIMI padding-mask output mismatch vs Perceiver non-default-resolution failure." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both concern offload/quantization, but one is loading a dequantized model with device_map=auto and the other is a broader offloading-not-working report; not the same concrete bug." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Broad offloading/quantization issue vs tokenization changes; no shared concrete failure 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": 17, + "review_unit_index": 12, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43749", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44008", + "issue:44186", + "issue:44387", + "issue:44423", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44610", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44918", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46050" + ], + "soft_pairs": [ + "issue:43526|issue:43883", + "issue:43454|issue:43720", + "issue:43881|issue:44991", + "issue:43296|issue:44991", + "issue:41720|issue:44977", + "issue:44466|issue:45357", + "issue:42617|issue:42915", + "issue:35707|issue:41720", + "issue:44008|issue:45823", + "issue:43577|issue:43708", + "issue:34634|issue:35707", + "issue:44877|issue:44991", + "issue:43450|issue:45701", + "issue:44186|issue:45072", + "issue:43550|issue:43720", + "issue:43404|issue:45357", + "issue:36331|issue:42915", + "issue:36331|issue:42617", + "issue:44423|issue:45406", + "issue:44387|issue:45704", + "issue:43873|issue:44756", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:43749|issue:44534", + "issue:44186|issue:44610", + "issue:45632|issue:45684", + "issue:43322|issue:46050", + "issue:43295|issue:45701", + "issue:42915|issue:43317", + "issue:44479|issue:44918", + "issue:41720|issue:43322", + "issue:40990|issue:45701", + "issue:41720|issue:42617", + "issue:41720|issue:45161" + ], + "prepared_review_unit_hash": "0cf55e411c7c99347dce64e8fdf7ada79a028d09dbd6e9ab99b30e8975fdfa82", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37954, + "estimated_input_tokens": 9489, + "estimated_eval_tokens": 19234 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a mixed bag of mostly unrelated Transformers issues. The closest shared theme is loading/offloading/quantization regressions, but none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43873 is the broadest representative of the loading/offloading + quantization subgroup and is already open with active discussion/inbound references.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 best matches the cluster\u2019s dominant technical theme and is the most plausible umbrella representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Different subsystems and failures: image processor label reduction vs missing tied-weights attribute on a model." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading, but they target different models and distinct code paths (weight tying vs packed-weight unpacking)." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated failures: one is a model load failure, the other is tokenizer loading for a different model family." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "Different stack and symptom: vLLM/PaddleOCR-VL load failure vs tokenizer regression for est-roberta." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Same model family, but different bugs: device-map CUDA assert vs flash-attention generation regression." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both touch save/serialization, but one is lm_head tying/serialization by device and the other is incorrect visual encoder keys." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Different failures entirely: 3d_parallel.py execution vs Qwen3Moe FP8 config load." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device-map CUDA asserts." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma-related, but the bugs are different: variable name collision vs device mismatch during forward." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Different areas: dtype handling in Blip2 vs Trainer resume/max_steps calculation." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset issue is unrelated to progressive generation with embeddings and cache state." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Different config/loader problems for different model types; no shared underlying defect." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batching shape bug is unrelated to a general tokenization change report." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer padding/truncation crash and bf16 dtype mismatches are different bugs." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Different code paths: torch.compile/SDPA behavior vs accelerate loading of packed BitNet weights." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Weight tying at model init is not the same as incorrect save_pretrained key serialization." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer API signature mismatch vs model loading failure; unrelated issues." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Unexpected trainer keyword argument is unrelated to 3d_parallel execution failures." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Continuous-batching multimodal crash and missing processor attribute are different failures." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both are memory-related, but one is quantization OOM and the other is an apex RMSNorm memory leak." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both involve memory/offload concerns, but one is quantization offloading and the other is mmap on Strix Halo causing OOM." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Different models and mechanisms: SDPA/torch.compile vs CPU-offload device mismatch." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both concern image resolution, but one is a processor output-size bug and the other is a Perceiver interpolation failure." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk are distinct loading/runtime issues." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash vs vision processor size mismatch; unrelated." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Both are filesystem/path related, but cache-path collisions and read-only permission propagation are different bugs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next load segfault and sinq save/load breakage are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "The former is a specific processor/tokenizer regression; the latter is too broad and not clearly the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FP8 load failure and dequantized device_map offload failure are different quantization/load paths." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Video-input regression for several Qwen models is not the same as TRL unpacking of input embeddings." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Different model families and error modes: device-map assert vs segmentation fault on Llava Next load." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity anomaly on GPT-OSS is not the same as a generic tokenization-change report." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto device mapping and 3d_parallel execution are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both mention large-model parallelism, but the bugs are different models and different parallelism 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": 17, + "review_unit_index": 13, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:42617", + "issue:42915", + "issue:43316", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:43927", + "issue:44186", + "issue:44361", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44671", + "issue:44743", + "issue:44898", + "issue:44913", + "issue:45059", + "issue:45072", + "issue:45081", + "issue:45406", + "issue:45464", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812" + ], + "soft_pairs": [ + "issue:36010|issue:42617", + "issue:43526|issue:43881", + "issue:45081|issue:45812", + "issue:43531|issue:43572", + "issue:43316|issue:44913", + "issue:43540|issue:43782", + "issue:44671|issue:44743", + "issue:44186|issue:44361", + "issue:43421|issue:45812", + "issue:29159|issue:36032", + "issue:45406|issue:45464", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:43421|issue:45701", + "issue:42915|issue:43322", + "issue:43606|issue:43653", + "issue:43531|issue:45759", + "issue:44492|issue:44509", + "issue:34567|issue:35707", + "issue:43388|issue:45059", + "issue:44514|issue:44545", + "issue:42490|issue:43329", + "issue:43873|issue:45704", + "issue:41720|issue:43746", + "issue:43421|issue:43927", + "issue:44898|issue:45072" + ], + "prepared_review_unit_hash": "1a2a74180d357118031690d9d6db50b2d7f9ec9a7662ce21466dc6979164a313", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35910, + "estimated_input_tokens": 8978, + "estimated_eval_tokens": 18212 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the only clear duplicate is the Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False (issues 44514 and 44545). The other soft pairs look like adjacent-but-distinct bugs affecting different models, code paths, or failure modes.", + "confidence": 0.95, + "canonical_issue_reason": "Pick issue:44514 as the canonical issue: it is the earlier, clearer report of the same exact batched padding=False crash later repeated in issue:44545.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue:44514 is the strongest representative because it has the same concrete repro as the duplicate, with a precise API/method target and a clear failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "ImportError for GenerationMixin and a 3d_parallel execution failure are different problems with no shared code path evidence." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are unrelated model/component issues." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Mistral regex patch crash in tokenizer loading and Granite token-id regression are different tokenizer bugs." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx regression are distinct config/model issues." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload regression are separate config serialization problems." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Video-input ValueError in Qwen3OmniMoe and Qwen3VL weight_only load error are different failure paths." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked LM regression and Qwen3.5 recurrent-state reset bug affect different models and mechanisms." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer/task failures, but one is LayoutLMv2 NER padding/truncation and the other is MLuke task AttributeError; not the same bug." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Runtime post-processor update for special tokens and Granite AutoTokenizer wrong IDs are related to token handling but are not the same concrete defect." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is not the same as T5Tokenizer's add_special_tokens method conflict on load." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are unrelated processor/runtime bugs." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation map processing and BarkProcessor voice_preset failure are different processor issues." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and BigBird mask-token decode issue do not share the same code path." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "CPU offload device mismatch for bark-small and BitNet packed-weight unpacking during accelerate loading are separate loading/offload bugs." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Special-token post-processor refresh and generic tokenization-version changes are too broad and do not indicate one bug." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated model-loading problems." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "bark-small CPU offload mismatch and BigBirdTokenizer mask-token registration are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen3 variants, but one is sliding_window behavior and the other is from_config text_config unwrapping; different code paths." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategies typo and docs mentioning removed pipeline tasks are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counting and progressive generation with inputs_embeds/past_key_values are distinct issues." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "gather_for_metrics label dropping and SAM3 text/bounding-box behavior are different subsystems." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same exact crash: Qwen2_5_VLProcessor.apply_chat_template on batched input with padding=False, reported twice." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Local-path trailing-slash import warning and multimodal token-count video branch bug are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and T5 apex FusedRMSNorm memory leak are different problems." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and GraniteSpeech PEFT local checkpoint loading are not the same bug." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both touch token IDs, but one is runtime post-processor updates while the other is save/load loss of custom token IDs." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation failure and bf16 dtype mismatch in SwitchTransformers/TimmWrapperModel are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and appropriately conservative. The only clear duplicate is issue:44514 / issue:44545, which have identical titles and matching crash context. The other soft-edge rejections stay on the cautious side and do not overstate duplication evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 14, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:30064", + "issue:33666", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43525", + "issue:43531", + "issue:43596", + "issue:43653", + "issue:43720", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:45923" + ], + "soft_pairs": [ + "issue:45759|issue:45859", + "issue:44451|issue:44991", + "issue:33666|issue:35707", + "issue:43531|issue:44589", + "issue:45588|issue:45759", + "issue:44987|issue:45042", + "issue:44062|issue:44451", + "issue:44368|issue:45357", + "issue:39692|issue:44610", + "issue:43421|issue:43653", + "issue:45704|issue:45923", + "issue:45663|issue:45923", + "issue:44918|issue:44936", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:34634|issue:41720", + "issue:45704|issue:45859", + "issue:43421|issue:44186", + "issue:43525|issue:43531", + "issue:43381|issue:45704", + "issue:43782|issue:45381", + "issue:45081|issue:45356", + "issue:39692|issue:43873", + "issue:44625|issue:45479", + "issue:43994|issue:44671", + "issue:43596|issue:44220", + "issue:45356|issue:45907", + "issue:30064|issue:36010", + "issue:43720|issue:43819" + ], + "prepared_review_unit_hash": "1d9e1b3e17328b089b98f3816f881e344fcc10e925421570d2bac92e8b234dee", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36091, + "estimated_input_tokens": 9023, + "estimated_eval_tokens": 18302 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly unrelated issue reports across different models/features. Only one soft edge looks like a true duplicate/same-bug pair: the HF_HUB_OFFLINE tokenizer cache failure and the Mistral regex/offline model_info call. Everything else appears to be distinct root causes or different model-specific regressions.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44843 is the clearest root-cause report in the accepted duplicate pair: it identifies the specific offline break in `_patch_mistral_regex` via an unconditional `model_info()` call, which is more actionable than the broader symptom report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the strongest representative issue overall because it names the concrete failing code path and the exact offline mode regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a config unwrapping bug and the other is a CPU-offload rotary embedding bug; different code paths." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but they affect different models and likely different failure causes; too broad to be the same bug." + }, + { + "left": "issue:33666", + "right": "issue:35707", + "accept": false, + "reason": "One is Qwen2-VL multi-GPU training and the other is progressive generation with inputs_embeds/past_key_values; unrelated problems." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to the Float8 storage import/runtime error." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Flash-attention sink-less crash and Qwen multimodal config unwrapping are different failures." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Both are loading/processor regressions, but one is a model loading failure and the other is an image processor torchvision dependency issue." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-argument error is not the same as a model tokenizer loading failure." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a tie_word_embeddings warning during LoRA fine-tuning; the other is a visual-encoder key save regression." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both mention processor/model mismatch, but they concern different models and different concrete failures." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both involve special tokens, but one is a runtime post-processor update feature gap and the other is a BigBirdTokenizer decode bug." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "RMSNorm memory leak in T5 is unrelated to a missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Gemma-4 FSDP2 shared_kv_states bug is unrelated to Nemotron trust_remote_code metadata." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "One is an inputs_embeds/past_key_values unpacking bug; the other is a train/evaluate lifecycle issue." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing in eval mode and abnormal gradient growth under ZeRO-3 are different training issues." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Both describe the same offline tokenizer-loading regression; 44843 gives the specific root cause (`model_info()` called unconditionally in `_patch_mistral_regex`)." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice preset failure and Qwen3 auto device-mapping cuda assert are unrelated." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "RMSNorm memory leak in T5 is unrelated to Qwen3.5 CPU-offload rotary embedding behavior." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime special-token post-processing and LayoutLMv2 NER padding/truncation crashes are different tokenizer issues." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config and Qwen3-MoE sliding_window behavior are distinct config/model bugs." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior is unrelated to T5's apex RMSNorm memory leak." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3VL weight_only loading and Qwen2.5-VL video vision_position_ids are different multimodal regressions." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both mention fix_mistral_regex/tokenizer loading, but one crashes on backend_tokenizer access while the other reports codec handling and warning issues; not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 documentation errors and quantization/offloading behavior are not the same underlying issue." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 and degenerate zero-loss classification across models are related only at a high level, not the same defect." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Incorrect masked-LM predictions in CamemBERT and nonsensical AutoModel/pipeline results for SigLIP2 are different model-specific regressions." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed ZeRO-3 init/BertModel index error is unrelated to `_torch_extract_fbank_features()`." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Kimi tokenizer regression and `list_repo_files` missing revision forwarding are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Void segmentation maps in image processing and GenerationMixin import errors are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading and DAC.from_latents STE mismatch are different code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well-grounded in the packet, and the one accepted duplicate pair (42886/44843) is supported by the titles and conservative reasoning. The other soft-edge rejections are cautious and do not overclaim duplication from broad topical similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 15, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:42915", + "issue:43295", + "issue:43335", + "issue:43388", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43746", + "issue:43756", + "issue:43782", + "issue:43819", + "issue:43873", + "issue:43950", + "issue:44008", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44448", + "issue:44488", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45834", + "issue:45987", + "issue:46032", + "issue:46050" + ], + "soft_pairs": [ + "issue:43746|issue:46050", + "issue:44077|issue:44625", + "issue:43653|issue:43720", + "issue:44220|issue:44661", + "issue:43540|issue:44008", + "issue:43618|issue:43697", + "issue:43550|issue:45072", + "issue:42915|issue:43746", + "issue:43746|issue:43873", + "issue:43295|issue:45834", + "issue:43550|issue:43819", + "issue:45183|issue:45479", + "issue:43388|issue:46032", + "issue:43716|issue:44008", + "issue:43335|issue:43756", + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:43531|issue:43643", + "issue:44062|issue:44488", + "issue:43950|issue:45987", + "issue:44843|issue:45081", + "issue:43643|issue:44877", + "issue:43540|issue:45381", + "issue:30064|issue:36331", + "issue:44448|issue:44671", + "issue:43421|issue:44811", + "issue:39692|issue:45834", + "issue:43782|issue:44977" + ], + "prepared_review_unit_hash": "0ea76ffc1dc20c6c18740151a57c7141b2bbb5450380c1aca0e3a6daf5a573b6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35917, + "estimated_input_tokens": 8980, + "estimated_eval_tokens": 18216 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous. I would reject all soft pairs as duplicates; they share at most a broad subsystem, not the same concrete bug/change. The most representative issue is the broad quantization/offloading report.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43873 is the broadest, most central report in the set, with the most discussion/inbound references and a cross-cutting loading/offloading/quantization scope.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best umbrella issue here because it is active, broadly scoped, and closer to an infrastructure-level regression than the other model-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "Both are loading-related, but one is PEFT adapter loading from local checkpoints and the other is sinq quantization save/load corruption; different models and different failure modes." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Different areas entirely: PatchTSMixer post_init policy vs Qwen3.5 num_labels propagation in text config." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug vs BitNet packed-weight unpacking during accelerate loading; unrelated code paths." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction helper failure vs tokenizer-mapping code generation failure; no shared underlying bug." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different multimodal models and different root causes: Qwen3OmniMoe video processing error vs Gemma 3n variable-name collision." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Both are v5 regressions, but one is CLIP output attentions assignment and the other is RTDetrV2 output drift; not the same concrete bug." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Torch.compile/SDPA failure in Bamba vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel; different symptoms and paths." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and GraniteSpeech PEFT checkpoint loading are unrelated beyond both touching model loading." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Both involve loading/quantization, but one is PEFT checkpoint loading while the other is offloading behavior under quantization; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer API regression vs long-OCR index error in Kosmos2.5; different subsystems and failures." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba compile/SDPA bug is unrelated to DAC.from_latents missing-STE mismatch." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Both concern input/label handling, but one is XOR validation messaging and the other is zero-loss classification logic; distinct bugs." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Distributed gather_for_metrics label truncation is unrelated to Mamba2Mixer use_cache incorrect results." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Different model families and different issues: image preprocessor/model dtype mismatch vs a forward-pass variable collision." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Both are model-configuration mistakes, but they affect different architectures and different misconfigurations." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer NER/padding crash vs image-classification interpolation failure; unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBird tokenizer special-token registration bug has nothing to do with DAC latent decoding mismatch." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE vs AutoConfig trust_remote_code field loss; different components." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument conflict vs model loading failure for cjvt/sleng-bert; not the same root cause." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Both are core API regressions, but one corrupts non-persistent buffers on from_pretrained while the other returns the wrong type from StaticCache.get_seq_length()." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper function, but different bugs: unconditional hub metadata lookup offline vs missing backend_tokenizer attribute crash." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-loading issues, but one is missing fields with trust_remote_code and the other is strict config blocking granite_speech; different failures." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video/multimodal Qwen models, but the concrete bugs differ: video input ValueError vs incorrect vision_position_ids." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Completely unrelated: image processor segmentation-map handling vs CustomTrainer compute_loss signature breakage." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are model output regressions in v5, but on different models and different code paths." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Tokenizer post-processor update bug vs Whisper batch_decode skip_special_tokens handling; related theme, different underlying issue." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 docs example errors and Kosmos2.5 OCR index error are separate model-specific problems." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different Qwen variants and different failures: weight_only load error vs flash-attention generation 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": 17, + "review_unit_index": 16, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36010", + "issue:42175", + "issue:43262", + "issue:43322", + "issue:43404", + "issue:43452", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43650", + "issue:43697", + "issue:43720", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43994", + "issue:44186", + "issue:44442", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44821", + "issue:44857", + "issue:44936", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45406", + "issue:45584", + "issue:45685", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:43577|issue:43818", + "issue:45584|issue:45834", + "issue:44821|issue:45084", + "issue:43322|issue:45685", + "issue:43606|issue:45072", + "issue:44479|issue:45834", + "issue:44186|issue:44442", + "issue:34634|issue:36010", + "issue:43825|issue:43994", + "issue:44561|issue:45020", + "issue:45987|issue:46192", + "issue:43577|issue:43582", + "issue:44485|issue:44509", + "issue:28218|issue:29159", + "issue:44610|issue:45072", + "issue:44805|issue:44936", + "issue:43720|issue:45072", + "issue:43452|issue:44991", + "issue:43825|issue:43827", + "issue:43404|issue:43782", + "issue:43582|issue:44857", + "issue:42175|issue:45081", + "issue:43906|issue:46209", + "issue:45081|issue:45406", + "issue:44671|issue:45701", + "issue:43262|issue:43819", + "issue:43526|issue:43697", + "issue:43650|issue:43906" + ], + "prepared_review_unit_hash": "5a846f64cfdcad9fc4ff2a566bbe40c6db1446198c4d5af9eef366e07c437590", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36112, + "estimated_input_tokens": 9028, + "estimated_eval_tokens": 18312 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous; the soft-linked pairs mostly share only broad subsystem similarity and do not look like true duplicates of the same bug/change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the broadest tokenizer-regression report and the closest thing to a representative anchor, but the cluster overall is too mixed to have a strong canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype remains float32 vs Video-LLaVA missing temporal attention / near-shared towers; different models and failure modes." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty transcription after align_special_tokens vs Kosmos2.5 OCR long-input index error; unrelated modalities and code paths." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading vs Jinja/template compilation error; unrelated loading vs template parsing issues." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next segfault on model load vs MPS histogram not implemented for Int in MOE; different subsystems and symptoms." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch in Bark-small vs bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel; both are device/dtype issues but not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Qwen VL video-input regression vs Kosmos2.5 OCR index error; different models and input pipelines." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2Tokenizer crashes on NER/batched padding vs AutoTokenizer failing to load a FastSpeech2ConformerTokenizer; both tokenizer-related but not the same failure." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset failure vs GenerationMixin import error; completely different APIs." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "pipeline() translation-support error message vs Siglip2 producing nonsensical outputs; docs/error-message issue vs model behavior bug." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both involve remote_code breakage, but one is a specific removed helper (is_torch_fx_available) and the other is a broad recent-version regression; not the same concrete bug." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache seq_length return-type violation vs return_tensors being ignored with explicit text_kwargs; unrelated API behaviors." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype stays float32 vs Apple Silicon caching_allocator_warmup TypeError; no shared code path." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs docs still mentioning removed pipeline tasks; unrelated implementation vs documentation issues." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Additional space after added token vs slow/fast tokenizer inconsistency; related area, but not clearly the same underlying bug." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "OmDet-Turbo image-size mismatch vs dtype mismatch in a different model family; unrelated regressions." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask shape IndexError vs trainer.evaluate() failing after train(); different failure sites and APIs." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed weights not unpacked during accelerate loading vs bfloat16 dtype mismatch in inference; different loading/inference bugs." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file breaking tokenizer/model loading vs a specific tokenizer load failure for EMBEDDIA/est-roberta; both loading regressions, but not the same concrete issue." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both relate to v5 pipeline removals, but one is an error message and the other is stale docs; not the same bug/change." + }, + { + "left": "issue:43404", + "right": "issue:43782", + "accept": false, + "reason": "Mistral3 lm_head weight tying bug vs Qwen3VL weight_only=True from_pretrained error; different models and root causes." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "AppleSilicon allocator warmup TypeError vs CUDA AMP crash in LwDetrImageLoss; unrelated platforms and subsystems." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Missing TensorFlow backend in pip extra vs Mistral regex patch AttributeError; unrelated packaging vs tokenizer internals." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "Isolated reproduction of an old issue vs FP8 inference failure on B200; no shared underlying bug." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral regex patch failing on tokenizer internals vs Gemma4Processor missing _tokenizer in transformers serve; different objects and code paths." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression vs a broad report that tokenization changes across versions; one is model prediction behavior, the other a tokenizer-level complaint." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio processor sampling-rate default mismatch vs DAC.from_latents / forward-pass mismatch; unrelated audio bugs." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug vs RTDetrV2 output changes in v5; both vision-related, but not the same concrete defect." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "Placeholder 'ADD THE DATA' issue vs an isolated reproduction report; no substantive overlap." + } + ] + }, + "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": 17, + "review_unit_index": 17, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43638", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43653", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43749", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:44060", + "issue:44291", + "issue:44368", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44811", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:45020", + "issue:45042", + "issue:45084", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45759", + "issue:45803", + "issue:45859" + ], + "soft_pairs": [ + "issue:43653|issue:44912", + "issue:44060|issue:44368", + "issue:43742|issue:45020", + "issue:45084|issue:45538", + "issue:43329|issue:45600", + "issue:44589|issue:45042", + "issue:43746|issue:44811", + "issue:43644|issue:43749", + "issue:43638|issue:44936", + "issue:43550|issue:44912", + "issue:43819|issue:43927", + "issue:36010|issue:43645", + "issue:34634|issue:36331", + "issue:43854|issue:44977", + "issue:39692|issue:44912", + "issue:43720|issue:44610", + "issue:38175|issue:44610", + "issue:45459|issue:45803", + "issue:43454|issue:43577", + "issue:36010|issue:36331", + "issue:43526|issue:44554", + "issue:43531|issue:45042", + "issue:43262|issue:44811", + "issue:43749|issue:44291", + "issue:43257|issue:43329", + "issue:43643|issue:45759", + "issue:43329|issue:43450", + "issue:43606|issue:45859", + "issue:42490|issue:43257" + ], + "prepared_review_unit_hash": "63d87f296b65d8c95569a02ffa41628308654ad2041bd163889cfda91399d8c6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36175, + "estimated_input_tokens": 9044, + "estimated_eval_tokens": 18344 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is loosely related by Transformers loading/processor regressions, but the soft pairs mostly describe different concrete bugs. I did not find any safe duplicate merges, and there are no PRs in the set.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45020 is the broadest and most central loading/regression report here, covering recent-version breakage around remote/custom model handling that overlaps the most with the rest of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best general exemplar because it is high-level, user-facing, and closest to the shared theme of recent Transformers compatibility regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Different subsystems: tokenizer special-token registration vs MXFP4 quantization loading." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-embedding warnings, but they describe different Qwen warning paths and symptoms." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "MobileLLM key error is a model-specific load failure, not the same as broad remote_code regressions." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated issues: template compilation error vs tokenizer max length constant." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Video token counting bug is unrelated to removed auto-mapping config references." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Storage-object lookup error is unrelated to PIL backend torchvision requirements." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT local checkpoint loading vs Whisper decode flag handling are different code paths." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both are loading-related, but one is buffer initialization corruption and the other is FSDP RAM-efficient loading." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero-3/BERT index error is unrelated to trainer evaluate-after-train behavior." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile+SDPA failure is not the same as quantization fallback during load." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent reconstruction mismatch and DiaConfig token-id persistence are unrelated bugs." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Importing GenerationMixin and Jupyter custom-model initialization are different failures." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "Bark voice_preset and CustomTrainer loss signature are unrelated issues." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash loading test failure is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 doc-example mismatch/quantization errors are different from MXFP4 load fallback." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "BitNet packed-weight unpacking and OmDet processor image-size mismatch are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to OmDet-Turbo input-shape mismatch." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both concern exception handling, but they are different functions and different failure modes." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head tying and Blip2 dtype propagation are separate model bugs." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin and trainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Beit reduce_labels bug and MPS attention correctness issue are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and PIL torchvision dependency are different problems." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting is unrelated to Whisper batch_decode special-token handling." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading and init_empty_weights argument handling are distinct loading regressions." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Accelerate/deepspeed weight conversion and multimodal token counting are different code paths." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config/model loading, but one is missing fields from trust_remote_code and the other is composite-config unwrapping." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined video-token helper calls and batched video processor shapes are not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "CPU offload device mismatch in Bark is different from Qwen3_5MoE rotary embedding offload incompatibility." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash path warning and Qwen3 MOE weight conversion 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": 17, + "review_unit_index": 18, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:24643", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43329", + "issue:43404", + "issue:43452", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43696", + "issue:43697", + "issue:43742", + "issue:43927", + "issue:43931", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44737", + "issue:44821", + "issue:44849", + "issue:44898", + "issue:44912", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45137", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45588", + "issue:45701", + "issue:45823", + "issue:45907", + "issue:45987" + ], + "soft_pairs": [ + "issue:34567|issue:40990", + "issue:44568|issue:45356", + "issue:43329|issue:44479", + "issue:29159|issue:34567", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:44849|issue:45701", + "issue:43257|issue:43931", + "issue:43452|issue:43582", + "issue:41720|issue:43540", + "issue:43927|issue:44912", + "issue:42915|issue:44821", + "issue:43697|issue:45357", + "issue:44568|issue:45488", + "issue:33357|issue:34567", + "issue:42915|issue:45538", + "issue:44987|issue:45538", + "issue:43540|issue:44479", + "issue:44479|issue:45381", + "issue:43540|issue:44560", + "issue:43317|issue:45538", + "issue:24643|issue:29159", + "issue:45137|issue:45405", + "issue:44987|issue:45907", + "issue:43742|issue:44291", + "issue:43577|issue:45072", + "issue:44062|issue:44991", + "issue:38175|issue:44898", + "issue:44737|issue:45588", + "issue:43540|issue:45823", + "issue:43404|issue:45987", + "issue:43257|issue:43696", + "issue:34567|issue:36331" + ], + "prepared_review_unit_hash": "92b018eae36bb2194fab2962c31dd7ca487a4886b306774885e0e6c40d8f9be5", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37766, + "estimated_input_tokens": 9442, + "estimated_eval_tokens": 19140 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidates are clearly unrelated despite superficial lexical overlap. The only strong duplicate-style match is the Qwen video-input regression pair, where the specific qwen2.5-vl vision_position_ids bug is plausibly the same underlying v5.3.0 video regression affecting multiple Qwen VL variants.", + "confidence": 0.73, + "canonical_issue_reason": "issue:44479 is the broadest, most central report: a v5.3.0 video-input regression spanning multiple Qwen VL variants, and it cleanly subsumes the more specific qwen2.5-vl symptom in issue:45381.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best representative issue because it describes the shared concrete regression at the right abstraction level and is broad enough to anchor the only convincing duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated symptoms: TrainerState token counters vs GPT-OSS perplexity on WikiText." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different models and different failure modes; no clear same code path." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both involve video processing, but 43329 is an undefined-name bug in multimodal token counting while 44479 is a broader Qwen video-input regression." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState's token accounting." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different loading failures: Qwen3 MoE FP8 config vs MobileLLM key error." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized is unrelated to PIL image processor torchvision requirements." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-state output bug is unrelated to a tokenizer version-change report." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen3 loading issues, but one is MoE weight conversion under accelerate+deepspeed and the other is a weight-shape mismatch for Qwen3-VL." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF/AutoTokenizer loading and Apple Silicon allocator warmup are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen3 failures: auto device mapping cuda assert vs video-input ValueError." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig token-id persistence is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3 MoE FP8 loading and AutoImageProcessor-from-URL are unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key regression are different problems." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different tokenizer classes and different underlying behavior." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token counts." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Qwen3 MoE FP8 loading is unrelated to CLIPTokenizer max_length handling." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Model loading failure for physical-intelligence/fast is unrelated to CLIPTokenizer model_max_length." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both mention video, but one is Qwen3OmniMoe video-input processing and the other is a broader Qwen VL regression." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "45381 looks like the same video-input regression as 44479, just reported as a specific qwen2.5-vl vision_position_ids symptom." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Different tokenizer issues: add_special_tokens/BOS-EOS behavior vs LlamaTokenizer overriding ByteLevel pre-tokenization." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU tensor creation bug is unrelated to flash_attention s_aux=None crash." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Loading a model called physical-intelligence/fast is unrelated to list_repo_files not forwarding revision." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Key error loading MobileLLM and init_empty_weights argument handling are different code paths." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but they affect different model families and likely different bugs." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken special-arg duplication is not the same as tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to Perceiver interpolation failure." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "Relative positional encoding on CPU is unrelated to sink-less flash-attention handling." + }, + { + "left": "issue:43540", + "right": "issue:45823", + "accept": false, + "reason": "Qwen3OmniMoe video-input failure is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "lm_head weight tying in Mistral3 image-text generation is unrelated to StaticCache sequence-length return type." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "MoE weight conversion/loading and CUDA OOM are not the same bug." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token tracking and custom compute_loss kwargs are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed weight-dimension error is unrelated to tokenizer slow/fast inconsistency." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (issue:44479 / issue:45381) is a conservative fit: both describe the same Qwen video-input regression at the same version line, with 44479 explicitly spanning qwen2_5_vl and related Qwen VL variants. The rejections also stay conservative and do not overclaim shared root causes from superficial lexical overlap." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 19, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:34689", + "issue:36010", + "issue:42175", + "issue:43257", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43582", + "issue:43643", + "issue:43646", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43824", + "issue:43906", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44488", + "issue:44514", + "issue:44589", + "issue:44625", + "issue:44756", + "issue:44792", + "issue:44821", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45600", + "issue:45920", + "issue:46121" + ], + "soft_pairs": [ + "issue:44913|issue:46121", + "issue:44062|issue:45081", + "issue:43906|issue:45381", + "issue:43646|issue:43749", + "issue:43257|issue:43756", + "issue:43493|issue:43994", + "issue:36010|issue:42175", + "issue:44291|issue:44821", + "issue:44792|issue:44964", + "issue:43749|issue:44756", + "issue:42175|issue:43824", + "issue:45042|issue:45081", + "issue:28282|issue:34689", + "issue:45357|issue:45920", + "issue:43742|issue:44589", + "issue:43452|issue:44488", + "issue:45405|issue:45600", + "issue:43526|issue:45479", + "issue:43643|issue:45375", + "issue:28282|issue:29127", + "issue:43531|issue:44821", + "issue:43756|issue:45325", + "issue:44625|issue:44877", + "issue:43582|issue:44977", + "issue:43720|issue:44898", + "issue:43697|issue:45216", + "issue:44514|issue:45381", + "issue:24643|issue:29127" + ], + "prepared_review_unit_hash": "cdd7115b88461beb19037a12c196bba7a32e91e2c953e8810ad0060f26b469b6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36371, + "estimated_input_tokens": 9093, + "estimated_eval_tokens": 18442 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues spanning loading, config propagation, tokenization, multimodal models, and backend/package regressions. I only see one plausible duplicate pair (SigLIP2 discrepancy vs nonsensical outputs); the rest are too different to merge safely.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43493 is the clearest root-cause style report and best matches the only plausible duplicate pair in the set; most other items are unrelated bug reports.", + "canonical_pr_reason": null, + "best_issue_reason": "43493 is the strongest representative issue for the cluster because it describes the underlying SigLIP2 behavior rather than a downstream symptom. It is still not a good umbrella for the rest of the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both concern config/rope handling, but one is a reload defaulting bug and the other is a JSON-list type error in rope validation; different code paths and failure modes." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Both involve tokenizers, but one is an AddedToken constructor clash and the other is a Mistral regex patch/backend_tokenizer crash; not the same bug." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "The first is an isolated reproduction of an older issue; the second is a Qwen2.5-VL video position-id regression. No concrete overlap." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "One is custom model initialization breaking in Transformers 5.0; the other is FSDP CPU RAM efficient loading breaking. Similar area, but not the same problem." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different models and different bugs: Qwen3 MOE accelerate/deepspeed weight conversion vs Smollm3 RoPE layer dropping." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both are about SigLIP2 behaving incorrectly: an HF-vs-JAX discrepancy and a symptom of wrong/nonsensical outputs through AutoModel/pipeline. Same underlying model implementation issue." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "An import error for GenerationMixin is unrelated to a packaging/backend omission of TensorFlow in the torch extra." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "The first is an init_empty_weights argument mismatch; the second is an AutoImageProcessor URL-loading issue. Different subsystems and failures." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both are multimodal/model-loading complaints, but they involve different models and different breakages." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are different memory/loading issues with no shared concrete bug." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "One is about backend packaging for pip extras; the other is a missing Qwen2_5_VLForConditionalGeneration import. Unrelated." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL backend torchvision dependency regression is unrelated to Mistral tokenizer regex patch crashes." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "A missing PyTorch ImportError is not the same as a specific Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 save_pretrained key corruption and AutoTokenizer token-id regression are different bugs affecting different APIs." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key error and Float8 storage lookup failure are distinct loading/runtime issues." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file handling in from_pretrained and sleng-bert load failure are not the same underlying defect." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version pinning and auto_mappings stale config references are separate CI/package maintenance problems." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and sequence-classification zero-loss behavior are unrelated label-handling bugs." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Generic trust_remote_code field loss and a specific Qwen3_5MoeVisionConfig strict-field drop are different config-loading issues." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Missing PyTorch and a LayoutLMv3 box-content clarity issue are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading have no shared code-path." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both mention RoPE-related behavior, but one drops layers in Smollm3 and the other mis-scales Qwen2.5-VL temporal positions; different bugs." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "num_labels propagation and strict config loading are separate config-handling issues." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "AppleSilicon caching_allocator_warmup TypeError is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver interpolation failures are unrelated model-specific regressions." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained corruption are different regression types." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL related, but one is batched chat-template padding behavior and the other is video position-id logic; not the same bug." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "A DeepSpeed weight-shape runtime error is unrelated to LayoutLMv3 box-information error messaging." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The single accepted pair (43493/43994) is a plausible duplicate on the face of the titles, and the rest of the soft-edge rejections are appropriately cautious rather than overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 20, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:33666", + "issue:38175", + "issue:41720", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43742", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44493", + "issue:44514", + "issue:44561", + "issue:44568", + "issue:44704", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45005", + "issue:45290", + "issue:45356", + "issue:45372", + "issue:45542", + "issue:45588", + "issue:45823", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46202", + "issue:46209" + ], + "soft_pairs": [ + "issue:41762|issue:44936", + "issue:43452|issue:43577", + "issue:43906|issue:46202", + "issue:44466|issue:45005", + "issue:43452|issue:44843", + "issue:45372|issue:46170", + "issue:43819|issue:44811", + "issue:44936|issue:45823", + "issue:30064|issue:33666", + "issue:43295|issue:45356", + "issue:44484|issue:44704", + "issue:43653|issue:44361", + "issue:44361|issue:44442", + "issue:45542|issue:46178", + "issue:44514|issue:45290", + "issue:38175|issue:43295", + "issue:43653|issue:44568", + "issue:43577|issue:43697", + "issue:44464|issue:44743", + "issue:44493|issue:46050", + "issue:44843|issue:44991", + "issue:29159|issue:33357", + "issue:43531|issue:43742", + "issue:43653|issue:43927", + "issue:41720|issue:43906", + "issue:44464|issue:44977", + "issue:43550|issue:43927", + "issue:43550|issue:44361", + "issue:44977|issue:46209", + "issue:43643|issue:44561", + "issue:43577|issue:45588" + ], + "prepared_review_unit_hash": "48341bb8404662b4cdb453bb2ae411495b50b2dcfb7b87cd52753c238fade7bd", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36587, + "estimated_input_tokens": 9147, + "estimated_eval_tokens": 18550 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are a heterogeneous set of unrelated Transformers bugs and regressions. The proposed soft pairs mostly share only a broad subsystem label (tokenizers, processors, generation, loading, or trust_remote_code) but not the same concrete failure mode, so none 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:41762", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: Gemma3 ZeRO-3 load-time IndexError vs trainer.evaluate() after train." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "Unrelated bugs: gguf_file loading regression vs Blip2 dtype staying float32." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "No same code path; one is a reproduction of an old issue, the other is a Gemma pipeline roundtrip/tokenization problem." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is serialization-by-device and the other is translation-model behavior in v5; not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both concern tokenizer loading, but one is gguf_file support and the other is offline model_info() usage in a regex patch." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Completely different issues: Gemma 4 processor import regression vs Windows ROCm torch.distributed import/type failure." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different features and symptoms: DAC latent decoding mismatch vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Different subsystems and failures: trainer lifecycle vs Gemma4 device-mapping mismatch." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "No overlap beyond being model-related; segmentation-map processing vs Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Different regressions: processor.tokenizer access/images-to-tokenizer breakage vs Kimi-K2.5 codec/fix_mistral_regex behavior." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "Save-pretrained shard-size question is unrelated to AutoProcessor kwargs propagation to cached_file." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Both tokenizer-related, but one is special-token registration/decode output and the other is an AttributeError in MLukeTokenizer tasks." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer loading failures with different models and causes." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both backend-related, but one is missing TensorFlow backend state and the other is an unsupported tensorflow_text backend mapping." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Different apply_chat_template crashes: batched padding=False vs assistant tool-call messages without content." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different model classes and failures: zero probabilities in SigLIP2 vs processor/tokenizer regression in v4.57.5." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both tokenizer issues, but one is BigBird special-token registration and the other is mdeberta BOS/EOS addition in v5." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Different bugs: BLIP2 dtype casting vs RTDetrV2 output changes in Transformers v5." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both generation/cache related, but one is chunked generation with compiled forward and the other is Qwen3 modular recurrent-state reset." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Different serialization/loading problems: unexpected position-id keys vs sinq quantization roundtrip breakage." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both tokenizer-loading regressions, but one is offline model_info() access and the other is a specific EMBEDDIA/est-roberta load failure." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer inconsistency vs MacOS bus error on a CLIP model are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Different model-specific issues: sliding_window in Qwen3-MoE vs loading MobileLLM-125M key error." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Different failures: BigBird special-token decode bug vs DiaConfig losing custom token IDs after save/load." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Different problems and code paths: auto device mapping cuda assert vs a reproduction of another issue." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Both affect generation, but compiled-forward chunking and Qwen3.5 flash-attention failures are not the same concrete bug." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Unrelated: torch.compile/SDPA failure in Bamba vs DiaConfig save/load token-id loss." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Completely different: Bamba compile/SDPA issue vs tokenizer task AttributeError." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Different runtime failures: Qwen3.5 flash-attention generation vs FP8 inference on B200." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve trust_remote_code, but one is missing fields from AutoConfig and the other is removal of is_torch_fx_available breaking models." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Different code paths: BLIP2 dtype loading vs flash_attention.py crash on s_aux=None." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 21, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28282", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:35141", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43262", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43825", + "issue:43927", + "issue:43994", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44554", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44849", + "issue:44913", + "issue:44964", + "issue:45357", + "issue:45405", + "issue:45440", + "issue:45491", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:32090|issue:34567", + "issue:28282|issue:35141", + "issue:43504|issue:43550", + "issue:43504|issue:43720", + "issue:41720|issue:44464", + "issue:43504|issue:44610", + "issue:45491|issue:46002", + "issue:32090|issue:33666", + "issue:33666|issue:36010", + "issue:44466|issue:46050", + "issue:43742|issue:44821", + "issue:43688|issue:45440", + "issue:45823|issue:46050", + "issue:43720|issue:43927", + "issue:43262|issue:43295", + "issue:43606|issue:43927", + "issue:28282|issue:32090", + "issue:43452|issue:44451", + "issue:44849|issue:46209", + "issue:28282|issue:33357", + "issue:43994|issue:46192", + "issue:43525|issue:43572", + "issue:33666|issue:36331", + "issue:43720|issue:44361", + "issue:43756|issue:44913", + "issue:43493|issue:43825", + "issue:44849|issue:45357", + "issue:44479|issue:44625", + "issue:44964|issue:45405", + "issue:39692|issue:43493", + "issue:43577|issue:44554" + ], + "prepared_review_unit_hash": "637eb13b9fb0799d4536528772077f5920a3029c66ca0f970685ee96f40c3c3d", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37132, + "estimated_input_tokens": 9283, + "estimated_eval_tokens": 18822 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly superficial similarities across unrelated bugs; none look like the same underlying issue or a mergeable duplicate PR. This set is a grab-bag of model-loading, trainer, config, and inference regressions rather than one coherent defect cluster.", + "confidence": 0.27, + "canonical_issue_reason": "Issue 43504 is a clean, self-contained load-time bug report with a concrete failure mode, making it the best anchor among these mostly unrelated reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43504 is the strongest representative issue here because it is specific, reproducible, and centered on a concrete code path rather than a vague symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and token-count tracking bug are different training subsystems and failure modes." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "Missing PyTorch import and output embedding reinitialization in post_init are unrelated bugs." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Legacy-field model loading failure vs torch.compile/SDPA failure in Bamba are different code paths and models." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "BEiT preset loading bug and BitNet packed-weight accelerate loading bug are distinct model-loading issues." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Auto device mapping cuda assert and chunked-generation compiled-forward inconsistency are separate runtime bugs." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Pretrained preset load failure is unrelated to OmDet processor/model input-size mismatch." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both involve NaNs, but one is a Gemma3 sliding-window GPU issue and the other is a classification-head init/checkout issue." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Custom Trainer broadcast failure and Qwen2-VL multi-GPU training are not the same bug." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training issue and GenerationMixin import error are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "lm_head serialization inconsistency and sinq quantization save/load breakage are different persistence bugs." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM checkpoint load failure and AutoImageProcessor URL loading are different object-loading paths." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Aux-loss normalization bug and DeepSeekV3 implementation divergence are different correctness issues." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 auto device-map mismatch and quantized model save/load breakage are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are different models and failure modes." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate default and processor.tokenizer regression are distinct processor bugs." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch and config token-ID loss on save/load are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing torch import and Trainer broadcast TypeError are completely different errors." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file loading breakage and inability to load a specific model are different load failures." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state bug and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "PyTorch import error and MacOS bus error on CLIP are not the same underlying bug." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 wrong outputs and return_tensors being ignored are separate API regressions." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are similar only superficially; they affect different configs and fields." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU training issue and custom compute_loss signature mismatch are unrelated trainer problems." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are different bugs." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Smollm3 RoPE-layer-count bug and GPTNeoXConfig rotary_pct reload bug are different configuration issues." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and pipeline translation-task messaging are unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5 hidden-states bug and save_pretrained visual-encoder key regression are different failure modes." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression and num_labels propagation bug are separate Qwen3.5 issues." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Model load failure and PEFT version pinning are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 doc-example failures and SigLIP2 implementation divergence are not the same concrete bug." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype-loading bug and MPS attention correctness issue are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 22, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:33357", + "issue:33666", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43749", + "issue:43756", + "issue:43824", + "issue:43844", + "issue:43927", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44734", + "issue:44792", + "issue:44843", + "issue:44977", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45405", + "issue:45520", + "issue:45584", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823" + ], + "soft_pairs": [ + "issue:43688|issue:43749", + "issue:44568|issue:45812", + "issue:43540|issue:44977", + "issue:45200|issue:45229", + "issue:44488|issue:44843", + "issue:43334|issue:43572", + "issue:39692|issue:43295", + "issue:43756|issue:45083", + "issue:43334|issue:45759", + "issue:43688|issue:45704", + "issue:43646|issue:44466", + "issue:43493|issue:43844", + "issue:43824|issue:44589", + "issue:45229|issue:45823", + "issue:33357|issue:33666", + "issue:43927|issue:44361", + "issue:43504|issue:43606", + "issue:44589|issue:45520", + "issue:43572|issue:44291", + "issue:43278|issue:43844", + "issue:44625|issue:45759", + "issue:43257|issue:43299", + "issue:43504|issue:43653", + "issue:43606|issue:44112", + "issue:43575|issue:45405", + "issue:43688|issue:45701", + "issue:44792|issue:45405", + "issue:44423|issue:44734", + "issue:44464|issue:45584" + ], + "prepared_review_unit_hash": "ed7803c2025dd1bca8dcb49e579a4537139bdeadc30724fa9ef6778271285ba3", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35988, + "estimated_input_tokens": 8997, + "estimated_eval_tokens": 18250 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly unrelated standalone issues; the soft edges are largely false positives based on broad surface similarity (tokenizers, Qwen, Gemma, loading regressions, CI/device issues) rather than the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the strongest standalone report here: it is concrete, well-scoped, has a clear regression boundary (v4\u2192v5), and describes a high-impact tokenizer bug with a specific affected family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Different problems: auxiliary-loss normalization vs FSDP RAM-efficient loading failure." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenization behavior." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Different code paths: Qwen3OmniMoe video preprocessing vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "One is a missing default for mm_token_type_ids; the other is a multi-GPU inference OOM." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both are model-loading failures, but one is a specific model compatibility issue and the other is an offline-mode hub lookup regression." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different model/config attributes and different failure modes; only superficial similarity around missing token-id fields." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Docs-example errors vs a runtime regression in custom processor/tokenizer code; not the same bug." + }, + { + "left": "issue:43756", + "right": "issue:45083", + "accept": false, + "reason": "Unrelated model architecture/blog discrepancy vs a helper-function bug in qwen3_omni_moe." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Different Qwen multimodal loading issues with different missing/unwrap logic." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Loss normalization bug is unrelated to the T5 RMSNorm memory leak." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but they hit different code paths: init behavior vs weight serialization." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "Different subsystems: SigLIP2 implementation mismatch vs DeepSpeed/ZeRO gradient explosion." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing a Qwen model class is unrelated to the Float8 storage lookup failure." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both involve Gemma4, but one is OOM and the other is a device-map mismatch during forward." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Completely different models and failure modes: MacOS bus error vs multi-GPU training question." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenization/model API bugs with no shared underlying code path." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy-field model loading bug vs CPU-offload device mismatch; not the same issue." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different failures: missing Float8 storage type vs missing flash_attn key on Python 3.13." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "A missing config field after 5.0 update is not the same as an init_empty_weights argument mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Training/eval dtype drift is unrelated to the ZeRO-3 gradient anomaly." + }, + { + "left": "issue:44625", + "right": "issue:45759", + "accept": false, + "reason": "Related model family, but different propagation bugs and different affected APIs." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE, but one is accelerate/deepspeed weight conversion and the other is dev loading breakage; not the same concrete bug." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Different models and failures: Beit legacy field loading vs BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A real CI test failure on GraniteSpeech is not the same as a CPU-offload device mismatch in bark-small." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "OOM during tensor-parallel loading is unrelated to a dependency version bump." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "The titles only share a broad theme of model behavior changes; the actual bugs are different." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A Janus test failure is unrelated to the PEFT version pin issue." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are server/runtime issues, but one is a multimodal continuous-batching crash and the other is KV-cache continuation indexing." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation compiled-forward inconsistency is unrelated to Whisper empty-transcription handling." + } + ] + }, + "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": 17, + "review_unit_index": 23, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:33357", + "issue:34567", + "issue:38175", + "issue:39692", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43572", + "issue:43688", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44186", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44849", + "issue:44964", + "issue:44987", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45325", + "issue:45356", + "issue:45381", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45823", + "issue:46050", + "issue:46121", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:29942|issue:30333", + "issue:43994|issue:44493", + "issue:45588|issue:46050", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:43688|issue:43873", + "issue:44521|issue:45290", + "issue:44554|issue:45446", + "issue:43540|issue:45200", + "issue:45084|issue:45520", + "issue:15354|issue:38175", + "issue:45200|issue:45823", + "issue:43873|issue:46050", + "issue:45072|issue:45198", + "issue:43278|issue:43701", + "issue:43232|issue:44849", + "issue:44186|issue:46192", + "issue:45406|issue:46121", + "issue:45823|issue:46192", + "issue:43493|issue:43701", + "issue:28218|issue:33357", + "issue:43295|issue:43450", + "issue:43295|issue:44610", + "issue:43572|issue:44987", + "issue:45381|issue:46209", + "issue:43572|issue:43957", + "issue:44554|issue:45356", + "issue:39692|issue:46192", + "issue:45200|issue:45325", + "issue:43295|issue:43873" + ], + "prepared_review_unit_hash": "f4e56951dfa82192a6270529af8b6b4ab9acc9a3961aebf6856a8b3164251451", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37030, + "estimated_input_tokens": 9258, + "estimated_eval_tokens": 18772 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are a heterogeneous set of distinct bugs and regressions across tokenization, generation, multimodal processing, loading, and training. The soft-similarity pairs are mostly broad-subs\u200bystem overlaps rather than the same concrete underlying issue, so none should be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Unrelated failures: Flash Attention test regressions vs MLflow reporting after a failed training job." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Both mention model correctness, but one is a SigLIP2 inference-quality bug and the other is a generic position-id key warning across many models." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Different problems: flash_attention sink-less crash on s_aux=None vs quantized model save/load corruption." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "DeepSpeed ZeRO-3 gradient growth and Phi-4 multimodal loading regression are separate issues." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState's num_input_tokens_seen tracking." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency does not match the Trainer NoneType broadcast crash." + }, + { + "left": "issue:43688", + "right": "issue:43873", + "accept": false, + "reason": "Aux-loss normalization and quantization/offloading behavior are different code paths and symptoms." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but one is all-zero assistant masks for multimodal inputs and the other is a crash on tool-call messages with no content." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "Attention correctness on MPS is unrelated to a PyTorch version check in flex_attention import logic." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video input validation in Qwen3OmniMoe is not the same as Gemma4 mm_token_type_ids defaulting." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Compilation of non-template nodes and flash_attn import KeyError are distinct failures." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero-probability inference." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 token-type-id defaulting and device-map forward placement issues are different bugs." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization offloading issues and sinq save/load breakage do not describe the same defect." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatches in inference are unrelated to Wav2Vec2 save_pretrained/tokenization failures." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "BF16-to-FP32 dtype drift in eval and resume_from_checkpoint key mismatch are separate training-state problems." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs sync bug and Qwen3.5 output_hidden_states bug affect different paths." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are not the same as return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor _tokenizer attribute regression is unrelated to rope-param list type handling." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device placement and return_tensors precedence are different behaviors." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation mismatch and checkpoint resume key mismatch are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Added-token spacing and CLIP MacOS bus error are unrelated tokenizer vs runtime crashes." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor.tokenizer regression and batched video shape bug are different multimodal issues." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Both are processor/model input regressions, but one is an attribute/API break and the other is a resolution mismatch." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is not the same as a specific model-load failure on transformers>=5.1.0." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL vision_position_ids bug and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Config missing pad_token_idx and meta-device loading regressions are different loading problems." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness issue is unrelated to Kimi-K2.5 tokenizer codec/regression handling." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 documentation/example errors are not the same as return_tensors being ignored." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma4 token-type-id defaults and Qwen2.5-VL temporal position-id scaling are distinct multimodal bugs." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor API regression and quantization offloading failure are 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": 17, + "review_unit_index": 24, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:37428", + "issue:38175", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43653", + "issue:43696", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43844", + "issue:43856", + "issue:43901", + "issue:43957", + "issue:43975", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44857", + "issue:44898", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:46170", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:44898|issue:45412", + "issue:43550|issue:43844", + "issue:44466|issue:45127", + "issue:43334|issue:43525", + "issue:37428|issue:43957", + "issue:43450|issue:43575", + "issue:45520|issue:46178", + "issue:43824|issue:43975", + "issue:42490|issue:43696", + "issue:43278|issue:43856", + "issue:43526|issue:46192", + "issue:42915|issue:43824", + "issue:43653|issue:44112", + "issue:38175|issue:43493", + "issue:44442|issue:45198", + "issue:43975|issue:45520", + "issue:43295|issue:43493", + "issue:44561|issue:46234", + "issue:37428|issue:43824", + "issue:38175|issue:45412", + "issue:44514|issue:44521", + "issue:43262|issue:43450", + "issue:41720|issue:46209", + "issue:45520|issue:46170", + "issue:43819|issue:44112", + "issue:44610|issue:44857", + "issue:43827|issue:43901", + "issue:43819|issue:45083", + "issue:44206|issue:45584", + "issue:43335|issue:43550", + "issue:37428|issue:43425" + ], + "prepared_review_unit_hash": "aa067815102bf5afc9630624fa74b6d56f6b27f7fd14f25c43ad07fff86b5532", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37167, + "estimated_input_tokens": 9292, + "estimated_eval_tokens": 18840 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft-similarity pairs span unrelated import errors, model-specific regressions, processor bugs, docs issues, and performance/OOM reports. None look like the same concrete underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No clear duplicate cluster exists; if a representative issue must be chosen, issue 44521 is the most active open item with the broadest follow-up potential around `apply_chat_template` multimodal behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44521 is the best representative because it is open, has the most discussion/inbound interest, and concerns a user-facing multimodal templating path likely to draw continued attention.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different bugs: Perceiver non-default resolution failure vs RT-DETR memory not being released." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "Different failure modes: torch.compile+SDPA crash vs gradient growth under ZeRO-3." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Different problems: lm_head serialization/tied weights vs LoRA merge collapse with extended vocab." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Same symptom class (`pad_token_id` missing) but different model families and config paths; not the same bug." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Import helper removal/import error vs meta-device loading regression; different code paths." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Unrelated issues: batched video processor shape vs tensor-parallel OOM." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both are import/platform issues, but they involve different backends and different root causes." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Import error for a VL model class vs incorrect detokenization in DeepSeek coder; unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Local-path trailing-slash warning vs GPT-oss-20b CUDA OOM; unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype regression in eval vs Qwen3 MoE memory inefficiency; different bugs." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs `return_tensors` being ignored with `text_kwargs`; different processors and behaviors." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to a Qwen2.5-VL import error." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration bug vs stale GraniteSpeech CI test; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is zero probabilities and the other is HF/JAX parity mismatch; not the same concrete bug." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Different ASR/tokenizer issues: FastSpeech2ConformerTokenizer load failure vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "DeepSeek detokenization bug vs import_utils flash_attn KeyError on Python 3.13; unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43493", + "accept": false, + "reason": "Processor.tokenizer / image-passing regression is not the same as SigLIP2 implementation discrepancy." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Different removed/missing import_utils symbols (`is_torch_fx_available` vs `is_falcon_mamba_ssm_available`) and different breakages." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different import errors for different missing symbols/models; not one underlying bug." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 output discrepancy and RT-DETR memory retention are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both hit `apply_chat_template`, but one crashes on batched padding=False and the other returns zero assistant masks; different bugs." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio sampling-rate default bug vs batched video input shape bug; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert vs FP8 inference failure on B200; different hardware/runtime issues." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Both are import-time/platform problems, but the affected dependencies and failure points differ." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC quantization/STE mismatch vs GraniteSpeech CI device override test; unrelated." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Processor output-size bug vs AMP float16 loss crash; different subsystems and causes." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they concern different pipelines and different deprecated parameters." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent conversion mismatch vs qwen3_omni_moe feature-length helper behavior; unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Audio feature-extractor regression vs Whisper empty-transcription generation bug; different code paths." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer configuration bug vs Bamba torch.compile+SDPA failure; unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "ImportError for a missing flash-attn helper vs generic Torch 2.10 incompatibility; not the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 25, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28218", + "issue:29942", + "issue:32090", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43388", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:43931", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44743", + "issue:44779", + "issue:44792", + "issue:44855", + "issue:44857", + "issue:45072", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859", + "issue:46234" + ], + "soft_pairs": [ + "issue:44857|issue:45593", + "issue:43606|issue:45083", + "issue:44792|issue:45563", + "issue:44448|issue:44743", + "issue:45127|issue:45245", + "issue:45290|issue:45381", + "issue:45200|issue:45588", + "issue:43335|issue:43653", + "issue:44206|issue:45823", + "issue:43611|issue:43931", + "issue:43550|issue:44112", + "issue:37428|issue:44351", + "issue:44855|issue:45083", + "issue:43819|issue:44493", + "issue:43335|issue:43504", + "issue:45588|issue:45859", + "issue:29942|issue:32090", + "issue:43278|issue:43825", + "issue:43388|issue:43526", + "issue:43335|issue:45083", + "issue:45459|issue:46234", + "issue:39692|issue:43526", + "issue:39692|issue:44493", + "issue:45310|issue:45478", + "issue:43720|issue:44112", + "issue:28218|issue:35141", + "issue:44448|issue:44779", + "issue:43335|issue:44112", + "issue:43335|issue:45072", + "issue:42490|issue:43854", + "issue:43606|issue:44857" + ], + "prepared_review_unit_hash": "373ada10a182d310e21ab9cfdb9b2bf2ad9c89d60e1be3867e7f6829bef5543b", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36707, + "estimated_input_tokens": 9177, + "estimated_eval_tokens": 18610 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items in this cluster are unrelated bug reports that only share broad Transformers subsystem vocabulary. The only clear duplicate pair is the Qwen3.5 MoE from_pretrained regression: issues 45310 and 45478. No pull requests are present.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 45310 is the clearest canonical issue because it is the earlier report for the same Qwen3.5 Moe from_pretrained failure on transformers>=5.4.0, and 45478 is a near-verbatim follow-up of the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "45310 is the best issue to keep as the representative duplicate target: it matches 45478 exactly on the failing model/version and appears slightly earlier and more canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different detection bugs: AMP/CUDA loss crash vs auxiliary-loss omission when denoising is off." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Unrelated code paths: CPU offload/device mismatch in Bark vs a qwen3_omni_moe feature-length helper." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "One is a Janus image generation test failure; the other is a paged generate warning for num_return_sequences." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Different models and failures: Pegasus v4/v5 output drift vs Qwen3 recurrent-state reset with cache." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/extended-vocab collapse is unrelated to the 2^24 categories runtime limit." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video position-id bug are separate features/paths." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 text-only mm_token_type_ids issue is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and BigBirdTokenizer special-token registration are different problems." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Feature extractor center-arg regression is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Base_model_prefix loading breakage is not the same as Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba and stale GraniteSpeech device-override test are different bugs." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but for different missing symbols and different modules/causes." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 indentation/JIT parsing issue is unrelated to the qwen3_omni_moe helper behavior." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC from_latents/STE mismatch has nothing in common with position-id key warnings." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Sparse-layer creation bug in SwitchTransformers is unrelated to BEiT semantic-segmentation legacy-field loading." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention sinkless-model crash and Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility are distinct." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate failure modes." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Embedding dtype drift during evaluate is unrelated to the pipeline translation-task warning regression." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Metric gathering drops tuple labels, while BeitImageProcessorFast reduce_labels returns only one label; not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe feature-length helper are unrelated." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer error masking without protobuf and missing is_falcon_mamba_ssm_available are different import_utils issues." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs/model mismatch is unrelated to BEiT fast processor label reduction." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Documentation example errors do not match the position-id warning issue." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same concrete bug: Qwen3.5 Moe from_pretrained fails on transformers>=5.4.0; 45478 is a duplicate report of 45310." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet packed-weights loading and GraniteSpeech stale device override are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer extra-space behavior and embedding reinitialization after resize are different code paths." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output differences and Deepseek tokenizer regressions are distinct model behaviors." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is unrelated to GraniteSpeech device override failures." + }, + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "Config-layer creation bug is unrelated to bfloat16 dtype mismatches in inference." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Slash-path import warning and GLM-4.7-Flash unit-test loading failure are different loading issues." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Device mismatch in Bark CPU offload is not the same as LwDetrImageLoss crashing under float16 AMP." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: the only strongly supported duplicate pair is 45310/45478, which have identical titles and clearly describe the same Qwen3.5 MoE from_pretrained regression. The other soft-edge verdicts are conservative and generally supported by the issue titles, and the note that there are no PRs is correct." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 26, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:34689", + "issue:35141", + "issue:36010", + "issue:37428", + "issue:38175", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43381", + "issue:43412", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43653", + "issue:43701", + "issue:43746", + "issue:43927", + "issue:44360", + "issue:44448", + "issue:44466", + "issue:44493", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45712", + "issue:45803", + "issue:45823", + "issue:45920", + "issue:46192", + "issue:46209" + ], + "soft_pairs": [ + "issue:45920|issue:46192", + "issue:38175|issue:44493", + "issue:44360|issue:44561", + "issue:44977|issue:45229", + "issue:44857|issue:45072", + "issue:44448|issue:44977", + "issue:34689|issue:35141", + "issue:43701|issue:45083", + "issue:29127|issue:32090", + "issue:43550|issue:45083", + "issue:40444|issue:43526", + "issue:43927|issue:45698", + "issue:29942|issue:35141", + "issue:42490|issue:43575", + "issue:44561|issue:44779", + "issue:44977|issue:45563", + "issue:44857|issue:45588", + "issue:43746|issue:46209", + "issue:44568|issue:45491", + "issue:45372|issue:45588", + "issue:43412|issue:45588", + "issue:45198|issue:45803", + "issue:43278|issue:43381", + "issue:45588|issue:45823", + "issue:36010|issue:37428", + "issue:44448|issue:44466", + "issue:43653|issue:44743", + "issue:25251|issue:32090", + "issue:45245|issue:45563", + "issue:45459|issue:45712" + ], + "prepared_review_unit_hash": "bb2afbdf2d45c7f91ff8b627236528b4e32c6ac93ca93f795524766078bcad9d", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36943, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 + }, + "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 set spans distinct regressions across tokenization, loading, generation, attention, training, and export. I\u2019d use issue 45920 as the best representative issue because it is open, broad, and describes a multi-model v5 tokenizer regression. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45920 is the broadest, most representative open issue here: it summarizes a v5 tokenizer regression affecting multiple model families, making it a stronger canonical anchor than the narrower single-model reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best global issue candidate because it is open, high-impact, and framed as a cross-model regression rather than a model-specific edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Both concern tokenizer/image processing inputs, but one is a multi-model token-id regression and the other is a `return_tensors` parameter precedence bug; different failure modes." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Siglip2 zero probabilities and unexpected position-id keys are unrelated inference bugs with different symptoms and code paths." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA ReLU discussion and removal of `is_torch_fx_available` are unrelated; no shared concrete bug." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and Gemma4 multi-GPU OOM are different issues and would not be merged into one fix." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "One is an AMP/CUDA crash in a loss module; the other is a dtype mismatch during bfloat16 inference." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Transformers v4/v5 output drift for Pegasus and Qwen3.5 flash-attention generation are unrelated model behaviors." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model loading breakage for Llama 3.2 vision and post-init reinitialization of embeddings are separate bugs." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint resume key mismatch and `_get_feat_extract_output_lengths` behavior do not describe the same underlying defect." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 box-validation messaging and Trainer `_gpu_broadcast_one` NoneType handling are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Bamba `torch.compile`/SDPA failure is not the same as the qwen3_omni_moe feature-length helper issue." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "IterableDataset multi-image finetuning failure and BeitImageProcessorFast `reduce_labels` bug affect different pipelines and code paths." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Broken custom token IDs after save/load and wrong custom module selection after save_pretrained are both persistence-related, but they are distinct bugs." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures and embedding reinitialization are not the same code-path issue." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Trailing-slash import warning and tensor-parallel OOM while loading Qwen2-57B are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "`is_torch_fx_available` removal breaking trust-remote-code models and DeepSeek tokenizer regressions are different regressions." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation correctness and a stale warning about `num_return_sequences` are not the same bug." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "Loss-module AMP crash and flash_attention `s_aux=None` crash involve different model components and failure conditions." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "PEFT adapter local checkpoint loading and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token addition regression and Gemma3 NaN embeddings under sliding-window attention are different issues." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor import failure and flash_attention `s_aux=None` crash are separate problems." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "ExecuTorch export constraints for Gemma3n and flash-attention sink-less crash do not share a concrete fix." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure and a bare-except swallowing errors in Fuyu are distinct bugs." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype drift in evaluation and gradient checkpointing in eval mode are related only by eval, not by root cause." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Flash_attention `s_aux=None` handling and Gemma4 PLE device placement mismatch are different runtime errors." + }, + { + "left": "issue:36010", + "right": "issue:37428", + "accept": false, + "reason": "Missing `GenerationMixin` import and missing `_flash_supports_window_size` import are separate import regressions." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus v4/v5 output differences and `lm_head.weight` serialization inconsistency are different model state bugs." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline `top_k` nesting change and Trainer broadcast of NoneType are not the same underlying issue." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-cardinality limit and stale `num_return_sequences` warning are unrelated generation/configuration issues." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking without protobuf and leftover dummy classes failing repo checks are different maintenance 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": 17, + "review_unit_index": 27, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:34567", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43421", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43873", + "issue:43976", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44442", + "issue:44462", + "issue:44479", + "issue:44610", + "issue:44734", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:46121" + ], + "soft_pairs": [ + "issue:43065|issue:45412", + "issue:44186|issue:45198", + "issue:43611|issue:46121", + "issue:44871|issue:44977", + "issue:44361|issue:45198", + "issue:44206|issue:44479", + "issue:33357|issue:39692", + "issue:45372|issue:45459", + "issue:29942|issue:30064", + "issue:29127|issue:34567", + "issue:43976|issue:45440", + "issue:42915|issue:42994", + "issue:43994|issue:45412", + "issue:44265|issue:45198", + "issue:43278|issue:43575", + "issue:10000|issue:15354", + "issue:43066|issue:44462", + "issue:42490|issue:43873", + "issue:43873|issue:45229", + "issue:44734|issue:45406", + "issue:43421|issue:45491", + "issue:43653|issue:44442", + "issue:43065|issue:43994", + "issue:43381|issue:43701", + "issue:44610|issue:45412", + "issue:41720|issue:45229", + "issue:32090|issue:33357", + "issue:43756|issue:45440", + "issue:45372|issue:45406", + "issue:29942|issue:36010", + "issue:43278|issue:43329" + ], + "prepared_review_unit_hash": "25084b9c9a816f144968988dbf9098f1e24bf5f206ac886266ff83a6c406f010", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of unrelated single-issue reports. None of the soft pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue cleanly represents this cluster; the items span unrelated subsystems, so there is no strong canonical duplicate anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45406 is the most self-contained and actionable concrete bug report in the set, with a specific crash path and direct repro.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Different subsystems and symptoms: a dummy conv layer in a decoder vs a model memory-leak cleanup issue." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer NER/padding crash and Wav2Vec2 save/tokenization failure are unrelated code paths." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Model loading with base_model_prefix is unrelated to rope-parameter validation type handling." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Different models and failures: eos_token_id config mismatch vs flash-attention generation bug." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLuke tokenizer AttributeError and Wav2Vec2 save/tokenization issues do not share the same bug." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Feature extractor argument incompatibility is unrelated to the video-input regression in Qwen VL models." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "One is a MacOS bus error on a community CLIP model; the other is a documentation example with different errors." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 processor import failure and protobuf-error masking are distinct failure modes with different causes." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention test failures and void segmentation map handling are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "Error-message clarity in LayoutLMv3 and Trainer token-count tracking are different bugs." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python-version compatibility regression is unrelated to DeepSeekV3MoE implementation divergence." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both mention quantization, but one is an FP8 config failure and the other is quantized saving; different concrete problems." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "Nonsensical SigLIP2 outputs and RT-DETR memory retention are separate issues." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures from torch_compilable_check are unrelated to Wav2Vec2 tokenizer/save issues." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype drift during eval and tensor-parallel OOM are not the same underlying bug." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "German DistilBertModel issue and TorchScript generator-expression incompatibility are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and ignoring tokenizer.json are different tokenizer-loading bugs." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Local-path import warning and quantization offloading malfunction are not the same issue." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Offloading/quantization behavior and Gemma4 multi-GPU OOM have different root causes and fixes." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation indexing in serve is unrelated to Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime post-processor updates and Gemma3 NaN embeddings are different tokenizer/model bugs." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are distinct tokenizer issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "A dummy conv in Sam3PixelDecoder is unrelated to SigLIP2 incorrect outputs." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode gradient checkpointing and resume checkpoint key mismatch are different training-state problems." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor size mismatch for OmDet-Turbo is unrelated to RT-DETR memory release." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "cudaErrorAssert with auto device mapping is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast error and MacOS bus error are unrelated runtime failures." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer count mismatch and DeepSeekV3MoE implementation divergence are different model-specific bugs." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both involve Gemma4 loading, but one is an external import dependency problem and the other is an internal processor attribute crash." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention test failures and GenerationMixin import errors are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype changes and undefined video-branch helpers in multimodal token counting are unrelated code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 28, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28218", + "issue:29127", + "issue:32090", + "issue:33357", + "issue:35141", + "issue:38175", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43334", + "issue:43381", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43646", + "issue:43701", + "issue:43723", + "issue:43761", + "issue:43824", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44206", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44464", + "issue:44554", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45245", + "issue:45372", + "issue:45440", + "issue:45491", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45834", + "issue:45859", + "issue:45907" + ], + "soft_pairs": [ + "issue:43526|issue:45834", + "issue:43761|issue:45859", + "issue:43723|issue:44554", + "issue:45103|issue:45372", + "issue:45440|issue:45593", + "issue:43526|issue:44186", + "issue:43334|issue:43531", + "issue:43575|issue:43701", + "issue:43526|issue:44857", + "issue:44977|issue:45245", + "issue:42490|issue:43381", + "issue:43575|issue:44387", + "issue:43824|issue:44279", + "issue:44442|issue:45072", + "issue:29127|issue:35141", + "issue:44743|issue:44811", + "issue:42915|issue:43931", + "issue:43646|issue:44077", + "issue:43262|issue:44206", + "issue:44977|issue:45584", + "issue:43452|issue:45588", + "issue:33357|issue:38175", + "issue:44062|issue:45907", + "issue:44464|issue:45491", + "issue:28218|issue:32090", + "issue:44442|issue:44898" + ], + "prepared_review_unit_hash": "752b27a85a78fdfff8ac2d8ef90a787cff16798e6e8f4b1d3a90a760c5970632", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34979, + "estimated_input_tokens": 8745, + "estimated_eval_tokens": 17746 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; it mixes many unrelated bugs across tokenization, model loading, attention, audio, and training. No soft edge looks like the same underlying change or bug.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated subsystems and failure modes, so there is no single representative duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global canonical for this cluster; any choice would be arbitrary because the issues are thematically unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "Both are vision/ocr-related at a high level, but one is a BeitImageProcessorFast label-reduction bug and the other is a Kosmos2.5 long-OCR index error; different code paths." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "Unrelated failures: CLIPVisionModel hidden_states handling versus Qwen3_5Moe rotary embedding CPU offload compatibility." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading in v5 and an MPS attention correctness issue are different problems in different subsystems." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation parsing crash is unrelated to a mistral_common import error in Gemma 4 processor loading." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "Different model implementations and bugs: DeepSeekV3 divergence versus D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "Both involve processor/tokenizer behavior, but the concrete bugs differ: label reduction versus NER/batched padding crashes." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Same broad Qwen prefix only; pad_token_id absence and sliding_window behavior are distinct bugs." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Qwen2 large-model TP OOM and resume_from_checkpoint key mismatch are unrelated training/loading issues." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Image processor label reduction and LwDetrImageLoss AMP/CUDA crash are different code paths and symptoms." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation bug and category cardinality limit error are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Local path import warning and gradient checkpointing eval-mode restriction are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory/OOM, but one is TP loading OOM for a specific Qwen2 model and the other is int4 CUDA reserved-memory regression." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "ImportError for a specific model class and a generic dependency issue are not the same underlying bug." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are separate failures." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging and embedding reinitialization on resize are different issues." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Recurrent cache reset in Qwen3.5 modular code and Whisper batch_decode skip_special_tokens handling are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are different model-loading problems." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Framework-wide custom initialization breakage is not the same as a patchtsmixer-specific post_init policy issue." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and unsupported center arg in LasrFeatureExtractor are unrelated audio-processing bugs." + }, + { + "left": "issue:44977", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issue and Whisper empty-transcription failure after align_special_tokens are different bugs." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "GGUF loading/AutoTokenizer behavior and flash_attention s_aux=None crash are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:38175", + "accept": false, + "reason": "MacOS bus error with CLIP and zero-probabilities in SigLIP2 are different model/runtime failures." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken special-argument duplication and list_repo_files revision forwarding are unrelated API bugs." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Chunked generation under compiled forward and Gemma3 NaNs with mixed-length sliding windows are different correctness issues." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer added-token spacing bug and Trainer _gpu_broadcast_one NoneType error are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading failure and Perceiver non-default-resolution failure are different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 29, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:29127", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36010", + "issue:42175", + "issue:42907", + "issue:43316", + "issue:43381", + "issue:43493", + "issue:43525", + "issue:43550", + "issue:43611", + "issue:43643", + "issue:43650", + "issue:43720", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44568", + "issue:44743", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44964", + "issue:44987", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45440", + "issue:45479", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46178", + "issue:46202", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:33290|issue:34567", + "issue:43381|issue:43873", + "issue:43316|issue:43525", + "issue:44479|issue:45290", + "issue:44964|issue:45440", + "issue:43720|issue:44442", + "issue:34689|issue:43611", + "issue:42175|issue:46234", + "issue:32090|issue:33290", + "issue:44442|issue:45803", + "issue:43550|issue:44743", + "issue:43931|issue:45216", + "issue:43493|issue:43827", + "issue:34689|issue:36010", + "issue:43316|issue:45759", + "issue:45290|issue:46209", + "issue:43761|issue:44188", + "issue:43873|issue:44871", + "issue:42907|issue:45698", + "issue:43643|issue:44913", + "issue:44987|issue:46178", + "issue:44077|issue:44479", + "issue:44568|issue:45479", + "issue:44855|issue:45137", + "issue:33290|issue:34689", + "issue:43650|issue:46202", + "issue:44190|issue:44496", + "issue:29127|issue:33357" + ], + "prepared_review_unit_hash": "c2dabccf7bf08a3a55d78aba11249d3ad6762f50c5895b50c9a25c9c0c98a0ff", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35759, + "estimated_input_tokens": 8940, + "estimated_eval_tokens": 18136 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous Transformers issues with only loose thematic overlap around loading/config/serialization. No pair appears to be the same underlying bug or change, so I rejected all soft edges. If a single representative issue is needed, the broadest fit is the from_pretrained/save_pretrained roundtrip issue.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45698 is the broadest and most representative open issue in the set: it concerns a general from_pretrained/save_pretrained loading-path bug, has inbound references, and is more central than the narrower model-specific or config-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best single issue to represent the cluster because it is general enough to anchor a serialization/loading regression theme, whereas the others are mostly unrelated one-off model, backend, or training failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Different subsystems and symptoms: DeepSpeed/optimizer OOM vs TrainerState token counter not updating." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Eval-mode gradient checkpointing limitation is unrelated to quantization offloading behavior." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API issues, but they affect different fields and different model families; not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Different multimodal/chat-template regressions with different failure modes and code paths." + }, + { + "left": "issue:44964", + "right": "issue:45440", + "accept": false, + "reason": "Different model-loading/divergence bugs; one is a loading failure, the other is implementation parity with remote code." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet accelerate-loading unpacking bug is unrelated to AutoTokenizer loading FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but for different root causes and versions; not plausibly one merged fix." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Packaging/backend inclusion problem is unrelated to a missing utility symbol in import_utils." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast NoneType error and DeepSpeed Adafactor OOM are distinct failure classes." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer-loading failure and a bare-except bug in Fuyu tensor conversion are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44743", + "accept": false, + "reason": "Different model-specific runtime bugs: torch.compile/SDPA vs recurrent state reset with cache." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both involve Qwen models, but one is a weight-shape mismatch and the other is a bad checkpoint save regression." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "Model parity discrepancy and docs pipeline-removal cleanup are unrelated." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Different loading failures: multimodal model breakage vs GenerationMixin import path regression." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both mention config structure, but one is a field-discrepancy issue and the other is missing text_config unwrapping in composite multimodal configs." + }, + { + "left": "issue:45290", + "right": "issue:46209", + "accept": false, + "reason": "Chat template crash is unrelated to FP8 inference failure on B200 hardware." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVision hidden_states regression is not the same as the attention-kernel divergence under torch.compile." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantized offloading behavior and Gemma eos_token_id inconsistency are different issues." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Both touch save/load paths, but one is dequantized model saving and the other is custom-module loading after save_pretrained." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are config persistence problems, but they involve different mechanisms and model classes." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Model loading regression and missing BACKENDS_MAPPING entry for tensorflow_text are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "PatchTSMixer post_init validation has no relation to Qwen video-input regression." + }, + { + "left": "issue:44568", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer special-token insertion bug and sequence-classification zero-loss bug are different code paths." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 decorator parsing issue is unrelated to DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "DeepSpeed optimizer OOM and multimodal model loading regression are not the same bug." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Very different problems: low-information issue title vs Gemma tokenization roundtrip failure." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an example script is unrelated to model-type recognition for an OLMo checkpoint." + } + ] + }, + "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": 17, + "review_unit_index": 30, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:25251", + "issue:28282", + "issue:33290", + "issue:35141", + "issue:41950", + "issue:42617", + "issue:43065", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43475", + "issue:43504", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:43825", + "issue:44162", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44617", + "issue:44792", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45200", + "issue:45542", + "issue:45584", + "issue:45656", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234" + ], + "soft_pairs": [ + "issue:25251|issue:28282", + "issue:43295|issue:43825", + "issue:43550|issue:43701", + "issue:42617|issue:43575", + "issue:43606|issue:43701", + "issue:43540|issue:44514", + "issue:43316|issue:43716", + "issue:44871|issue:45584", + "issue:44162|issue:44792", + "issue:43316|issue:44877", + "issue:33290|issue:35141", + "issue:43575|issue:43742", + "issue:33290|issue:45656", + "issue:45803|issue:46170", + "issue:43404|issue:43441", + "issue:43295|issue:44871", + "issue:43065|issue:44617", + "issue:41950|issue:43475", + "issue:43646|issue:44913", + "issue:43452|issue:44496", + "issue:44871|issue:46050", + "issue:43761|issue:44479", + "issue:45127|issue:45200", + "issue:43572|issue:46178", + "issue:45542|issue:46234", + "issue:43232|issue:43388", + "issue:43504|issue:45698" + ], + "prepared_review_unit_hash": "79bf1557f817774166aa57d2a9eefb70e976ea4be07aee7b1cc7bc80edd03a05", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35191, + "estimated_input_tokens": 8798, + "estimated_eval_tokens": 17852 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated standalone issues spanning different models, pipelines, and backend/runtime regressions. None of the soft pairs look like true duplicates or the same concrete fix path, so all soft edges should be rejected.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must anchor the cluster, issue 45698 is the most broadly framed open regression and the most representative of a general load/save model-resolution problem.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45698 is the best cluster representative: it is open, broadly applicable, and concerns a common from_pretrained/save_pretrained failure mode rather than a model-specific edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Different failures: pipeline output nesting vs missing PyTorch import." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both touch v5 regressions, but one is custom processor/tokenizer access and the other is a pipeline error message issue." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA OOM for a model versus checkpoint key mismatch during resume are unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py runtime failure and TP load OOM are different code paths and symptoms." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both involve video/vision processors, but the concrete bugs differ: video input validation vs batched chat template padding behavior." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Config API discrepancy for Gemma3 text config is unrelated to image preprocessor/model dtype mismatch." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "EOS token config mismatch and empty-transcription generation failure are different problems." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 runtime breakage and janus image-generation test failure are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Different config issues: one API inconsistency, one strict config loading failure." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Deepspeed AdaFactor OOM is not the same as embedding resize reinitialization." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP load OOM and a model-loading key error are distinct failure modes." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention deepspeed, but one is OOM during optimization and the other is double optimizer stepping." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Bare except swallowing errors is unrelated to Windows ROCm import/dependency issues." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Mistral3 weight tying and FlashAttention failures are separate bugs." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression does not match an EOS token configuration inconsistency." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is a model implementation issue, not the same as Sam3Video OOM." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Pipeline choosing image processors and a missing vision encoder attribute are different defects." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Transformers 5 custom initialization breakage is unrelated to GPTNeoX rotary_pct reload behavior." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "GGUF loading regression and unrecognized model/config.json validation are not the same bug." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "EOS token mismatch and quantized model save/load breakage are different persistence/config issues." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIP hidden_states regression and Qwen video-input regression affect different models and code paths." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with extended vocab is unrelated to mm_token_type_ids defaults for text-only fine-tuning." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLmConfig pad_token_idx and tensorflow_text backend mapping are unrelated config/import issues." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend handling is unrelated to a missing import_utils helper." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync and metric gathering label truncation are different distributed-training bugs." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy field loading in Beit semantic segmentation is unrelated to incorrect custom module resolution after save/load." + } + ] + }, + "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": 17, + "review_unit_index": 31, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:35141", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43335", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43673", + "issue:43761", + "issue:43827", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44242", + "issue:44462", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45357", + "issue:45381", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:45987", + "issue:46082", + "issue:46121", + "issue:46234" + ], + "soft_pairs": [ + "issue:44077|issue:44534", + "issue:43673|issue:44743", + "issue:43761|issue:44242", + "issue:44792|issue:44977", + "issue:44617|issue:45584", + "issue:44534|issue:44913", + "issue:44479|issue:44734", + "issue:44190|issue:44964", + "issue:44792|issue:45600", + "issue:44855|issue:45588", + "issue:43827|issue:43994", + "issue:41720|issue:45245", + "issue:43065|issue:44610", + "issue:45803|issue:45987", + "issue:43931|issue:44060", + "issue:35141|issue:42490", + "issue:43335|issue:44855", + "issue:43644|issue:44913", + "issue:45632|issue:46234", + "issue:44077|issue:44877", + "issue:43550|issue:44190", + "issue:43976|issue:44855", + "issue:43540|issue:44734", + "issue:44521|issue:45381", + "issue:45127|issue:45357", + "issue:44462|issue:45907", + "issue:44737|issue:45803", + "issue:43335|issue:44190", + "issue:44112|issue:45600", + "issue:46082|issue:46121" + ], + "prepared_review_unit_hash": "f6aeace59e45c407cce127fad06c174fe09770a3aeb2a8f2f29738bdd417745e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36832, + "estimated_input_tokens": 9208, + "estimated_eval_tokens": 18672 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly noisy: the soft-similarity pairs share broad themes (generation, multimodal, config, CI) but do not look like the same underlying bug or change. The only near-duplicate in the item set is the non-persistent buffers regression (43644/44534), which is the best anchor issue, but none of the provided soft edges should be merged.", + "confidence": 0.93, + "canonical_issue_reason": "issue:43644 is the clearest and earliest report of the recurring 'non-persistent buffers filled with junk' regression, and issue:44534 is effectively a reworded follow-up of the same bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best cluster anchor because it has the most precise title and a direct duplicate-style re-report exists in the set; it is the strongest representative of the only obvious duplicate theme here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Both mention model init/post-init behavior, but one is about optional post_init in patchtsmixer and the other about non-persistent buffers being initialized with junk; different bugs." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both involve cache/state handling, but one is a missing GenerationMixin cache in chunked_prefill and the other is recurrent state reset in modular_qwen3_5; not the same code-path failure." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "Different subsystems and symptoms: CLIP vision hidden_states output vs load-balancing loss computation in router logits." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation failures, but they affect different models and appear to come from different mechanisms (test failure vs flash-attention behavior)." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Completely different issues: Sam3Video OOM versus Whisper empty-transcription failure after align_special_tokens." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is a v5 buffer-initialization regression; the other is a GPTNeoXConfig reload bug for rotary_pct. No shared underlying fix." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both are v5-era regressions, but one is multimodal video input handling and the other is KV-cache continuation indexing in transformers serve." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Unrelated: local dataset loading in a training script versus loading a specific multimodal model." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "A failed janus test and stale auto_mappings references are both CI-related only in the broadest sense; not the same bug." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Different failure modes: Python 3.13 IndentationError in torch.jit parsing versus flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is not the same as SigLIP2 producing nonsensical outputs with AutoModel/pipeline." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Unrelated: CUDA assert on A800 with auto device mapping versus a category-count limit error." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Different model/component bugs: a dummy Conv2d in Sam3PixelDecoder versus a processor producing the wrong image resolution for OmDet-Turbo." + }, + { + "left": "issue:45803", + "right": "issue:45987", + "accept": false, + "reason": "Bare except swallowing interrupts is unrelated to StaticCache.get_seq_length returning the wrong type." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Both concern Qwen3-family weight handling, but one is a model loading shape mismatch and the other is an incorrect tied-weights warning across layers." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinit after resize and path-with-trailing-slash import warning are unrelated bugs." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and DebertaV2 Python 3.13 import parsing are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "The first is a buffer initialization regression; the second is a config value not persisting on reload. Different root causes." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Cache-path collision for local trust_remote_code models is unrelated to the missing falcon-mamba availability helper." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Optional post_init in patchtsmixer and strict config blocking granite_speech are different config/init problems." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile+SDPA failure is unrelated to loading a local dataset in the image classification script." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Python version compatibility problem versus an IndentationError in a scripted model import; not the same issue." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Video-input ValueError in Qwen3OmniMoe is not the same as serve-time KV-cache continuation tensor indexing." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal-template regressions, but assistant_masks all zero and wrong vision_position_ids are different bugs and likely different fixes." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocabulary is unrelated to incorrect visual encoder keys saved in Qwen3.5." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Repository tokenizer resolution and list_repo_files revision forwarding are separate hub API behaviors." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet CPU-side torch.arange and Fuyu bare-except error handling are unrelated implementation bugs." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers config behavior and a dataset-loading script issue do not share the same underlying defect." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both are CI failures, but one is a stale test and the other is stale generated auto_mappings entries; different maintenance issues." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "LlamaConfig head_dim validation and rope-parameter JSON list handling are unrelated config-validation 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": 17, + "review_unit_index": 32, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42907", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43404", + "issue:43441", + "issue:43479", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43901", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44336", + "issue:44464", + "issue:44496", + "issue:44561", + "issue:44683", + "issue:44734", + "issue:44933", + "issue:44938", + "issue:44964", + "issue:45003", + "issue:45042", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45335", + "issue:45362", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45600", + "issue:45632" + ], + "soft_pairs": [ + "issue:43116|issue:44190", + "issue:45276|issue:45584", + "issue:44938|issue:45542", + "issue:43295|issue:45276", + "issue:43479|issue:45375", + "issue:45245|issue:45584", + "issue:44464|issue:45563", + "issue:36246|issue:41628", + "issue:44186|issue:44734", + "issue:43650|issue:44220", + "issue:44246|issue:45003", + "issue:44683|issue:45446", + "issue:43404|issue:45216", + "issue:44933|issue:45362", + "issue:38175|issue:43065", + "issue:43643|issue:44496", + "issue:44112|issue:44964", + "issue:44964|issue:45600", + "issue:45563|issue:45584", + "issue:43854|issue:44964", + "issue:36246|issue:38617", + "issue:44734|issue:45632", + "issue:44561|issue:45632", + "issue:43531|issue:45600", + "issue:43295|issue:45335", + "issue:42907|issue:43441", + "issue:43526|issue:43901", + "issue:41950|issue:44016", + "issue:44496|issue:45042", + "issue:43673|issue:44336" + ], + "prepared_review_unit_hash": "8f450094836d89b70db97e7500dbe86d9f2c475bea56f591631d27c49f82797a", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36575, + "estimated_input_tokens": 9144, + "estimated_eval_tokens": 18544 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated Transformers bug reports with a few topical similarities, but none of the soft pairs look like the same concrete bug/change. I would not merge any of the proposed pairs.", + "confidence": 0.82, + "canonical_issue_reason": "issue:41628 is the broadest, most user-facing regression in the set: a top-level import failure affecting the core `transformers` entrypoint.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:41628 is the best single anchor because it is generic, high-impact, and clearly described compared with the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Both are example-script failures, but one is multi-label classification output behavior and the other is local dataset loading in image-classification; different code paths." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated model behaviors: Gemma4 embedding resize propagation vs Whisper empty-transcription generation after token alignment." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "One is a Python 3.14 import/load failure; the other is a TensorBoard-without-TensorFlow backend issue. Different failures and subsystems." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Both mention tokenizer/embedding-related regressions, but the concrete bug is different: processor/tokenizer API breakage vs token-embedding resize not propagating in Gemma4." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-shape issues, but they affect different model families and different missing/default fields." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category-count runtime error is unrelated to Whisper empty transcription generation." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Both are generation-related, but one is inconsistent outputs under compiled forward and the other is a stale warning about `num_return_sequences`." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules; no evidence they stem from the same export bug." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer crashes on NER/padding are unrelated to KV-cache tensor indexing in the serve `/v1/responses` path." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No meaningful overlap: one is a vague data request, the other is an audio feature extraction bug." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance regression vs unsafe `sys.modules` access; related to import time, but not the same bug or fix." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both involve `flex_attention`, but one is compiled flex attention failing on torch>=2.9 and the other is an incorrect version check for AuxRequest import." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different model-specific regressions: tied `lm_head` weights in Mistral3 vs incorrect save_pretrained checkpoint state in Qwen3.5." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Nonexistent `image_utils` import and Qwen3.5 chat crash are unrelated failures." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "Different model families and symptoms: zero probabilities in SigLIP2 vs dummy Conv2d in Sam3PixelDecoder." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Config loading with `trust_remote_code` missing fields is distinct from an unrecognized model missing `model_type`." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI failure in GraniteSpeech tests is unrelated to Phi-4-multimodal model loading." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Phi-4-multimodal load failure and stale auto_mappings references to removed Sam3LiteText configs are separate issues." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale generation warning does not share the same code path as Whisper failing on empty transcription." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model loading issues, but they concern different checkpoints and different failure modes." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import errors, but they involve different missing names and different modules; not the same regression." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Server KV-cache continuation crash is unrelated to local trust_remote_code cache-path collisions." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Removal of `is_torch_fx_available` breaking remote-code models is unrelated to cache-path collisions for local models." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and stale Sam3LiteText auto-mapping references are different code paths." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both are embedding/processor-style regressions, but one is processor/tokenizer API breakage and the other is decoder embedding resize propagation in t5gemma." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral-3 are separate issues." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BEiT image processor label reduction bug is not the same as a documentation mismatch in text classification pipeline behavior." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and a notebook syntax error are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading and PIL backend torchvision dependency regression are different failures." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache missing in v5 chunked prefill is unrelated to ANSI codes emitted by loading_report." + } + ] + }, + "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": 17, + "review_unit_index": 33, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:32090", + "issue:35141", + "issue:38175", + "issue:39692", + "issue:41950", + "issue:42175", + "issue:43295", + "issue:43299", + "issue:43335", + "issue:43381", + "issue:43388", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43577", + "issue:43650", + "issue:43688", + "issue:43824", + "issue:43901", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44220", + "issue:44242", + "issue:44279", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44617", + "issue:44779", + "issue:44933", + "issue:45020", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45372", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:43335|issue:45593", + "issue:43441|issue:45588", + "issue:42175|issue:44279", + "issue:43388|issue:43688", + "issue:44075|issue:44368", + "issue:44016|issue:44220", + "issue:43299|issue:45216", + "issue:43381|issue:45593", + "issue:43901|issue:44496", + "issue:43650|issue:46209", + "issue:44496|issue:45759", + "issue:45335|issue:45584", + "issue:45372|issue:46234", + "issue:32090|issue:35141", + "issue:43479|issue:43525", + "issue:44242|issue:45200", + "issue:38175|issue:44617", + "issue:43577|issue:45216", + "issue:43299|issue:45357", + "issue:41950|issue:45020", + "issue:43824|issue:44933", + "issue:41950|issue:45823", + "issue:44779|issue:45812", + "issue:44279|issue:45405", + "issue:44351|issue:45278", + "issue:39692|issue:43901", + "issue:44016|issue:44062", + "issue:43295|issue:44779" + ], + "prepared_review_unit_hash": "a41bd36a97de021b518022b58d1be54f8c9ac9ddcbfa2db133258604bf477892", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35609, + "estimated_input_tokens": 8903, + "estimated_eval_tokens": 18062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: mostly unrelated Transformers bugs spanning model loading, tokenizer behavior, config/init issues, docs, and dependency/import regressions. None of the soft pairs look like the same underlying bug or change, so all soft edges should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43295 is the best anchor only in a weak sense: it is a concrete, user-facing regression report with clear reproduction, while the rest of the cluster is a mix of unrelated model-, config-, and docs-level problems.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43295 is the strongest single triage candidate because it is a specific regression with direct breakage and clearer scope than the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Different bugs: SwitchTransformers sparse-layer construction vs D-FINE auxiliary-loss omission." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Different flash-attention failures with different root causes and code paths." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Backend install/dependency packaging issue vs a generic transformers dependency problem." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metrics label truncation vs auxiliary-loss normalization are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD optimizer args is unrelated to a Qwen tie_word_embeddings warning." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error vs audio feature extraction bug; not the same issue." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both are Qwen regressions, but one is model loading and the other is save_pretrained checkpoint corruption." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to auxiliary-loss behavior." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs mismatch for return_all_scores vs an unrecognized-model config error." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "No meaningful relation; one title is malformed, the other is an FP8 inference failure." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Both involve model config/loading, but one is unknown-model detection and the other is AutoModel text_config unwrapping." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Different model behaviors: T5Gemma embedding resize vs Whisper empty-transcription generation." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Different missing imports/symbols affecting different subsystems and models." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast TypeError vs token-embedding reinitialization after resize are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Multimodal config defaults being created vs a missing pad_token_id attribute are different config bugs." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss handling vs Gemma 4 mm_token_type_ids defaulting are different concrete issues." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3Video OOM are unrelated failures." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Dtype propagation in Blip2 vs Qwen3.5 save_pretrained checkpoint bugs are different code paths." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 multimodal, but one is loading failure and the other is incorrect visual-encoder keys on save." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup bug vs remote_code compatibility regressions are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Different import failures from different modules and symbols." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Pipeline processor-selection bug vs Gemma4 device-map mismatch are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different implementations." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue vs an overly new MIN_PEFT_VERSION requirement are not the same bug." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "Single missing HybridCache export vs broad import errors after upgrade are not the same underlying change." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Different documentation problems in different examples and API surfaces." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error vs AddedToken keyword collision are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Different regression scopes: custom multimodal processor/tokenizer handling vs DeepSeek tokenizer output correctness." + } + ] + }, + "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": 17, + "review_unit_index": 34, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:42994", + "issue:43065", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43723", + "issue:43784", + "issue:43854", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44297", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:44987", + "issue:45081", + "issue:45276", + "issue:45325", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45712", + "issue:45823", + "issue:45834", + "issue:45907", + "issue:46132" + ], + "soft_pairs": [ + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:43299|issue:43854", + "issue:43650|issue:46132", + "issue:43976|issue:44987", + "issue:44462|issue:45081", + "issue:43450|issue:44279", + "issue:43976|issue:44279", + "issue:44297|issue:44462", + "issue:45081|issue:45907", + "issue:44466|issue:45276", + "issue:44521|issue:45325", + "issue:43441|issue:45834", + "issue:45600|issue:45712", + "issue:42994|issue:43317", + "issue:43653|issue:45593", + "issue:44623|issue:45397", + "issue:44623|issue:44749", + "issue:43065|issue:44898", + "issue:44261|issue:44263", + "issue:44016|issue:45823", + "issue:44361|issue:45593", + "issue:44671|issue:44779", + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:43901|issue:43994", + "issue:43450|issue:44617", + "issue:44297|issue:45397", + "issue:43784|issue:45542", + "issue:43525|issue:43881", + "issue:43901|issue:44246" + ], + "prepared_review_unit_hash": "75d501850e0977cd81c33423de60caa7a08c4ce636a3a92a5ab6504baed26f1d", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36711, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: the soft-linked pairs mostly share only broad vocabulary like loading, tokenizer, or model regressions, but they point to different subsystems and root causes. I would not merge any of the suggested soft edges.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue stands out because the items span unrelated bugs, regressions, docs, CI, and dependency problems rather than one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global canonical anchor for this cluster; the set is too diverse to represent one concrete bug or change.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Different failures: compiled flex_attention on torch>=2.9 vs missing TensorFlow backend handling for tensorboard-only installs." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Unrelated dependency/version problems: Python 3.9 runtime incompatibility vs an unreleased PEFT minimum-version bump." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Both are model-loading regressions, but for different models and distinct code paths; not the same underlying bug." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "Completely unrelated: a vague data-related issue vs AttentionInterface.register changing behavior." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Different issues: Python version support vs loading a specific model repository." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer-related, but separate bugs: tokenizer.json selection vs a Mistral regex patch crash." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor shape bug vs generic dependency issue; no shared code path." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Different classes of problems: interpreter support vs dependency resolution/installation." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Both involve tokenizers, but one is save_pretrained serialization metadata and the other is AutoTokenizer loading behavior." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regex crash vs repo file listing not forwarding revision; unrelated subsystems." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both touch weights/vocab resizing, but one is serialization inconsistency and the other is embedding resizing propagation." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal bugs: assistant mask generation vs rope index temporal scaling." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Different model-specific inference bugs: FlashAttention failure vs OCR index error." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are cleanup/CI-related, but one is stale auto_mappings references and the other is dummy class leakage." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantization/offload related, but one is saving quantized models and the other is loading dequantized models with device_map=auto." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration vs D-FINE auxiliary-loss logic; unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Different save/load issues: processor.save_pretrained missing files vs gemma-4 ZeRO3 from_pretrained behavior." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor saving bug vs unrelated data-filtering slowdown; no common defect." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model issues: dummy Conv2d in Sam3PixelDecoder vs Perceiver non-default resolution failure." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Unrelated bugs: missing rms_norm_eps precision issue vs torch.split return-value handling." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device mismatch during forward; unrelated." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError on tasks vs D-FINE auxiliary-loss behavior; unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions, but one is CamemBERT masked-LM output and the other is DeepSeek tokenizer behavior." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Different loading regressions for different model families; no evidence they are the same bug." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both mention attention kernels, but one is a model-specific FlashAttention issue and the other is compiled flex_attention on newer torch." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch vs incorrect SigLIP2 model outputs; unrelated." + }, + { + "left": "issue:43450", + "right": "issue:44617", + "accept": false, + "reason": "Video processor shape bug vs Sam3Video memory exhaustion; different failure modes." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save metadata issue vs gemma-4 ZeRO3 loading issue; unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError in sentence-transformers integration vs missing TensorFlow backend handling; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading/config issues: missing pad_token_id on Llama4Config vs GLM-4V loading failure." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs/behavior mismatch vs import performance regression; 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": 17, + "review_unit_index": 35, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43262", + "issue:43299", + "issue:43352", + "issue:43525", + "issue:43643", + "issue:43673", + "issue:43761", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44291", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44521", + "issue:44561", + "issue:44589", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44964", + "issue:45003", + "issue:45081", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45362", + "issue:45397", + "issue:45593", + "issue:46132" + ], + "soft_pairs": [ + "issue:43525|issue:43643", + "issue:44016|issue:44291", + "issue:43262|issue:44521", + "issue:44188|issue:44336", + "issue:41628|issue:43976", + "issue:44589|issue:45362", + "issue:43673|issue:43761", + "issue:44336|issue:44373", + "issue:16998|issue:22355", + "issue:42491|issue:43299", + "issue:44295|issue:45397", + "issue:44246|issue:44496", + "issue:44242|issue:45593", + "issue:44246|issue:44871", + "issue:41950|issue:44220", + "issue:44336|issue:44485", + "issue:43824|issue:43976", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43976|issue:44964", + "issue:42886|issue:46132", + "issue:44779|issue:45292", + "issue:44295|issue:44749", + "issue:43673|issue:44561", + "issue:44242|issue:44857", + "issue:45081|issue:45362", + "issue:43352|issue:45250", + "issue:44162|issue:45200", + "issue:43901|issue:45003" + ], + "prepared_review_unit_hash": "551a385fb73b4e9cb3b4d961f0f81cd1dcd47b73a05ecf4b17d1862e40dd43e1", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36333, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is loosely related around Transformers v5 regressions, especially import/export and model-loading breakages, but the individual items are mostly distinct bugs or docs issues. I do not see any true duplicate pairs here.", + "confidence": 0.37, + "canonical_issue_reason": "issue:41628 is the broadest and most generic user-facing regression in the set: a top-level import failure from `transformers`, which matches the general import/export theme better than the more model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:41628 is the best representative issue because it is concrete, broadly scoped, and closest to the common import/export regression pattern seen across the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both are config-related, but one is a missing `pad_token_id` field on Llama4 and the other is missing fields from `AutoConfig` with `trust_remote_code`; different failure modes and code paths." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "A notebook syntax error and an `init_empty_weights` TypeError are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both involve `apply_chat_template`, but one is an audio sampling-rate default and the other is multimodal assistant mask generation; not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence on `torch.compile` is unrelated to ANSI codes in loading reports." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Top-level import failure vs Python-version compatibility regression; not the same underlying issue." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Different model/chat failures with no shared concrete code path." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Chunked-prefill cache handling and CLIP hidden-states behavior are separate regressions." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI formatting and a docstring problem are unrelated." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A `model_max_length` question on DeBERTa-V3 is unrelated to a missing `transformers.onnx` module import." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE, but one is LoRA portability across versions and the other is model loading on v5; too different to be the same bug." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "A `position_ids` buffer access error is unrelated to the Gemma-4 zero3 `from_pretrained` issue." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowness is not the same as an unrecognized model/config loading failure." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both concern losses not being applied, but one is MoE load-balancing loss and the other is D-FINE auxiliary losses; different model families and mechanisms." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import performance and Gemma eos-token configuration are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video pipeline image-processor lookup and `_torch_extract_fbank_features()` are different subsystems." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Missing Qwen2.5 VL export and Python 3.9 compatibility are different issues." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols from different packages and likely different fixes." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Missing model-specific export vs missing `HybridCache` export; similar symptom, not the same concrete bug." + }, + { + "left": "issue:43976", + "right": "issue:44964", + "accept": false, + "reason": "Python support regression and inability to load a specific multimodal model are not the same underlying problem." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading and `AttentionInterface.register` behavior change are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer correctness and `resize_token_embeddings` output-weight syncing are different bugs." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "`position_ids` buffer access and a reported post-upgrade slowdown are not the same issue." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Both are v5 regressions, but one is chunked-prefill cache handling and the other is `trust_remote_code` breakage from a removed helper; not mergeable as one fix." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "MoE load-balancing loss omission and an AMP/CUDA crash in LwDetrImageLoss are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45362", + "accept": false, + "reason": "Mistral tokenizer regex patching and Qwen3.5 chat crashes are different code paths." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "The first is a specific unsupported-model warning for Flash Attention 2.0; the second is just a generic topic title with no evidence of the same bug." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 token-type defaults are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "A pipeline docs mismatch and unsafe `sys.modules[]` access are separate concerns." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 36, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36246", + "issue:43065", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43408", + "issue:43475", + "issue:43526", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44336", + "issue:44368", + "issue:44466", + "issue:44479", + "issue:44483", + "issue:44496", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45412", + "issue:45459", + "issue:45561", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45656", + "issue:45701", + "issue:45759", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46170", + "issue:46202" + ], + "soft_pairs": [ + "issue:36246|issue:43596", + "issue:43526|issue:45003", + "issue:44496|issue:45003", + "issue:45412|issue:46123", + "issue:45292|issue:45584", + "issue:43262|issue:44246", + "issue:45375|issue:45759", + "issue:43257|issue:45137", + "issue:45292|issue:45701", + "issue:44877|issue:45375", + "issue:43232|issue:44336", + "issue:45987|issue:46082", + "issue:45561|issue:45563", + "issue:44987|issue:45459", + "issue:44479|issue:46170", + "issue:43065|issue:43408", + "issue:44995|issue:45200", + "issue:43475|issue:43540", + "issue:44466|issue:45335", + "issue:45397|issue:45656", + "issue:44483|issue:46202", + "issue:45137|issue:45600", + "issue:44368|issue:44977", + "issue:43653|issue:44162", + "issue:44995|issue:45588", + "issue:44368|issue:45563", + "issue:44877|issue:46170", + "issue:43931|issue:44368", + "issue:44060|issue:44077" + ], + "prepared_review_unit_hash": "836ecc2a4da635fd2b18b8fb60d4386ff055ef6cfd79b464e2778510b7c9789f", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36613, + "estimated_input_tokens": 9154, + "estimated_eval_tokens": 18564 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly broad theme matches (DeepSpeed, Qwen, resize/tokenization, config handling) but not the same concrete bug or change. I would reject all proposed duplicate edges and keep these issues separate.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Different failures in unrelated code paths: a Qwen2.5-VL import error vs a DeepSpeed ZeRO3/BERT index error." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "One is a BeitImageProcessorFast label-reduction bug; the other is unsafe sys.modules access in modeling_utils. No shared underlying defect." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized-model loading for a specific OLMo checkpoint is distinct from the generic sys.modules access bug." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory not being released is unrelated to MaskGenerationPipeline dropping the final partial batch." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "resize_token_embeddings/output_embeddings mismatch is unrelated to Whisper empty transcription after align_special_tokens." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio sampling-rate defaults and import latency are unrelated problems." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen 3.5/3.6 multimodal configs, but one is a missing strict field and the other is AutoModelForCausalLM.from_config text_config unwrapping; not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both mention DeepSpeed/ZeRO, but one is Qwen3 MoE weight conversion and the other is an empty-deque IndexError. Different code paths and fixes." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "resize_token_embeddings behavior is unrelated to tokenizer-version-dependent tokenization changes." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both are config/strictness related, but they concern different models and different failure modes." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI codes in loading_report." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache return type contract and LlamaConfig head_dim validation are separate issues." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "A pytest-xdist file race is unrelated to a stale num_return_sequences warning." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "A model-loading regression for physical-intelligence/fast is unrelated to protobuf error masking in tokenizer decoding." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "A video-input regression in Qwen models is unrelated to Windows ROCm/torch.distributed import issues." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Both are Sam3-related, but a dummy Conv2d in the pixel decoder is not the same bug as a sam3_video vs sam3_tracker model-type warning." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale cache crash and Gemma 4 mm_token_type_ids defaults are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 VisionEncoderOutput missing a field and Qwen3OmniMoe video input validation are different failures." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both touch embeddings/weight tying, but one is serialization inconsistency and the other is resize_token_embeddings not propagating to decoder.embed_tokens." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Both mention DeepSpeed, but one is Gemma-4 zero3 loading and the other is double optimizer stepping. Different concrete problems." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "API request rejection and Gemma pipeline roundtrip failure on '\u2581' are unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "An empty-deque ZeRO3 crash is unrelated to stale auto_mappings references to removed Sam3LiteText configs." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tied-word-embeddings warning is unrelated to a flash-attention generation failure." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "A tokenizer special-token registration bug is unrelated to the broad ESM2 breakage report." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "A second-forward stale indexer cache crash is unrelated to a flash_attention.py s_aux=None AttributeError." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "The tied-weights warning and the stale num_return_sequences warning are different issues." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Granite speech strict config loading and Windows ROCm import problems do not share a concrete bug." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "A Qwen3-VL weight-shape mismatch is unrelated to tied-word-embedding warnings." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning in Qwen3-Next and patchtsmixer post_init policy 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": 17, + "review_unit_index": 37, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42371", + "issue:42886", + "issue:42994", + "issue:43232", + "issue:43452", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43827", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44530", + "issue:44623", + "issue:44661", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44933", + "issue:45003", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45229", + "issue:45230", + "issue:45310", + "issue:45356", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45478", + "issue:45561", + "issue:45656", + "issue:46077", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202" + ], + "soft_pairs": [ + "issue:43673|issue:44734", + "issue:42886|issue:45229", + "issue:45003|issue:45561", + "issue:44530|issue:46170", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44869|issue:45356", + "issue:43452|issue:45003", + "issue:43646|issue:44162", + "issue:44623|issue:45656", + "issue:44749|issue:45656", + "issue:43673|issue:43827", + "issue:43701|issue:44749", + "issue:44871|issue:46132", + "issue:44188|issue:46202", + "issue:45230|issue:45397", + "issue:43475|issue:45381", + "issue:43475|issue:43643", + "issue:42994|issue:43482", + "issue:44038|issue:44623", + "issue:28282|issue:30990", + "issue:46077|issue:46123", + "issue:43232|issue:43673", + "issue:42371|issue:45092", + "issue:43475|issue:44661", + "issue:41628|issue:44933", + "issue:42994|issue:43525", + "issue:45375|issue:45478", + "issue:45081|issue:45103", + "issue:45310|issue:45375" + ], + "prepared_review_unit_hash": "b9806d58f2825e31e82e8e7664d57a4640c9acdb36cf816dff9051ee5bf23a42", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36582, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: mostly Transformers v5-era regressions around model/config loading, tokenizers, generation, and save/load paths, but the soft pairs generally point to different concrete bugs rather than true duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43646 is the broadest, most central regression report in the set: it covers v5 model initialization breakage and can act as a representative umbrella for several adjacent load/config failures, even though the cluster is not a clean duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single issue to anchor the cluster because it is the most general and broadly applicable bug report among otherwise distinct failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both are generation/cache-related, but one is chunked prefill cache handling and the other is KV-cache continuation indexing; different code paths and symptoms." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are unrelated failure modes." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils and an xdist race on captured_info.txt are different bugs." + }, + { + "left": "issue:44530", + "right": "issue:46170", + "accept": false, + "reason": "PagedAttentionCache group-type crash on Qwen3.5 models is unrelated to Windows ROCm import/dependency issues." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Both involve save_pretrained, but tokenizer config mismatch and missing processor files are different save-time defects." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer task AttributeError and a Chinese performance regression are not the same bug." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer-related, but one is Whisper timestamp decoding and the other is Kimi-K2.5 codec/fix_mistral_regex behavior; not the same underlying issue." + }, + { + "left": "issue:43452", + "right": "issue:45003", + "accept": false, + "reason": "gguf_file loading failures and unsafe sys.modules access do not share the same concrete code path." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss and ESM2 breakage are different model families and failure mechanisms." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save_pretrained missing files and deepspeed optimizer double-step are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "A throughput regression is not the same bug as an optimizer step being called twice." + }, + { + "left": "issue:43673", + "right": "issue:43827", + "accept": false, + "reason": "Generation cache corruption and stale docs referencing removed pipeline APIs are unrelated." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint resume key mismatch and a data-processing slowdown are different problems." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior changes affect different subsystems." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile branching in attention kernels and Gemma pipeline roundtrip failure are distinct issues." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report provides no concrete match, and gemma-4 zero3 from_pretrained is a specific loading regression." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "SAM 3 Video missing vision output fields and Qwen2.5-VL video position-id errors are different multimodal bugs." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM 3 Video output-field failure and trust_remote_code AutoConfig field loss are unrelated." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and GGUF loading failure are opposite directions and different code paths." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe v5 breakage and processor save_pretrained missing files are not the same bug." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Missing PyTorch ImportError and Sentence Transformers loading hang are unrelated." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Gemma4 causal-mask construction and partial-batch pipeline result dropping are different behaviors." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and missing cache during chunked_prefill are separate generation bugs." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 settings API guidance and InternVL2 meta-init incompatibility are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 Video output shape bug and add-new-model-like mapping failure are different model-registration problems." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both mention import issues, but one is missing AutoImageProcessor export and the other is a nonexistent image_utils import; not the same defect." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving and missing pad_token_id on Llama4Config are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5 Moe-related, but one is a missing vision-config field dropped by strict serialization and the other is a generic from_pretrained error; too little evidence they are the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Mistral regex patch tokenizer crash and auto_docstring future-annotations crash are separate failures." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both concern Qwen3.5 Moe, but one is a generic from_pretrained error and the other is a missing deepstack_visual_indexes field in a vision config; not clearly 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": 1, + "component_count": 17, + "review_unit_index": 38, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43901", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44373", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44964", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45406", + "issue:45561", + "issue:45600", + "issue:45712", + "issue:45812", + "issue:46123", + "issue:46202" + ], + "soft_pairs": [ + "issue:45406|issue:46202", + "issue:41628|issue:44351", + "issue:42673|issue:43704", + "issue:44964|issue:45137", + "issue:42994|issue:43824", + "issue:44222|issue:45310", + "issue:42831|issue:44222", + "issue:36010|issue:41628", + "issue:44038|issue:45310", + "issue:42175|issue:45712", + "issue:44496|issue:45561", + "issue:43377|issue:44242", + "issue:45561|issue:45600", + "issue:43232|issue:44792", + "issue:42898|issue:44568", + "issue:43901|issue:45561", + "issue:42222|issue:45200", + "issue:30990|issue:35141", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:44188|issue:44373", + "issue:43065|issue:46123", + "issue:30333|issue:33290", + "issue:43232|issue:44188", + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:45356", + "issue:42898|issue:44779", + "issue:43874|issue:45092", + "issue:44734|issue:45103" + ], + "prepared_review_unit_hash": "9f99593b328e4a7ca99cac58dbbe212dde7eeb7d99e7609bb5e69a92c15e5d50", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36962, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a heterogeneous mix of unrelated Transformers bugs and regressions, with only one clear duplicate pair. Most soft edges should be rejected because they share only a broad subsystem or release/version theme, not the same concrete failure mode.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 42898 is the strongest anchor in this mixed set: it describes a concrete, reproducible v5 tokenizer regression with clear impact, while most other candidates are narrower model-specific breakages or unrelated failures.", + "canonical_pr_reason": null, + "best_issue_reason": "42898 is the best standalone issue here because it is a well-scoped framework-level regression with a specific symptom and broader user impact than the other, more fragmented reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Both are Gemma-related, but one is a missing processor tokenizer attribute and the other is a token roundtrip/pipeline mismatch; different bugs and code paths." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different symbols (`AutoImageProcessor` vs `HybridCache`) and different underlying breakages." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same reported bug: Qwen3ForCausalLM VRAM leak when used in multiple dataloader threads." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "One is a model load failure for Phi-4 multimodal, the other is a DeepSpeed ZeRO3 deque error; unrelated failures." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized model saving failure versus a missing model import; different functionality and failure modes." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both mention MoE/FP8-era models, but one is `save_pretrained` on FP8 MoE and the other is a `from_pretrained` load error for Qwen3.5 MoE; not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Accuracy regression in FineGrainedFP8 versus serialization failure in FP8 MoE; different concrete problems." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different missing imports (`GenerationMixin` vs `AutoImageProcessor`) and different parts of the API." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen MoE-related, but they describe different models/versions and different failure points." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Packaging dependency omission is unrelated to leftover dummy classes and repo-check leakage." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "A model config/load compatibility issue is unrelated to a pytest-xdist file race." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask batched-vs-single mismatch is unrelated to MoE load-balancing loss being skipped when router logits are off." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "A concurrent test-file race is not the same as stale `auto_mappings.py` references to removed configs." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug versus a Janus image-generation test failure; different code paths and symptoms." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but one is about `clean_up_tokenization_spaces` and the other about `add_special_tokens` not inserting BOS/EOS; not the same bug." + }, + { + "left": "issue:43901", + "right": "issue:45561", + "accept": false, + "reason": "Documentation mismatch for `TextClassificationPipeline` is unrelated to the xdist captured-info race." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose model breakage and Gemma 4 token-type-id handling are different model families and different failure mechanisms." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "A loading hang for Sentence Transformers is unrelated to embedding reinitialization after resizing token embeddings." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Training/eval dtype drift is unrelated to a DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Missing files in `processor.save_pretrained(...)` is a persistence bug, not the same as a model load error." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Attention-kernel divergence under `torch.compile` is unrelated to a docstring typo for `position_ids`." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "A dummy `nn.Conv2d` placeholder is unrelated to the final-partial-batch `is_last` handling bug." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting failure after training job failure is unrelated to an Adafactor/DeepSpeed OOM." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation-state update bug and attention-kernel branching divergence are different code paths." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one is cleanup-space behavior and the other is Granite token-ID mapping; different underlying issues." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "A missing `nn` import during sentence-transformers import is unrelated to a Qwen2.5-VL import error." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading in v5 and special-token insertion behavior in v5 are related only at a high level; they are distinct bugs." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "`clean_up_tokenization_spaces` regression is unrelated to Kimi tokenizer codec handling and warning behavior." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "A cleanup-space regression is not the same as Deepseek tokenizer producing incorrect results." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-processor helper methods and remote-code/meta-init incompatibility are different multimodal loading issues." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation tensor indexing in `transformers serve` is unrelated to the auto-docstring crash under future annotations." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is well-grounded in the packet: the only clear duplicate pair is 42673/43704, and the other soft edges are rejected conservatively based on distinct titles, symptoms, or code paths. The choice of 42898 as the anchor issue is reasonable given the title-level evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 39, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:41628", + "issue:42175", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43122", + "issue:43232", + "issue:43479", + "issue:43723", + "issue:43873", + "issue:43976", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44448", + "issue:44484", + "issue:44623", + "issue:44869", + "issue:44933", + "issue:44945", + "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" + ], + "soft_pairs": [ + "issue:44016|issue:46231", + "issue:44075|issue:46231", + "issue:44295|issue:45656", + "issue:44060|issue:45125", + "issue:42898|issue:44448", + "issue:45405|issue:45712", + "issue:42175|issue:44933", + "issue:44448|issue:46132", + "issue:44222|issue:44297", + "issue:43479|issue:44079", + "issue:44933|issue:45712", + "issue:44297|issue:45656", + "issue:43976|issue:44933", + "issue:44222|issue:44623", + "issue:44190|issue:46082", + "issue:30333|issue:34689", + "issue:42831|issue:44623", + "issue:43479|issue:44869", + "issue:30333|issue:36331", + "issue:45245|issue:46132", + "issue:30333|issue:35141", + "issue:28282|issue:30333", + "issue:45071|issue:46082", + "issue:41628|issue:42757", + "issue:44297|issue:45310", + "issue:44206|issue:45103", + "issue:43723|issue:44977", + "issue:43122|issue:46209", + "issue:44263|issue:45092", + "issue:43232|issue:45834", + "issue:43873|issue:44945", + "issue:44222|issue:44484", + "issue:44295|issue:44623" + ], + "prepared_review_unit_hash": "b91d287695169573644d24a8895dc7d346dd06f6891444ca2b2f50f20345a9e1", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36771, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose cluster of unrelated Transformers issues rather than a duplicate set. The soft-similarity pairs mostly share broad keywords like v5 regressions, imports, save_pretrained, or model loading, but they do not appear to describe the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.24, + "canonical_issue_reason": "No clear duplicate canon emerges; if one representative is needed, issue 43873 is the broadest active report and best cluster anchor, but it is still not a true duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the most general and still-open report in the set, so it is the least misleading representative, even though the surrounding issues are mostly distinct bugs or documentation problems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error vs documentation parameter-name mismatch; different artifacts and different failure modes." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer args being ignored is unrelated to a docs typo about set_training() parameters." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access error and double optimizer-step under deepspeed are separate code paths." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen model-specific bugs: tied-weights warning vs missing tensor-parallel plan." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both involve v5 behavior changes, but one is tokenization-space cleanup and the other is Pegasus output differences; not the same bug." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "PEFT version pinning issue is unrelated to leftover dummy classes leaking into dir(transformers)." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend/extra installation packaging issue differs from a missing image_utils import export." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Model output regression and attention registration behavior change are distinct problems." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained bug and tokenizer_config class mismatch are unrelated save_pretrained issues." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Phi4 multimodal config initialization is unrelated to ModelOutput key assignment." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Both touch import surface area, but one is a missing symbol and the other is dummy class leakage without torch." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config mismatch and deepspeed double optimizer step have no shared underlying cause." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility regression is not the same as a missing import from image_utils." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "FP8 MoE save_pretrained bug is different from missing files in processor.save_pretrained()." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow reporting on failed training jobs and Llama 3.2 Vision model loading are different subsystems and failures." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue is unrelated to processor save_pretrained missing files." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Multimodal config default-init bug and Whisper timestamp decode crash are separate model-path issues." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "Training job reporting and custom trainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-limit runtime error and attention registration behavior change are not the same underlying bug." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLflow job-state reporting and token embedding reinitialization are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "AutoModel missing PyTorch import error is unrelated to failed-job MLflow reporting." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are different config-layer problems." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different packages; both are import errors but not the same symbol or fix." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save_pretrained metadata bug and Qwen3.5 MoE from_pretrained regression are unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "LastFeatureExtractor center-arg regression and auto_docstring annotation crash are different failure modes." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading issue in v5 and Qwen3.5 flash-attention generation bug do not share a concrete root cause." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization drift across versions and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return-value handling in one model is unrelated to remote-code checkpoint meta-init incompatibility." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwarg update after sync_gpus and Kosmos2.5 OCR index error are different code paths." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading issue and incorrect pipeline-parallel output are not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained MoE bug and max_shard_size question are unrelated save_pretrained topics." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Registering position_ids as a buffer and processor missing files are separate serialization/model-state 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": 17, + "review_unit_index": 40, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30990", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43479", + "issue:43550", + "issue:43576", + "issue:43644", + "issue:43646", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:44038", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44485", + "issue:44514", + "issue:44568", + "issue:44704", + "issue:44829", + "issue:44871", + "issue:44913", + "issue:44945", + "issue:45092", + "issue:45103", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45406", + "issue:45753", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46132" + ], + "soft_pairs": [ + "issue:44373|issue:44485", + "issue:43761|issue:46082", + "issue:43479|issue:43644", + "issue:44483|issue:45406", + "issue:45092|issue:45357", + "issue:43723|issue:44871", + "issue:44263|issue:44360", + "issue:30990|issue:38175", + "issue:44514|issue:45103", + "issue:44261|issue:45092", + "issue:42907|issue:46050", + "issue:44829|issue:46002", + "issue:43479|issue:44913", + "issue:42831|issue:45310", + "issue:22355|issue:28282", + "issue:41628|issue:43824", + "issue:44263|issue:44373", + "issue:42831|issue:44038", + "issue:39401|issue:44568", + "issue:44945|issue:45753", + "issue:42886|issue:43066", + "issue:42757|issue:45278", + "issue:42886|issue:42994", + "issue:44117|issue:44704", + "issue:43576|issue:43784", + "issue:43550|issue:46132", + "issue:42371|issue:44263", + "issue:43646|issue:46002" + ], + "prepared_review_unit_hash": "c0121a48f777bee04abcbb2e4441ce17dd59e1a495f8b333d35ea863550960c0", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35714, + "estimated_input_tokens": 8929, + "estimated_eval_tokens": 18114 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is highly heterogeneous: most items are unrelated import/API regressions, tokenizer/model-specific bugs, or saving/loading issues. I did not find any soft pair that looks like the same concrete bug/change, so all soft edges should be rejected. If forced to anchor the cluster, 42886 is the cleanest narrow regression report to use as a representative issue.", + "confidence": 0.92, + "canonical_issue_reason": "issue:42886 is the most concrete, self-contained regression report with a narrow repro (tokenizer cache loading under HF_HUB_OFFLINE=1) and clear scope; the other issues are mostly unrelated bug classes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42886 is the most triage-friendly representative because it describes one specific failure mode and environment, making it a better anchor than the broader or more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo vs GLM-5 RoPE behavior; different bug classes and code paths." + }, + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "CLIP hidden_states regression vs LlamaConfig validation; unrelated functionality." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Multimodal config initialization vs non-persistent buffer junk; not the same underlying defect." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request rejection vs Gemma4Processor missing _tokenizer; different processor/server failures." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "InternVL2 meta-init incompatibility vs Qwen3.5 save_pretrained visual key bug; same broad area, different concrete failures." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading regression vs eos_token_id mismatch; related to tokenization but not the same bug." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Two different DSA/indexer issues with different symptoms (split outputs vs missing ReLU)." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Loading stall vs zero-probability output; different model/runtime failures." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Qwen2_5_VL batched template crash vs auto_docstring annotation crash; unrelated subsystems." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Layernorm epsilon precision issue vs meta initialization incompatibility; not the same code-path problem." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Dequantized model save failure vs sinq quantization save/load breakage; same broad area, different quantization flow." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Flash-attention training degeneration vs NaNs during checkout/init; unrelated training/init bugs." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Default multimodal config init vs GPTNeoX rotary_pct reload behavior; different config bugs." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 Moe from_pretrained error; different model and failure mode." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module vs missing PyTorch dependency; both import errors but not the same root cause." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Two separate missing imports for different symbols; no evidence they are the same underlying change." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Indexer tensor-splitting bug vs docstring mistake; clearly unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FP8 accuracy problem vs Qwen3-VL-Moe bug; different model behavior and code path." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer offset mapping bug vs BOS/EOS special-token regression; both tokenizer-related but distinct." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline parallelism output error vs image-specific Qwen3_5 infinite loop; different symptoms and fixes." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache load failure vs wrong decoder type in v5; separate tokenizer/config issues." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Single missing hub symbol vs broad upgrade-related import failures; 45278 is an aggregate report, not the same concrete bug." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache load bug vs quantized model saving failure; different operations and root causes." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping None handling vs AutoProcessor kwargs forwarding; different from_pretrained internals." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command vs sentence-transformers import NameError; unrelated CLI vs import regression." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile SDPA bug vs AttentionInterface.register behavior change; different APIs and code paths." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings warning/behavior vs DSA indexer tensor split issue; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization regression vs NaNs after checkout/init; different failure mechanisms and scope." + } + ] + }, + "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": 17, + "review_unit_index": 41, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:22355", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43723", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:44038", + "issue:44060", + "issue:44222", + "issue:44261", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45216", + "issue:45245", + "issue:45278", + "issue:45290", + "issue:45446", + "issue:45478", + "issue:45593", + "issue:45600" + ], + "soft_pairs": [ + "issue:44295|issue:44297", + "issue:30990|issue:33357", + "issue:43232|issue:45593", + "issue:44704|issue:45290", + "issue:44261|issue:44373", + "issue:42907|issue:44945", + "issue:43723|issue:44442", + "issue:43784|issue:45600", + "issue:43723|issue:44779", + "issue:43756|issue:44060", + "issue:41628|issue:45278", + "issue:43122|issue:44514", + "issue:42907|issue:43937", + "issue:43408|issue:43475", + "issue:43116|issue:45200", + "issue:43122|issue:44521", + "issue:42907|issue:44964", + "issue:44360|issue:44485", + "issue:44038|issue:44297", + "issue:44625|issue:44829", + "issue:43232|issue:45245", + "issue:44038|issue:44295", + "issue:28218|issue:30990", + "issue:42907|issue:43452", + "issue:43122|issue:44779", + "issue:36296|issue:43116", + "issue:44038|issue:44222", + "issue:39401|issue:44488", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:39401|issue:44521", + "issue:43784|issue:45446", + "issue:22355|issue:28218" + ], + "prepared_review_unit_hash": "16f7fe824ce783ac7981a43f34b5e530f6046d5168300e5eecab44e19ae5a36c", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37193, + "estimated_input_tokens": 9299, + "estimated_eval_tokens": 18854 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grab-bag of Transformers bug reports, mostly around imports, tokenizers, model loading, and generation. The soft pairs look superficially similar by keyword, but they do not describe the same concrete underlying bug.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical issue emerges: the items are mostly distinct bugs with only superficial overlap in subsystem or wording.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43723 is the broadest tokenizer-loading regression and is the closest thing to a thematic anchor, but it still does not represent the rest as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Different bugs: position_ids buffer access vs tokenizer_class metadata mismatch in save_pretrained." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Unrelated failures: Sentence Transformers loading hang vs MacOS bus error on a CLIP model." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug is unrelated to D-FINE auxiliary loss behavior." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Different code paths: AutoProcessor kwargs forwarding vs chat template crashing on tool-call messages." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA layernorm epsilon precision issue is unrelated to a docstring problem for position_ids." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Different problems: dequantized save_pretrained failure vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "Both involve AutoTokenizer loading, but one is a v5 generic issue and the other is a specific FastSpeech2Conformer tokenizer failure." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Importing sentence-transformers with missing nn is unrelated to stale auto_mappings references breaking CI." + }, + { + "left": "issue:43723", + "right": "issue:44779", + "accept": false, + "reason": "Generic tokenizer loading regression is not the same as Deepseek tokenizer producing wrong results." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different model-specific logic: Smollm3 RoPE layer count vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing AutoImageProcessor import is not the same as broad import errors after version upgrade." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer behavior change across versions is unrelated to Qwen2_5_VLProcessor batched chat-template crashes." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Model saving failure is unrelated to GLM-5 invalid GenerationConfig." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "These are different SAM 3 video issues: model-type warning vs missing fpn_position_embeddings attribute." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label classification example output bug is unrelated to Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer version drift is not the same as assistant_masks being all zero for multimodal chat templates." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Dequantized save_pretrained failure and Phi-4 multimodal load failure are separate model-loading/saving bugs." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer ReLU discussion is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe bug is not the same as tokenizer_class metadata mismatch on save_pretrained." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 degenerate training are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:45245", + "accept": false, + "reason": "Generation kwargs/sync_gpus behavior is unrelated to the 2^24 categories runtime limit." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Different issues: Qwen3-VL-Moe behavior vs reading position_ids after registering a buffer." + }, + { + "left": "issue:28218", + "right": "issue:30990", + "accept": false, + "reason": "Tokenizer adds a space after an added token is unrelated to Sentence Transformers loading getting stuck." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Dequantized save_pretrained failure is unrelated to gguf_file breaking AutoTokenizer/AutoModel loading." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Version-to-version tokenization drift is not the same as Deepseek tokenizer producing incorrect results." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug is unrelated to the multi-label classification example script returning empty results." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe bug and FP8 save_pretrained moe are distinct MoE-related problems, not one concrete fix." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 tokenizer offset_mapping mismatch is unrelated to failing to load cjvt/sleng-bert." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5, but one is save_pretrained checkpoint regression and the other is from_pretrained load error." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Different missing-symbol import errors: Qwen2_5_VLForConditionalGeneration vs HybridCache." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping bug is unrelated to assistant_masks being all zero for multimodal inputs." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "Sentence-transformers import NameError and flex_attention version-check bug are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing transformers.onnx module import and tokenizer added-space behavior are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 42, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44297", + "issue:44351", + "issue:44355", + "issue:44380", + "issue:44485", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44938", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45753", + "issue:46002", + "issue:46077", + "issue:46231" + ], + "soft_pairs": [ + "issue:42757|issue:44351", + "issue:22355|issue:29159", + "issue:43122|issue:43906", + "issue:44355|issue:46231", + "issue:43992|issue:44704", + "issue:43066|issue:44779", + "issue:43611|issue:46002", + "issue:43937|issue:44485", + "issue:43976|issue:45446", + "issue:42371|issue:43874", + "issue:42222|issue:43441", + "issue:42831|issue:44297", + "issue:44075|issue:44936", + "issue:45357|issue:45478", + "issue:44038|issue:45478", + "issue:43116|issue:43825", + "issue:43493|issue:45753", + "issue:43116|issue:43493", + "issue:45083|issue:45446", + "issue:44380|issue:46077", + "issue:43937|issue:45563", + "issue:22355|issue:29127", + "issue:44117|issue:45588", + "issue:39401|issue:44779", + "issue:33453|issue:39401", + "issue:38617|issue:45278", + "issue:44625|issue:46002", + "issue:42175|issue:44938", + "issue:43827|issue:45446" + ], + "prepared_review_unit_hash": "eda97432bc9648aaee3dfb393bf3a30d67d6981b095fb84aa2dd5d8ce101a2e8", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35908, + "estimated_input_tokens": 8977, + "estimated_eval_tokens": 18210 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous: tokenizer regressions, import/API breakages, model-specific bugs, and unrelated docs/tests. I don\u2019t see any soft pair that clearly describes the same concrete bug or change, so all soft edges should be rejected.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43122 is the broadest tokenizer-regression report in the set and best matches the recurring v4\u2192v5 behavior-change theme; the other issues are either narrower model-specific bugs or unrelated API/docs/import failures.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43122 is the best umbrella issue for this cluster because it is the most general and central tokenizer behavior regression, whereas the rest are mostly distinct one-off failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different packages (`huggingface_hub` vs `transformers`) and do not indicate the same bug." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "One is a missing module import for `transformers.onnx`; the other is a tokenizer consistency complaint. Different codepaths and failures." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Both mention tokenizer behavior, but one is a version-regression report and the other is an isolated reproduction of a different issue. Not the same concrete bug." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors and a documentation parameter-name mismatch are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "`UMT5Encoder` missing weights and `AutoProcessor` dropping kwargs are distinct loader/processor bugs." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes (decoder type vs incorrect results)." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "A `base_model_prefix` loading regression and NaNs in classification heads are unrelated issues." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "These are separate GLM-5 problems: invalid generation config versus RoPE implementation behavior." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version support failure and an incorrect PyTorch version check are different compatibility bugs." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing image-patch helper in GLM46V are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Different models and different failures: vitpose breakage versus FlashAttention support for Ministral-3." + }, + { + "left": "issue:42831", + "right": "issue:44297", + "accept": false, + "reason": "FineGrainedFP8 accuracy issues are unrelated to tokenizer save/load metadata mismatches." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer argument handling and `trainer.evaluate()` failing after training are separate trainer issues." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a save/load key regression and the other is a `from_pretrained` error. Not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "These are different Qwen/MoE model-loading problems with no clear shared codepath." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "A multi-label classification example bug and a misleading pipeline error message are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation drift and a Qwen3_5 infinite loop are different model-specific bugs." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Classification example behavior and SigLIP2 implementation discrepancy do not share the same underlying failure." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A Qwen3_omni_moe helper-function issue and a flex_attention import check are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both involve attention behavior, but they concern different models and different mechanics (scaling vs mask construction)." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig validation and a stale warning about `num_return_sequences` are not the same issue." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Missing module import and an unclear LayoutLMv3 error message are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping assumptions and a flash-attention `s_aux=None` crash affect different subsystems." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but one is an offset-mapping bug in Qwen3 and the other is a broader incorrect-results report for Deepseek v5." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression and wrong offset mapping are distinct tokenizer bugs with different symptoms." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import and a broad set of import errors after upgrade are related only at the theme level, not the same bug." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "`num_labels` propagation and NaNs in classification heads are separate classification-related problems." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Torch-only install backend coverage and Python 3.14 loading failures are different compatibility/install issues." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing `pipeline()` and an incorrect PyTorch version check are unrelated changes." + } + ] + }, + "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": 17, + "review_unit_index": 43, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43122", + "issue:43381", + "issue:43408", + "issue:43421", + "issue:43450", + "issue:43577", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:43986", + "issue:44075", + "issue:44117", + "issue:44261", + "issue:44291", + "issue:44410", + "issue:44466", + "issue:44483", + "issue:44484", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44945", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45072", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45405", + "issue:45446", + "issue:45464", + "issue:45542", + "issue:45656", + "issue:46002", + "issue:46178", + "issue:46231" + ], + "soft_pairs": [ + "issue:43064|issue:43673", + "issue:43381|issue:44945", + "issue:43450|issue:43986", + "issue:43937|issue:44977", + "issue:44466|issue:46002", + "issue:45405|issue:45446", + "issue:42673|issue:45072", + "issue:44117|issue:45081", + "issue:42222|issue:46178", + "issue:44410|issue:45310", + "issue:44829|issue:45216", + "issue:39401|issue:43421", + "issue:44291|issue:45071", + "issue:38617|issue:41628", + "issue:43408|issue:43742", + "issue:44877|issue:46231", + "issue:44484|issue:45310", + "issue:43937|issue:44496", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:44964|issue:45278", + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:44075|issue:45656", + "issue:42371|issue:44261", + "issue:43012|issue:43577" + ], + "prepared_review_unit_hash": "78b9a1e0f7a004ece0facdf09d4f1f1fa74118370f94adf9b8bedc0ad7d33420", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35557, + "estimated_input_tokens": 8890, + "estimated_eval_tokens": 18036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most pairs are only loosely related by theme (tokenizers, loading, training, video/multimodal, backend/dependency issues) and do not describe the same underlying bug or change.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45278 is the broadest umbrella report here (\u201cMany import errors after update from 4.57.0 to 5.5.0\u201d), so it is the best representative if one must be chosen, though the set is not a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 has the widest scope and is the most plausible sink for related version-regression reports, but it still does not cover the other concrete failures in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Different bugs: optimizer-state corruption under FSDP2/PEFT vs missing generation cache during chunked prefill." + }, + { + "left": "issue:43381", + "right": "issue:44945", + "accept": false, + "reason": "Eval-mode gradient checkpointing limitation is unrelated to pipeline-parallel output corruption." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video shape bug vs crash from missing torchvision dependency are different failure modes." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Invalid GenerationConfig for GLM-5 is not the same as Qwen3.5 flash-attention generation breakage." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization of tied weights depends on device; NaNs after checkout/init is a separate model-init issue." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Dependency version bump problem is unrelated to the flex_attention PyTorch version check bug." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multithreaded Qwen3.5 inference is distinct from bfloat16 dtype mismatches." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping None is a different issue from the Mistral regex patch crashing on backend_tokenizer access." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "Broken vitpose models and missing tensorflow_text backend registration are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing qwen3next projections is a different model-architecture bug from Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "flash_attention_3 training degradation is unrelated to the save_pretrained checkpoint regression." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Wrong offset_mapping and stale post-processor updates are separate tokenizer issues." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights argument handling differs from PretrainedConfig type-checking regression." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and code paths." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Model-type warning for sam3 is not the same as the MobileLLM key-error load failure." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "These both touch chat/completions, but one is a generic endpoint rejection and the other is a Qwen3.5 MoE model load error." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig and unrecognized model/config.json are different loading failures." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Wrong tokenizer decoder type and version-to-version tokenization changes are related only broadly, not the same bug." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Vitpose model breakage is unrelated to the undefined TensorFlow backend caused by tensorboard-only installation." + }, + { + "left": "issue:44964", + "right": "issue:45278", + "accept": false, + "reason": "Loading a specific multimodal model is not the same as a broad import-error regression after upgrading transformers." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but the concrete failures differ: endpoint validation vs streaming inference on Qwen3.5-0.8B." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Missing TensorFlow in the torch extra is a packaging/backend issue, not the vitpose model breakage." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD args being ignored is a different optimizer bug from Deepspeed calling step twice." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 behavior settings and missing rms_norm_eps precision error are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "PyTorch precision warning during compile is not the same as incorrect dtype propagation in Blip2 loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 44, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:38617", + "issue:41628", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42831", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43335", + "issue:43408", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44360", + "issue:44368", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44829", + "issue:44933", + "issue:44945", + "issue:44977", + "issue:44987", + "issue:45071", + "issue:45200", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45542", + "issue:45588", + "issue:46050", + "issue:46172", + "issue:46231" + ], + "soft_pairs": [ + "issue:44410|issue:45478", + "issue:38617|issue:44220", + "issue:43097|issue:44493", + "issue:43874|issue:44261", + "issue:41628|issue:45071", + "issue:44263|issue:44485", + "issue:42673|issue:44977", + "issue:45071|issue:45542", + "issue:42673|issue:43906", + "issue:44263|issue:46172", + "issue:44360|issue:46172", + "issue:43012|issue:43408", + "issue:44484|issue:45478", + "issue:43441|issue:44829", + "issue:42371|issue:42831", + "issue:43335|issue:45071", + "issue:43931|issue:44038", + "issue:43097|issue:43452", + "issue:44933|issue:46231", + "issue:44117|issue:44987", + "issue:43874|issue:44263", + "issue:43986|issue:44484", + "issue:43408|issue:44368", + "issue:43976|issue:44945", + "issue:45440|issue:46172", + "issue:43097|issue:46050", + "issue:43097|issue:45588", + "issue:43064|issue:44315", + "issue:42673|issue:45200", + "issue:44062|issue:44117", + "issue:38617|issue:42757", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:43931|issue:44410", + "issue:43502|issue:43992" + ], + "prepared_review_unit_hash": "3f3bcf1d9ccb496ed817481eb1112286a4d52961fbe7ac15fba77cb4490c6fa1", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38090, + "estimated_input_tokens": 9523, + "estimated_eval_tokens": 19302 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: the soft-similarity pairs mostly reflect broad subsystem or symptom overlap, not the same underlying bug/change. I would not merge any of the proposed soft edges, and there is no clear canonical issue/PR for this set.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out; the items span unrelated failures across imports, model config, attention, tokenizer, and training paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No best issue for duplicate consolidation: none of the issues appears to subsume the others as the same concrete bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen-family models, but one is a missing projection in qwen3next and the other is a Qwen3.5 MoE loading error; different code paths." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "Different failures: a missing transformers export vs an audio feature extraction issue." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "One is a removed config API warning; the other is an unexpected position-id key during model loading." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Unrelated model components: image processor method missing vs a layer-norm epsilon precision bug." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Import of AutoImageProcessor is unrelated to PretrainedConfig type-checking breakage." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Both touch GLM-related code, but the reported defects are different operators/modules and not the same bug." + }, + { + "left": "issue:42673", + "right": "issue:44977", + "accept": false, + "reason": "VRAM leak in dataloader threads is unrelated to flash-attention generation failure." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type-checking regression and TensorBoard/TF backend error are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "The second is an isolated reproduction of a separate issue, not the same VRAM leak." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different model families and different failure modes; no shared concrete bug." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "DSA ReLU concern vs a DeepSeek-V4 dtype/transpose issue; unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "A PyTorch precision warning while compiling bfloat16 is not the same as a sam3 model-type warning." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Save-pretrained shard size question is unrelated to Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both mention flash attention, but they describe different concrete bugs: inference failure vs degenerate training." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API guidance and FineGrainedFP8 accuracy regression are distinct topics." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "Sparse-layer construction bug and type-checking regression are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Both involve Qwen3-VL variants, but one is a weight-shape mismatch and the other is a broader v5/MoE loading issue." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "A removed embedding-tie config warning does not match a gguf_file loading breakage." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Import error and documentation parameter-name mismatch are not the same bug." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None is a different failure from loading physical-intelligence/fast on >=5.1.0." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Missing image-processor method and torch.split return-value issue are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Missing torchvision crash in AutoProcessor is unrelated to max_shard_size behavior." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "These are both warnings, but they concern different models and different config mismatches." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version incompatibility and incorrect pipeline-parallel output are different problems." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Different DeepSeek variants and different implementation details; not the same concrete defect." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "Embedding-tie removal warning is unrelated to quantized save/load corruption." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Config warning about removed API is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Distributed optimizer-state corruption in FSDP2/PEFT is unrelated to Liger Kernel application with model_init." + }, + { + "left": "issue:42673", + "right": "issue:45200", + "accept": false, + "reason": "VRAM leak in multithreaded loading and Gemma 4 mm_token_type_ids defaulting are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer AddedToken keyword collision is different from TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different missing symbols from different packages and code paths." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing projections in qwen3next is unrelated to a strict config field being dropped." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "One is a Qwen3-VL-Moe bug and the other is qwen3next layer-projection omission; not the same issue." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight mismatch on Qwen3-VL-30B is a different failure mode from missing attn projections in qwen3next." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "The first is unintended network access with local_files_only, while the second is a missing embed_tokens weight on UMT5Encoder." + } + ] + }, + "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": 17, + "review_unit_index": 45, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:38617", + "issue:42490", + "issue:42617", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43526", + "issue:43550", + "issue:43792", + "issue:43824", + "issue:43825", + "issue:43873", + "issue:43906", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44117", + "issue:44162", + "issue:44222", + "issue:44230", + "issue:44261", + "issue:44351", + "issue:44368", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44704", + "issue:44938", + "issue:44964", + "issue:45103", + "issue:45290", + "issue:45341", + "issue:45372", + "issue:45478", + "issue:45561", + "issue:45588", + "issue:45698", + "issue:45753", + "issue:45859", + "issue:46123", + "issue:46172", + "issue:46178" + ], + "soft_pairs": [ + "issue:45341|issue:45859", + "issue:38617|issue:44351", + "issue:43097|issue:43873", + "issue:44261|issue:44485", + "issue:44964|issue:45341", + "issue:44704|issue:45103", + "issue:43986|issue:44117", + "issue:42617|issue:45341", + "issue:43986|issue:45290", + "issue:43066|issue:43906", + "issue:43986|issue:45372", + "issue:43208|issue:43550", + "issue:45341|issue:45588", + "issue:45341|issue:45561", + "issue:43992|issue:44484", + "issue:43986|issue:45341", + "issue:43012|issue:43550", + "issue:43825|issue:44655", + "issue:43012|issue:44368", + "issue:42907|issue:44655", + "issue:42907|issue:45753", + "issue:44162|issue:46172", + "issue:43792|issue:46123", + "issue:44938|issue:46178", + "issue:43208|issue:43526", + "issue:44483|issue:44514", + "issue:44222|issue:45478", + "issue:43012|issue:44683", + "issue:42490|issue:43012", + "issue:42757|issue:43824", + "issue:45698|issue:46172", + "issue:44038|issue:44230" + ], + "prepared_review_unit_hash": "83118e4eea221276b9093dff609782c7110b7a12886a8ddb27319078cd49d741", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36960, + "estimated_input_tokens": 9240, + "estimated_eval_tokens": 18736 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: import errors, processor/chat-template bugs, save/load regressions, compiler warnings, and model-specific failures. None of the soft pairs look like the same underlying bug or change, so no duplicates should be merged here. No PRs are present.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45341", + "right": "issue:45859", + "accept": false, + "reason": "Different subsystems: a testing_utils bug versus a Qwen3.5 RoPE CPU-offload issue." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols in different modules (`layer_type_validation` vs `HybridCache`)." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "One is a removed API warning/regression, the other is an offloading/quantization behavior issue; not the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "MLA q_a_layernorm epsilon handling and GLM-5 RoPE implementation are unrelated code paths." + }, + { + "left": "issue:44964", + "right": "issue:45341", + "accept": false, + "reason": "Model loading regression versus a testing_utils bug; no shared failure mode." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and auto_docstring annotation parsing are separate problems." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Different failures: missing torchvision for video model loading versus TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "A 3d_parallel.py runtime failure is unrelated to the testing_utils bug." + }, + { + "left": "issue:43986", + "right": "issue:45290", + "accept": false, + "reason": "Processor loading without torchvision and chat-template tool-call handling are different code paths." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and a reproduction of another issue are not the same concrete bug." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Video-model loading without torchvision is unrelated to the ReasoningEffort import crash in Gemma 4 loading." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs and a Bamba torch.compile/SDPA failure are distinct model-specific issues." + }, + { + "left": "issue:45341", + "right": "issue:45588", + "accept": false, + "reason": "Testing-utils bug versus flash_attention s_aux=None crash; different code paths and symptoms." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Testing-utils bug versus pytest-xdist race on captured_info.txt; not the same issue." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "UMT5 from_pretrained missing weights and save_pretrained max_shard_size default are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Video-model loading failure is unrelated to the testing_utils bug." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A precision warning during compile is not the same as a Bamba SDPA compile failure." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task-support messaging and Pipeline save_pretrained failure are different defects." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "These are different warnings with different triggers: bf16 compile precision versus tie_word_embeddings in LoRA fine-tuning." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized models and saving Pipeline objects are different save_pretrained behaviors." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Model save failure and Qwen3.5 infinite-loop inference bug are unrelated." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and deepseek-v4 tensor-op failure are separate model-specific bugs." + }, + { + "left": "issue:43792", + "right": "issue:46123", + "accept": false, + "reason": "Whisper not running and MaskGenerationPipeline dropping final partial-batch results are different subsystems." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 load failure and a tensorflow_text backend mapping error have different causes." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs do not match the BeitImageProcessorFast reduce_labels bug." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Both involve chat/completions behavior, but one is API acceptance and the other is batched apply_chat_template crashing; not the same fix." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained moe and Qwen3.5 moe from_pretrained errors are different load/save paths." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Both mention compilation, but one is a bf16 warning and the other is a flex_attention failure on torch>=2.9." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "A trailing-slash local model import warning is unrelated to a bf16 compile warning." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different missing names and different import surfaces." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Wrong custom module after save_pretrained and a deepseek-v4 tensor operation bug are not the same underlying issue." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both mention Qwen/MoE/FP8, but one is a generic transformers 5.0 bug and the other is FP8 inference support; not one concrete duplicate." + } + ] + }, + "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": 17, + "review_unit_index": 46, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42994", + "issue:43064", + "issue:43066", + "issue:43299", + "issue:43352", + "issue:43441", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43673", + "issue:43761", + "issue:43867", + "issue:43881", + "issue:43901", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44117", + "issue:44242", + "issue:44355", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:45216", + "issue:45276", + "issue:45290", + "issue:45310", + "issue:45375", + "issue:45478", + "issue:45479", + "issue:45753", + "issue:46077", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46202", + "issue:46209" + ], + "soft_pairs": [ + "issue:38617|issue:43502", + "issue:43673|issue:46123", + "issue:44483|issue:45290", + "issue:43618|issue:43761", + "issue:43064|issue:44380", + "issue:43066|issue:46209", + "issue:43299|issue:43867", + "issue:36683|issue:43475", + "issue:45276|issue:46077", + "issue:42994|issue:44655", + "issue:43901|issue:44655", + "issue:36010|issue:42757", + "issue:44625|issue:46077", + "issue:43611|issue:44655", + "issue:43352|issue:43937", + "issue:43867|issue:45216", + "issue:46077|issue:46082", + "issue:43502|issue:44485", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:44380|issue:46132", + "issue:36683|issue:44242", + "issue:36683|issue:43881", + "issue:43906|issue:45753", + "issue:43299|issue:45753", + "issue:43352|issue:43976", + "issue:43992|issue:45310", + "issue:43441|issue:44380", + "issue:43352|issue:45478", + "issue:43867|issue:46202", + "issue:44355|issue:44855", + "issue:44380|issue:45479", + "issue:43352|issue:44380", + "issue:43867|issue:45375", + "issue:36683|issue:44117", + "issue:43618|issue:44380" + ], + "prepared_review_unit_hash": "8d3b813d737a6a08dede34cf795a7aff10880d55875552143f80453892b9427e", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "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 grab-bag of distinct Transformers v5 regression reports (imports, config fields, model loading, attention backends, save/load paths, and a few model-specific bugs). The soft pairs are thematically similar in places, but none look like the same concrete underlying bug/change.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43611 is the broadest, clearest regression-framed report in the set and a reasonable anchor for the model-loading breakage theme, but the cluster is too heterogeneous for a strong duplicate canon.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the most representative single issue to keep as a triage anchor because it is concise, actionable, and closest to the common v5 loader-regression surface area.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Both are import-related, but the missing symbol (`layer_type_validation`) and the local-files-only network leak are different APIs and different failure modes." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Both involve generation/pipeline behavior, but one is a missing cache during chunked prefill and the other is final partial-batch result dropping in MaskGenerationPipeline." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Chat/completions request handling vs `apply_chat_template` crash on tool-call assistant messages are different code paths and bugs." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one is missing `attentions` and the other is `hidden_states=None`; different outputs and likely different fixes." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "Trainer/FSDP/PEFT optimizer-state corruption is unrelated to GPT-2 attention scaling being ignored under SDPA/FlashAttention." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Wrong tokenizer decoder type and FP8 inference failure on B200 are unrelated issues." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are loading failures, but one is Qwen3VL MoE model loading and the other is a generic sorted-state_dict load error; not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Missing `vocab_size` in Gemma3Config vs missing `fpn_position_embeddings` in Sam3VisionEncoderOutput are distinct model-specific attribute errors." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4-related, but resize-token-embeddings propagation and bidirectional-attention mask construction are different behaviors." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving and pipeline object serialization both involve saving, but they are different failure paths with different objects and APIs." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Documentation mismatch for `return_all_scores` is unrelated to pipeline `save_pretrained` failures." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Separate import errors from different packages/symbols (`GenerationMixin` vs `is_offline_mode`)." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 `num_labels` propagation and Gemma4 attention-mask behavior are unrelated config/model bugs." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Model loading with `base_model_prefix` and pipeline save serialization are different code paths and fixes." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Nemotron FlashAttention 2 support gating and GLM-5 `GenerationConfig` validation are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Generic load error with sorted state_dict and Qwen3.5 checkpoint corruption from `save_pretrained` are distinct save/load bugs." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 causal-mask bug and LlamaConfig `head_dim` validation are unrelated config/attention issues." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Local-files-only still making API requests is unrelated to GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing-config-attribute reports, but they concern different models and different fields (`vocab_size` vs `pad_token_id`)." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attention output regression and Qwen3.5 save_pretrained corruption are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Attention scaling in GPT-2 backends and `AttentionInterface.register` behavior changes are different problems." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Missing `vocab_size` and load-balancing loss omission are different subsystems and failure modes." + }, + { + "left": "issue:36683", + "right": "issue:43881", + "accept": false, + "reason": "A missing config attribute in Gemma3 is not the same as a GLM-4V model loading failure." + }, + { + "left": "issue:43906", + "right": "issue:45753", + "accept": false, + "reason": "An isolated reproduction of an earlier issue and a Qwen3_5 infinite-loop-on-image report are not the same bug." + }, + { + "left": "issue:43299", + "right": "issue:45753", + "accept": false, + "reason": "Qwen3VL MoE loading failure and Qwen3_5 image infinite loop are different model/runtime issues." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "FlashAttention support issue and Python-version compatibility issue are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Missing `embed_tokens.weight` in UMT5Encoder loading is not the same as Qwen3.5 MoE `from_pretrained` failure." + }, + { + "left": "issue:43441", + "right": "issue:44380", + "accept": false, + "reason": "Both mention attention/FlashAttention, but one is unsupported FA2 for Ministral-3 and the other is ignored GPT-2 attention scaling; different concrete bugs." + }, + { + "left": "issue:43352", + "right": "issue:45478", + "accept": false, + "reason": "These are different issues entirely: Nemotron FlashAttention support vs Qwen3.5 MoE `from_pretrained` error." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Sorted-state_dict load failure and Gemma pipeline roundtrip on the underscore token are unrelated save/load behaviors." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled Python-file errors and Python 3.13 `@torch.jit.script` parsing/indentation errors are different import-time problems." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Ignored attention scaling and degenerate zero loss with `num_labels=1` are unrelated model-training bugs." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "FlashAttention support gating is unrelated to GPT-2 scaling being ignored in other attention backends." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Load failure on sorted state_dict and missing `deepstack_visual_indexes` in Qwen3_5MoeVisionConfig are different issues." + }, + { + "left": "issue:36683", + "right": "issue:44117", + "accept": false, + "reason": "Missing config field in Gemma3 and tokenizer-mapping returning `None` are unrelated failure modes." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP attentions missing and GPT-2 attention scaling ignored are different regressions affecting different models 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": 17, + "review_unit_index": 47, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43097", + "issue:43352", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45125", + "issue:45278", + "issue:45341", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:46132" + ], + "soft_pairs": [ + "issue:41669|issue:44246", + "issue:43097|issue:43352", + "issue:42890|issue:45561", + "issue:43064|issue:44492", + "issue:43976|issue:44938", + "issue:43867|issue:45357", + "issue:43992|issue:45478", + "issue:36296|issue:43575", + "issue:43064|issue:44655", + "issue:28282|issue:33453", + "issue:43618|issue:44857", + "issue:44655|issue:44938", + "issue:43618|issue:44977", + "issue:43618|issue:44077", + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:43502|issue:44062", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:44222|issue:45125", + "issue:42890|issue:45341", + "issue:43576|issue:46132", + "issue:42548|issue:44351", + "issue:44355|issue:44749", + "issue:45125|issue:45478", + "issue:30333|issue:33453", + "issue:44387|issue:44556", + "issue:45125|issue:45375", + "issue:43010|issue:45278", + "issue:30990|issue:31795", + "issue:43723|issue:44556", + "issue:43502|issue:44222" + ], + "prepared_review_unit_hash": "7a1ebaf494bfc02a1f5fabd23bcca71e73106e8a8b0c5b8505535d5fe8f368b6", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36754, + "estimated_input_tokens": 9189, + "estimated_eval_tokens": 18634 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a duplicate cluster; the issues span unrelated import, loading, training, CI, docs, and model-specific regressions. None of the soft pairs looks like the same underlying bug/change.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the set is heterogeneous and does not form one duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong best issue: there isn't a single report that plausibly subsumes the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both mention import slowdown, but one is a concrete import-star cleanup proposal and the other is an intermittent performance complaint; not the same bug." + }, + { + "left": "issue:43097", + "right": "issue:43352", + "accept": false, + "reason": "Different problems: removed config flag behavior vs Flash Attention support for NemotronH." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Both are test/CI failures, but one is missing set_seed flakiness and the other is an xdist file race." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Unrelated: distributed optimizer state corruption vs a typo in cache strategy docs/code." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version compatibility complaints, but they target different releases and failure modes." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Both involve loading/saving state, but one is a sorted-state_dict load error and the other is incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different model-specific load bugs; no clear shared code path." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Both mention tensor parallelism, but one is a training bug and the other is an OOM on model load." + }, + { + "left": "issue:43064", + "right": "issue:44655", + "accept": false, + "reason": "Optimizer-state mismatch in training is unrelated to Pipeline save_pretrained support." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "AutoModel missing PyTorch and tokenizer loading regression are unrelated components." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIPOutput attentions regression and an AMP/CUDA crash in LwDetrImageLoss are different bugs." + }, + { + "left": "issue:44655", + "right": "issue:44938", + "accept": false, + "reason": "Pipeline serialization and Python 3.14 import/load failures are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "CLIPOutput attentions assignment is unrelated to Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "No shared bug: CLIPOutput output fields vs patchtsmixer post_init allowance." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "Both discuss model internals, but the DSA indexer ReLU issue and DeepSeekV3MoE divergence are not the same change." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Test seeding flakiness is unrelated to a model loading failure." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Local-files-only network leak and AddedToken keyword collision are different bugs." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are CI/test issues, but the flaky seed problem and GraniteSpeech device-override failure are distinct." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Test flakiness and auto_mappings breaking CI due to removed configs are not the same underlying problem." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Both concern Qwen3.5/MoE, but save_pretrained FP8 behavior and missing _tp_plan for tensor parallelism are different code paths." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "A generic testing-utils bug and SamHQ set_seed flakiness are not clearly the same defect." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken env command and changed AttentionInterface.register behavior are unrelated API issues." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and likely different export paths." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled-Python-file errors and post-upgrade filtering slowdown are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but missing _tp_plan and generic from_pretrained error are not enough to treat as one concrete bug." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow job reporting on failure and tokenizer loading regression are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory growth and checkpoint reload compatibility are different regressions." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "A missing tensor-parallel plan and a dropped vision-config field are separate config/schema issues." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache update no_grad semantics and broad import errors after upgrade are not the same bug." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers hanging on load and confusing forward-arg docs are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "AutoTokenizer.from_pretrained failure and v4-to-v5 checkpoint reload failure are different loading problems." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Network calls despite local_files_only and FP8 save_pretrained on MoE 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": 17, + "review_unit_index": 48, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:30333", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43010", + "issue:43299", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43576", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:43976", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44355", + "issue:44360", + "issue:44368", + "issue:44373", + "issue:44393", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45468", + "issue:45542", + "issue:45656", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46209" + ], + "soft_pairs": [ + "issue:42831|issue:43874", + "issue:43827|issue:44961", + "issue:43502|issue:44961", + "issue:43010|issue:45850", + "issue:43010|issue:44368", + "issue:30333|issue:36296", + "issue:44355|issue:45071", + "issue:44315|issue:44360", + "issue:44355|issue:44655", + "issue:33453|issue:34689", + "issue:43299|issue:44556", + "issue:43576|issue:45071", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43010|issue:45656", + "issue:43576|issue:43824", + "issue:44222|issue:44230", + "issue:44961|issue:45468", + "issue:31795|issue:38617", + "issue:41628|issue:42548", + "issue:43010|issue:43976", + "issue:43576|issue:44938", + "issue:44908|issue:44961", + "issue:44315|issue:46209", + "issue:36296|issue:43901", + "issue:36296|issue:43526", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:44038|issue:45125", + "issue:44230|issue:46095", + "issue:44273|issue:45003", + "issue:44393|issue:45071", + "issue:44373|issue:44961", + "issue:43576|issue:44556", + "issue:43525|issue:45850", + "issue:44393|issue:45753" + ], + "prepared_review_unit_hash": "512ad6c1090d0d50a0c4038ee9c747f6d7f0ec2b85a75814d4ce233e20a3ecf3", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37808, + "estimated_input_tokens": 9452, + "estimated_eval_tokens": 19160 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous transformer v5 regressions, import/load failures, and docs-only issues; the soft links look like broad topical similarity rather than true duplicates. No PRs are present.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43576 is the cleanest concrete regression report in the set and the closest thing to an umbrella user-facing failure, though it still does not truly subsume the rest of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43576 is the best single representative because it is specific and actionable; the other issues are either model-specific, documentation-only, or unrelated runtime regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "Different problems: FP8 accuracy vs a missing image-processor method/AttributeError." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs about pipeline removals are unrelated to the unrelated 'racoon' report." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "local_files_only network leakage is a distinct bug from the unrelated 'racoon' issue." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A no_grad decorator change for cache/layer updates is not the same bug as a broad 'v5 breaks many models' complaint." + }, + { + "left": "issue:43010", + "right": "issue:44368", + "accept": false, + "reason": "No_grad/update semantics are unrelated to the tie_word_embeddings warning during LoRA fine-tuning." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow reporting on failed jobs and tensor-parallel training are unrelated code paths." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file execution errors do not match the separate PretrainedConfig type-checking regression." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application in model_init is unrelated to the DSA indexer/ReLU discussion." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file errors are not the same as pipeline save_pretrained failures." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both involve loading, but one is tokenizer loading regression and the other is a specific Llama 3.2 Vision model-loading failure." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are load failures, but for different versions and model families; not the same concrete bug." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI env command breakage and PretrainedConfig type checking are different failures." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "An import error from huggingface_hub is unrelated to the transformers env command breakage." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training." + }, + { + "left": "issue:43010", + "right": "issue:45656", + "accept": false, + "reason": "Cache/layer update decoration and double optimizer stepping under deepspeed are different bugs." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "A broken env command and a missing Qwen2.5-VL export are separate issues." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; different bugs." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "The 'racoon' placeholder issue is unrelated to Gemma4 audio positional encoding." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Doc confusion on forward arguments is unrelated to the layer_type_validation import error." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Two different missing-symbol import errors for different APIs; not one underlying bug." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache update decoration and Python-version compatibility are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "The env command issue is not the same as Python 3.14 load failures." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs handling is unrelated to the 'racoon' report." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger Kernel/model_init and FP8 inference on B200 are different code paths." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training and text-classification docs are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training and BeitImageProcessorFast label reduction are different bugs." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "A missing-files save_pretrained bug is not the same as a vague placeholder bug report." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading and a TensorFlow backend error from tensorboard-only installs are separate issues." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Qwen3-VL-Moe breakage and missing _tp_plan for Qwen3.5 MoE are related only by theme, not the same bug." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a DeepSeek weight-conversion question are unrelated." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading malfunction is not the same as unsafe sys.modules access." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output problems are unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:44373", + "right": "issue:44961", + "accept": false, + "reason": "A docstring fix for position_ids is unrelated to the 'racoon' issue." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different version/regression reports with different symptoms and upgrade scenarios." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a model-specific config bug, not the broad v5 breakage complaint." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Both involve Qwen3_5/Qwen3-VL, but one is bounding-box output and the other is an infinite loop on a specific image." + } + ] + }, + "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": 17, + "review_unit_index": 49, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43519", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43704", + "issue:43827", + "issue:43874", + "issue:43976", + "issue:44077", + "issue:44246", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45071", + "issue:45092", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45468", + "issue:45478", + "issue:45850", + "issue:46018", + "issue:46143", + "issue:46144", + "issue:46172", + "issue:46218" + ], + "soft_pairs": [ + "issue:41669|issue:43673", + "issue:44556|issue:45092", + "issue:43572|issue:45850", + "issue:45362|issue:45850", + "issue:41669|issue:46218", + "issue:44246|issue:44273", + "issue:42890|issue:44908", + "issue:43408|issue:44315", + "issue:44704|issue:44908", + "issue:44297|issue:45230", + "issue:43976|issue:44393", + "issue:44295|issue:45230", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:44908|issue:45071", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:45278|issue:45850", + "issue:43673|issue:44492", + "issue:43874|issue:45850", + "issue:46143|issue:46144", + "issue:43976|issue:46218", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44492", + "issue:44315|issue:44829", + "issue:42371|issue:44492", + "issue:43576|issue:44273", + "issue:44655|issue:44861", + "issue:44360|issue:45468", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:43576|issue:46218", + "issue:43827|issue:44492", + "issue:46018|issue:46172", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689" + ], + "prepared_review_unit_hash": "9d863233e03de2f7b1ae51ecec64557cc2168aa341fe224cbe4e45d1eb4ae9b9", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39266, + "estimated_input_tokens": 9817, + "estimated_eval_tokens": 19890 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose grab-bag of Transformers issues, mostly v5 regressions and import/model-loading bugs, but the soft pairs generally describe different concrete code paths rather than duplicates.", + "confidence": 0.63, + "canonical_issue_reason": "issue:45850 is the broadest umbrella report about v5 breaking many models, so it best represents the cluster theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best triage anchor because it captures the widest cross-model impact among the v5 regression reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Both mention import/loading behavior, but one is about import * bloat and the other about chunked_prefill cache behavior." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 upgrade breakages, but they affect different model-loading paths and different failure modes." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "One is a StableLmConfig missing field regression; the other is a broad complaint about many models breaking." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Qwen chat crash is a specific model/API failure, not the same bug as the broad v5 breakage report." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import-time slowdown from model imports is unrelated to a missing Qwen3_5 module import error." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Both concern imports/lazy loading, but one reports slow import timing and the other says lazy loading is broken." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seeding flakiness and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Different subsystems: model-type mismatch warning versus Liger Kernel application in model_init." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and inverse_sqrt scheduler kwargs are unrelated bugs." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer_class save metadata mismatch is not the same as an unspecified bug report." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility and Qwen3-VL bounding box output are different problems." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer access error is unrelated to the generic bug report." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention support warning and Qwen3-VL output errors are different code paths." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env CLI and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs omission and PretrainedConfig type checking are different issues." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import * cleanup/import overhead is not the same as a missing HybridCache export." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Model import bloat and a testing_utils bug do not share the same underlying change." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad v5 regressions, but one is about import errors and the other about general model breakage." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Chunked_prefill cache missing and a cache-strategy typo are not the same bug." + }, + { + "left": "issue:43874", + "right": "issue:45850", + "accept": false, + "reason": "Missing image-patch helper in GLM46V is a specific multimodal API gap, not the broad v5 regression." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "One is missing kwargs passthrough, the other is attention implementation wiring; related area but different fixes." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python compatibility and a Qwen3_5 import failure are unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different model families and different failure reports." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init validation and tied weight key AttributeError are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Documentation still referencing pipeline() is unrelated to scheduler kwargs handling." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion and tokenizer loading regression are different issues." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Import * cleanup and a typo in cache strategies are not duplicates." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application and flash_attention_3 training degeneration are different concrete bugs." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API guidance and a cache-strategy typo do not match." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command and lazy loading failure are adjacent but distinct issues." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained and tied weight key handling are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer ReLU discussion and Gemma4 audio positional encoding are unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Generic bug report versus a concrete Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading failure and inverse_sqrt scheduler kwargs are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Env CLI breakage and Qwen3_5 module import failure are different bugs." + }, + { + "left": "issue:43827", + "right": "issue:44492", + "accept": false, + "reason": "Docs referencing pipeline() and a cache-strategy typo are not the same issue." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Both are DeepSeek-V4 related, but one is about gating and the other about tensor transpose/casting." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp math and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation and scheduler kwargs are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import cleanup/performance and VRAM leak in multithreaded use are different failure modes." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion and Llama 3.2 Vision model loading breakage are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 50, + "review_unit_count": 50, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-433", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:37934", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43668", + "issue:43704", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44351", + "issue:44485", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46144", + "issue:46218" + ], + "soft_pairs": [ + "issue:31515|issue:31795", + "issue:44485|issue:45468", + "issue:43298|issue:46095", + "issue:43704|issue:44485", + "issue:45850|issue:46218", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:43606|issue:45468", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44861|issue:45290", + "issue:43525|issue:44861", + "issue:43704|issue:44998", + "issue:43572|issue:44861", + "issue:43066|issue:44998", + "issue:41669|issue:44998", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:44998|issue:46218", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "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": "c810d020f658a421d92c689d57ce5b3691e8e6fe33e5f65186aea132e6efcdac", + "component_budget": { + "node_count": 433, + "item_count": 433, + "soft_pair_count": 1595, + "serialized_chars": 826503, + "estimated_input_tokens": 206626, + "estimated_eval_tokens": 413508 + }, + "budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 42, + "serialized_chars": 36087, + "estimated_input_tokens": 9022, + "estimated_eval_tokens": 18300 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the issues span unrelated model bugs, tokenizer regressions, config/import errors, docs, and performance complaints. None of the soft links look like the same underlying bug or change, so all are rejected.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Different problems: slow checkpoint loading vs confusing forward() docs for input_ids/past_key_values." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Unrelated model internals: GLM-5 RoPE implementation vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Different topics: continuous batching support for audio models vs a DeepSeek-V4 weight conversion question." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in multithreaded dataloading is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Generic v5 breakage report vs a specific Qwen3.5 import failure; not the same bug." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Local-files-only network leakage vs a Qwen3VL timestamp calculation bug." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings deprecation is unrelated to the VRAM leak report." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Different processor bugs: timestamp math vs save_pretrained missing files." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Timestamp calculation bug and VRAM leak are unrelated code paths." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return values in GlmMoeDsaIndexer is unrelated to Gemma-4 audio encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Different model components: MLA q_a_layernorm epsilon handling vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Different missing symbols: HybridCache import vs Qwen3_5ForCausalLM import." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support gating and a module import failure are distinct issues." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both tokenizer-related, but one is a general v4-v5 behavior change and the other is a specific decoder-type mismatch." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to the unrelated 'Unemployment' issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key handling and chat template tool-call crashes are separate failures." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to _get_tied_weight_keys crashing." + }, + { + "left": "issue:43704", + "right": "issue:44998", + "accept": false, + "reason": "VRAM leak in dataloader threads is unrelated to the unrelated issue 44998." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key AttributeError." + }, + { + "left": "issue:43066", + "right": "issue:44998", + "accept": false, + "reason": "Tokenizer decoder type regression is unrelated to the unrelated issue 44998." + }, + { + "left": "issue:41669", + "right": "issue:44998", + "accept": false, + "reason": "Import-star performance cleanup is unrelated to the unrelated issue 44998." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Checkpoint load slowness and missing layer_type_validation import are different bugs." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:44998", + "right": "issue:46218", + "accept": false, + "reason": "The unrelated issue 44998 has nothing in common with a Qwen3.5 import error." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow reporting on failed training jobs is unrelated to slow from_pretrained checkpoint loading." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading performance and tensor-parallel training bugs are different code paths." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request is not the same as a Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Slow from_pretrained loading and tokenizer loading regression are distinct issues." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching support for audio models is unrelated to fp8 inference support." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention interface are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value usage and DeepSeek-V4 shared expert gating are different model bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "ModernBERT norm_eps type hint and tokenizer regex warnings are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Unused ignore_value and kwargs not passed to loss calculation are different defects." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss kwargs propagation is unrelated to RoFormer attention implementation." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value bug does not match the tokenizer regex warning regression." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to RoFormer attention implementation." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Unused ignore_value and a typo in quantization_operation are unrelated." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Passing kwargs to loss and ModernBERT type hints are separate issues." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Incorrect type hint and a typo in a conversion string are unrelated." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4TextL2Norm normalization question is unrelated to ModernBERT norm_eps typing." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value usage is unrelated to a ModernBERT type-hint bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "review_unit_index": 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": "174914c79743347044c48a7bcd83514e9ffd8ed6e3fab526224639a2284c3f4e", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three items are unrelated feature requests: truncation behavior in tokenization, ReduceLROnPlateau-like scheduler functionality, and Wav2Vec2 word time offsets. The soft similarities are superficial, so none should be merged as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "issue:10519 is the clearest and most self-contained feature request, with a specific API change request and explicit scope for both truncation modes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:10519 is the best standalone representative because it is concrete, narrowly scoped, and describes an actionable enhancement rather than a broad question.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different domains and problems: truncation strategy vs Wav2Vec2 word boundary timestamps. No shared code path or concrete duplicate relationship." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "One asks for tokenizer truncation-from-beginning behavior; the other asks for learning-rate scheduler functionality. These are unrelated features." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Scheduler feature request and Wav2Vec2 timestamp extraction are distinct tasks with no overlapping implementation target." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "eb758d59b360229cf07a5ccab9b6bb0bd7ba3f7e0815a712eec08c04a5715017", + "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": "No duplicate-quality match here. The only related pair is two model-loading complaints, but they describe different failure modes; the tokenization issue is unrelated.", + "confidence": 0.78, + "canonical_issue_reason": "Issue #16618 is the most general and representative of the shared model-loading theme, so it works best as the cluster\u2019s anchor despite not being a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #16618 is the strongest standalone issue: it describes a concrete user-facing failure to load a named model, which is more central than the narrower auth/error-message report and far more relevant than the tokenization bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different subsystems and symptoms: tokenization parity vs. model loading failure. No evidence they stem from the same bug." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both are model-download/load related, but one is specifically a private-model login/error-message problem and the other is a generic failure loading bert-base-uncased; not the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "93d12371ab4d5ce713c3f024ddea9c008b91c986115303281b2fbc79149aa99b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two distinct community-contribution feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share a similar framing but not the same underlying change or code path.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 26350 is the better cluster anchor because it is older and has more discussion activity, but it is not a duplicate of 33260.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 26350 is the most suitable representative for the cluster only by age/activity; the requests themselves are unrelated features.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different feature requests in different subsystems: Flash Attention 2 support vs GGUF support. Similar contribution template, not the same bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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:46222", + "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:46222", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:46222", + "issue:43834|issue:44365", + "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": "d2d9dea9ea57b8bc1e6293bfc4ed48f2a83022d4a60f235294aa023180637702", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 25, + "serialized_chars": 13675, + "estimated_input_tokens": 3419, + "estimated_eval_tokens": 7094 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 25, + "serialized_chars": 13675, + "estimated_input_tokens": 3419, + "estimated_eval_tokens": 7094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of i18n translation-request issues. The specific-language tickets (Turkish, Spanish, Romanian) are distinct from each other, while the placeholder-title i18n issues are near-identical and look like duplicate template instances.", + "confidence": 0.78, + "canonical_issue_reason": "issue:46222 is the best representative of the duplicate placeholder i18n template issues: it is current, open, and structurally matches the other generic translation-template tickets more closely than the language-specific requests.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46222 is the strongest cluster anchor because it represents the generic translation-template form and appears to be the latest active instance among the near-identical placeholder issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different concrete translation requests; Turkish vs placeholder generic template are not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:46222", + "accept": false, + "reason": "Turkish translation request is distinct from the generic placeholder translation template." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language/request scope; not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Specific Turkish translation request does not duplicate the generic placeholder ticket." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Turkish and Romanian translation requests are separate language tasks." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish and Spanish translation requests are separate issues, not duplicates." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request vs generic placeholder template; different underlying ticket." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Spanish translation request is distinct from the generic placeholder issue." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian translation requests are different language-specific issues." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Spanish request does not match the generic placeholder translation template as the same issue." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian translation request vs placeholder template; not the same concrete issue." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Romanian request is distinct from the generic placeholder translation ticket." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian translation task is separate from the placeholder template issue." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Different language-specific request versus generic placeholder template." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian translation request and the Thai-titled issue do not look like the same underlying ticket." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Both are near-identical placeholder i18n translation-template issues with the same title/body pattern." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": true, + "reason": "These look like duplicate instances of the same generic translation-template issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same placeholder translation-template issue, no meaningful distinction visible." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Thai-titled issue is not clearly the same generic placeholder translation template." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Near-identical placeholder i18n translation-template issues." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": true, + "reason": "Same generic placeholder translation issue, likely duplicate." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "No clear evidence that the Thai-titled issue is the same underlying placeholder ticket." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": true, + "reason": "These are effectively the same placeholder translation-template issue." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "The Thai-titled issue does not match the placeholder generic translation template closely enough to merge." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder translation-template issue and the Thai-titled issue are not clearly the same ticket." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the generic template issues ([i18n-] Translating docs to ) are extremely close to each other, while the language-specific tickets are clearly different by title and are reasonably kept separate. The soft-edge verdicts are conservative overall, and the canonical issue choice (46222) is supported by its open state and recency among the near-identical template issues." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-16", + "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" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43837|issue:45322", + "issue:43837|issue:43909", + "issue:43909|issue:44205", + "issue:43671|issue:43909", + "issue:44121|issue:45322", + "issue:43837|issue:44121", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133" + ], + "prepared_review_unit_hash": "41501de06e65798bd4a694373a44742731df0e6039b194238226a014400d318c", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are all separate model-request issues, not duplicates. The similarity is only at a broad topical level (\u201cadd support for a model\u201d), while the requested models and modalities differ materially.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue: each item requests a different model or feature (e.g., EuroBert, RF-DETR, Qwen3-TTS/ASR, ModernVBERT, Voxtral TTS, etc.), so they are not the same underlying request.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue for deduplication: none of these issues subsumes the others, and the overlap is only the generic 'add model support' theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model requests: EuroBert config support vs DEIMv2. Same broad theme, but not the same underlying change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config support and Qwen3-TTS support are unrelated model additions with different modalities." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert vs RF-DETR are different models and different code paths; not duplicates." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are distinct model requests, not one concrete duplicate." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR (vision detection) and LFM2.5 Audio 1.5B are different modalities and changes." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and OpenAI weight-sparse transformer are separate model requests with no shared concrete implementation target." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR (vision) vs Qwen3-ASR (audio speech recognition) are different underlying features." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are different models and different task domains." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions; only the broad 'add support' theme matches." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT vs OpenAI weight-sparse transformer are different architectures and request targets." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT is a text model; LFM2.5 Audio 1.5B is an audio model request." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are different model requests, not the same underlying bug/change." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT models are distinct requests despite both being BERT-like names." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different model features." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and the weight-sparse transformer request are separate model additions." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 architecture and ModernVBERT are different concrete model support requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral-4B-TTS-2603 are different model families and modalities." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models and Qwen3-ASR are different additions; not mergeable as one duplicate." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are distinct model support requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and BERT-like models with RoPE overlap only loosely on BERT naming, but request different implementations." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like models with RoPE and Voxtral TTS are unrelated features." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS vs Qwen3-ASR are related product names but different concrete modalities and code paths." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different TTS model requests, not the same model or change." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are different task domains (speech vs OCR)." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and the weight-sparse transformer request are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are different model requests in different domains." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are both audio-related but not the same concrete model or implementation." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model requests; no shared code-path bug/change." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are different audio model requests, not duplicates." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "OpenAI weight-sparse transformer and EUPE are distinct model requests with different targets." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and OpenAI weight-sparse transformer are unrelated requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral-4B-TTS-2603 are different speech models and different support changes." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral-4B-TTS-2603 are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and implementation areas." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are different model requests from different authors/tasks; only broad vision-language similarity exists." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "143e8a03e19b65b70ab0116efff78e12ac55d0fb431defe7b2fbf0a0402e669c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one is a specific decoding feature request, the other is a model-support request for DeepSeek v3.2. The only similarity is very broad AI/model infrastructure overlap, which is not enough to merge them.", + "confidence": 0.97, + "canonical_issue_reason": "Issue #41084 is the more concrete, narrowly scoped request and is easier to use as a representative anchor for this cluster, but it is not actually the same problem as #43037.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #41084 is the better canonical-style representative because it is more specific and actionable than the broader model-support request in #43037.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Reject: the titles describe different underlying needs. 'Set Block Decoding' is a decoding behavior/feature request, while 'DeepSeek v3.2 support' is about adding support for a specific model. They do not appear to be 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": 8, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "a88e3b9538761bfd0a28116276a700b7863ccf192e03b3bc73ec5db9bec7c3d1", + "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 about different Trainer concerns: one asks to revisit `remove_unused_column` for customizability, while the other asks to account for custom trainers when estimating FLOPS. They share a broad Trainer/customization theme but not the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": "No canonical issue: the reported behaviors are different code paths and different requested fixes.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a duplicate of the other; if one must be kept as the cluster representative, 43598 is the more concrete Trainer API/customization request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different problems: `remove_unused_column` behavior versus FLOPS estimation for custom trainers. Similar subsystem, but not the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "78f0280bf96ec4989303cfd373eb4e7f3f3107ec015e52b0d259b08ea937d8b1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: one requests Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5/Qwen3_5Config.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43684 is broader and more directly about model registration support, making it the better representative of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43684 is the strongest standalone issue because it describes a concrete registration gap for a specific model family and auto-class pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families and different auto classes; they do not describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43835-4", + "nodes": [ + "issue:43835", + "issue:43845", + "issue:43846", + "issue:43859" + ], + "soft_pairs": [ + "issue:43835|issue:43845", + "issue:43835|issue:43846", + "issue:43835|issue:43859", + "issue:43845|issue:43846", + "issue:43845|issue:43859", + "issue:43846|issue:43859" + ], + "prepared_review_unit_hash": "eb593e8fb5177a9792ad2fa3e98c968bdac801c3daf61f4157d4e17fc62ce838", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4267, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly low-signal: three issues share the identical vague title \"huggingface \" and one issue is titled \"Fett\". With no bodies, targets, filenames, or references, there isn\u2019t enough evidence to confirm any pair as the same underlying bug, so no soft matches are accepted.", + "confidence": 0.36, + "canonical_issue_reason": "Issue 43845 is the earliest of the repeated \"huggingface \" issues and is the most representative placeholder entry in this cluster; 43835 has a different title and looks less central.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the best single issue to anchor the cluster because it is one of the repeated identical titles and appears earliest among them; however, the issue data is too sparse to make a stronger choice.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no descriptive content; soft similarity alone is not enough to conclude the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no supporting details such as bodies, targets, or references." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no evidence they describe the same bug or change." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "Identical vague titles, but there is no body or other evidence showing they refer to the same underlying problem." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same vague title, but insufficient concrete evidence to merge them as duplicates." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Same vague title, but no substantive details to establish a 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": 11, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "9fd6c2da38d2735c4c4f21f448eb8c77497184da309263d9d8b07fb4a3e4bd46", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like a set of unrelated low-quality/spam issues with generic titles; there is no evidence of a shared bug report or duplicate workflow.", + "confidence": 0.96, + "canonical_issue_reason": "No issue clearly represents a real underlying problem, so there is no sensible canonical issue for this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44418 is the only one with a mildly descriptive non-spam title, but it still does not identify a concrete bug and is not a good duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are titled \"\" and provide no concrete bug detail; similarity is not enough to indicate the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders with no actionable description, so they cannot be treated as duplicates of a specific bug." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam and the other is a generic documentation suggestion; no shared underlying defect is evident." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "A spam placeholder and a generic company-name title do not indicate the same issue or change." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A vague documentation suggestion and spam placeholder are unrelated and do not describe the same problem." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "The titles point to different kinds of low-signal reports; there is no concrete common bug or request." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "They are both spam placeholders, but there is no evidence they refer to the same underlying issue rather than separate junk reports." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "No concrete defect or matching request is described, so the pair is not a valid duplicate match." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "The records do not share a specific bug, code path, or actionable request; reject as duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "5d3e530f4bba004dd2a5e30158177fff5349308e2c8d9feb71cac95f26d6ff93", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues look highly similar and likely represent the same topic, but the evidence is not strong enough to call them a definitive duplicate. Issue 45145 remains the best canonical candidate because it is earlier and already has an inbound reference.", + "confidence": 0.72, + "canonical_issue_reason": "45145 is the earlier issue and has the only inbound reference, so it is the best canonical record if these turn out to be duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "45145 is the strongest issue to keep as the primary record because it predates 45151 and has inbound reference activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "The titles are identical and the bodies are similar, but the packet provides only soft-similarity evidence with no direct linkage or verbatim confirmation, so treating them as the same report is too strong." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The canonical choice is reasonably grounded (45145 is earlier and has an inbound reference), but the duplicate/soft-edge claim is too strong for the packet. We only have identical titles, similar body lengths, and a soft-similarity signal; there is no direct evidence that the issues are 'verbatim' or definitively the same report. A more conservative verdict would be warranted." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45672-2", + "nodes": [ + "issue:45672", + "issue:45674" + ], + "soft_pairs": [ + "issue:45672|issue:45674" + ], + "prepared_review_unit_hash": "19a37480826bcbdc63fe8f18ced2749e9a1be302b5fb5c67addc51ecd45192a4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate 4-bit inference bugs in different code paths: one in Gemma4 audio modules hitting torch.finfo() on uint8 weights, and one in BitsAndBytesConfig module skipping that breaks the default lm_head exclusion. They should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Both mention 4-bit inference, but the failures are different: one is a torch.finfo() TypeError on uint8 weights in Gemma4 audio modules, the other is an AssertionError caused by llm_int8_skip_modules clearing the default lm_head exclusion. Different triggers, code paths, and fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "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": "b62a40d5708b6ac6f81d39025f5cd791366c1d4f5c0cb7276cf0112d2d302f82", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two closed, body-less issues that only contain different personal names and birthdates. They do not describe a product bug or change, and there is no evidence they share the same underlying problem. Treat them as unrelated, low-signal/spam-like reports rather than duplicates.", + "confidence": 0.99, + "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": "The titles are different personal-name/date strings with no shared bug description, code path, or issue context; they do not look like the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "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": "1cebfc293e8f894ebde778985bad8f565ebc9a49b5a2843d30eeffb22a92e29b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues are not duplicates: they are unrelated name/date entries with no shared problem statement, links, or discussion context.", + "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": "Different named entities and dates; no evidence they describe the same bug, request, or code issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45869-5", + "nodes": [ + "issue:45869", + "issue:45870", + "issue:45871", + "issue:45872", + "issue:45873" + ], + "soft_pairs": [ + "issue:45869|issue:45870", + "issue:45869|issue:45871", + "issue:45870|issue:45871", + "issue:45870|issue:45872", + "issue:45870|issue:45873", + "issue:45871|issue:45872", + "issue:45871|issue:45873", + "issue:45872|issue:45873" + ], + "prepared_review_unit_hash": "0f018cdb21c1db38a5160dfabaee3c072f88390111d3c732395fd92343e0308a", + "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": "This cluster splits into two exact duplicate pairs plus one unrelated dataset promo issue: 45870/45871 are the same example issue, 45872/45873 are the same DataForge scraping-service issue, and 45869 is a different financial data bundle request.", + "confidence": 0.96, + "canonical_issue_reason": "45870 is the oldest and cleanest representative of the exact duplicate pair for \"Example: Financial time series prediction with DataForge\"; 45872/45873 form a separate duplicate pair, and 45869 is unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is a strong global representative because it is earliest among the duplicate example issues and has an exact-title twin; however, this cluster is really two separate issue groups rather than one unified topic.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different subjects: a financial data bundle promo/request versus a DataForge example issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different subjects: a financial data bundle promo/request versus a DataForge example issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Exact same title and apparent same underlying example issue; clear duplicate." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both mention financial/time-series, but they describe different artifacts: an example issue versus a scraping-service integration." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Different issue concepts: DataForge example versus DataForge scraping-service integration." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Different issue concepts despite shared financial/DataForge wording; not the same underlying request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different issue concepts despite shared financial/DataForge wording; not the same underlying request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Exact same title and same underlying DataForge scraping-service issue; clear duplicate." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: the two exact-title pairs are reasonable duplicates, and 45869 is plausibly separate from the DataForge example/service issues. No soft-edge verdict appears overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-13", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "4e7698c4ce846fd7e6f972378e3dd979f4b575c950d6100c8cd833e246be2358", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34773, + "estimated_input_tokens": 8694, + "estimated_eval_tokens": 17644 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34773, + "estimated_input_tokens": 8694, + "estimated_eval_tokens": 17644 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly unrelated PRs that share broad themes like model onboarding, tokenizer fixes, or test updates. The only true near-duplicate is the pair of `fixed_cross_entropy` PRs, which both implement the same loss kwargs support for the same issue target.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "`pull_request:43251` is the best representative for the loss fix: it has the explicit issue target `43240`, touches the exact same code path as the companion PR, and presents the broader, more direct API change (`weight` and `label_smoothing`) in `fixed_cross_entropy`.", + "best_issue_reason": null, + "best_pr_reason": "`pull_request:43251` is the strongest global representative because it is the only PR in the cluster with an explicit issue target and it cleanly captures the concrete bug/fix pair around `fixed_cross_entropy`.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Both are unrelated model/runtime additions; they share only a loose theme of Transformers feature work, not the same underlying bug or change." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer method additions vs. executorch dynamic-shape testing are different changes in different code paths." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "A Mistral tokenizer API gap and adding MobileLLM are unrelated model-specific changes." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "These are both processor/model-pretrained related only at a broad level; they address different models and different implementation problems." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs support is unrelated to adding a new MobileLLM model implementation." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "The loss utility fix and Mistral4 test/model work do not target the same bug or code path." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape export tests and Mistral4 test fixes are separate test concerns, not duplicate fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "They cover different subsystems and different failure modes; no shared concrete change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "Adding MobileLLM and fixing `cls.config` resolution are unrelated model vs. core inheritance behavior changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "A new model implementation and Falcon output-interface refactoring are different tasks." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM addition and a model-creation skill/meta-PR are not the same underlying code change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "These are unrelated; one adds a model, the other changes class annotation/config resolution semantics." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output-collection refactor and Mistral4 fixes are different model-family changes with no shared concrete bug." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactoring is unrelated to the model-creation skill/meta automation PR." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test/model fixes and the `layrnorm` rename sweep are different bugs/changes, even if both touch model code." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "A model-specific fix PR and a model-creation skill PR are not duplicate fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 fixes and `cls.config` resolution order are unrelated internal behaviors." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 fixes and adding the TDT loss kernel are different subsystems and different concrete changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "The shared theme is only model development infrastructure; the concrete edits are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "A new model PR and a meta skill PR are not the same underlying change." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs modify `src/transformers/loss/loss_utils.py` to make `fixed_cross_entropy` pass supported kwargs to `nn.functional.cross_entropy`; they target the same bug and could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is well grounded in the packet. The only accepted soft-edge pair, `pull_request:43251` vs `pull_request:43254`, is clearly mergeable as a single bugfix PR: both touch `src/transformers/loss/loss_utils.py`, both target issue `43240`, and both extend `fixed_cross_entropy` to pass supported kwargs to `nn.functional.cross_entropy`. The other pair verdicts are conservative and do not overclaim overlap beyond broad thematic similarity. The summary also stays within the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 29, + "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": "deaa24b9e3f43f9b535319b7e8a7557041c8ca66d01759c6256e788b950f5afa", + "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": "Two related but distinct pull requests for the VibeVoice family: one adds the base model support, the other adds a realtime variant with additional tokenizer/feature-extractor/generation work. They overlap heavily in shared scaffolding, but they are not the same concrete change set.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42781 is the broader and more feature-complete VibeVoice-related change, but it is still a different deliverable from the base VibeVoice implementation in 40546 (realtime extension plus extra model components).", + "best_issue_reason": null, + "best_pr_reason": "PR 42781 is the best representative of the cluster because it covers the larger, newer VibeVoice integration surface and includes the realtime-specific additions beyond the base model support.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "They share the VibeVoice codebase and docs, but 40546 is the base implementation while 42781 adds a realtime variant with extra tokenizer/feature-extractor/generation changes. Related, but not the same underlying fix/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 29, + "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": "9226e9a4c51f760b804ac0222b00df236dc24483bf11e45dd1b816c348815219", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4676, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4676, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated feature PRs in different model areas: TimesFM covariate forecasting vs. DINOv3 image-classification support. They do not look like the same underlying change and should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR fits: the items address different models, code paths, and user-facing features.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a good global representative for the cluster because they are not duplicates and solve unrelated problems.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different subsystems and goals: TimesFM forecasting with covariates vs. DINOv3 image classification. The filenames, APIs, and diffs indicate separate feature work, not one 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": 29, + "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": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-3", + "nodes": [ + "pull_request:41797", + "pull_request:45186", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186", + "pull_request:45186|pull_request:45630" + ], + "prepared_review_unit_hash": "d9ad6072d6668006e9497862c02483ce96fda9a263dd89dc7644d0c1f62c487b", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three PRs are same-pattern model additions, but they target different models and code paths (DeepSeek OCR, Isaac, Kimi2-6), so they are not duplicates. Shared auto/docs plumbing is too generic to merge them together.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45630 is the broadest and most complete single model-addition PR in this cluster, but it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different models and implementations (DeepSeek OCR vs Isaac); shared docs/auto-registration changes are generic boilerplate, not the same underlying fix." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Different model families and code paths (Isaac vs Kimi2-6); they overlap only in standard integration files, so they should not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 29, + "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": "59615d1a5f05b4993bb991574c894c769c2d4cebcd871eee69ad460bc054b026", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two distinct workflow-maintenance PRs: one is a broad GitHub Actions upgrade for Node 24 compatibility, the other updates Docker build-related actions and a few workflow action versions. They overlap in CI files but do not appear to fix the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the cleaner representative of the cluster because it has a single, explicit goal (Node 24 compatibility) across many workflows; PR 42976 is a separate action-version refresh centered on Docker/build workflows.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the best single PR for this cluster because its scope is broader and more clearly defined, while PR 42976 is narrower and targets a different dependency set.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Reject: both touch GitHub Actions workflow files, but 42975 upgrades core action versions for Node 24 compatibility, while 42976 specifically bumps docker/build-push-action and related workflow actions. They are related maintenance changes, not the same concrete 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 + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 29, + "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-42978-5", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490", + "pull_request:46162" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490", + "pull_request:45490|pull_request:46162" + ], + "prepared_review_unit_hash": "d6d280af2cb9dbaee7905e6a902fe1815007c3adf05ef299dd5f488e1bd0a02f", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 16552, + "estimated_input_tokens": 4138, + "estimated_eval_tokens": 8532 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 16552, + "estimated_input_tokens": 4138, + "estimated_eval_tokens": 8532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "No duplicate cluster here: the PRs are separate model additions (ViT NEPA, Molmo, Molmo2, CTSM) plus an unrelated XPU test-only tweak. The overlapping docs/auto-registration/test scaffolding files are generic and don\u2019t indicate the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No PR is a good canonical duplicate candidate because these are distinct model-addition or test-only changes. If forced to pick a representative, 43451 is the most substantial single PR, but it is still not duplicate-equivalent to the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto files are standard registration plumbing, not the same bug/change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model additions: ViT NEPA vs Molmo2. The overlap is only generic model-registration/docs scaffolding." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different domains and models: ViT NEPA vs CTSM. Shared auto/docs edits are broad infrastructure, not one concrete fix." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are separate model implementations with different files and behavior; they are not the same change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated model additions; the common auto/docs files are incidental." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are distinct model additions with different code paths and artifacts; not mergeable as one duplicate PR." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46162", + "accept": false, + "reason": "CTSM model implementation vs a small XPU test-annotation tweak across existing tests; these are unrelated changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 29, + "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-43291-3", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43665|pull_request:44815" + ], + "prepared_review_unit_hash": "ccc2e9ca3ff8c58e2bf3a436ba44baa03e33736283049f1c66d4f09479856ad0", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7093, + "estimated_input_tokens": 1774, + "estimated_eval_tokens": 3804 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7093, + "estimated_input_tokens": 1774, + "estimated_eval_tokens": 3804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated pull requests: one adjusts Whisper tokenization/tests, one is test-only noise for CLIP/ViT, and one is a substantive FP8/dequantization fix. The superficial similarity is just that they are all PRs touching tests or model code, but they do not target the same bug or change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44815 is the strongest canonical PR because it contains the most concrete production fix: a dequant/loading-path change plus FP8 integration work across core code and a related model test. The other PRs are either test-only noise (#43665) or a narrower Whisper test/tokenization adjustment (#43291).", + "best_issue_reason": null, + "best_pr_reason": "PR #44815 is the best representative of the cluster because it is the most substantive and self-contained code change. It fixes a specific code path, unlike #43665 which is clearly not a real fix, and #43291 which is a narrower Whisper-specific test/tokenization tweak.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenization/test changes and dummy CLIP/ViT test edits are unrelated fixes; they do not address the same underlying bug or code path." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper decode/tokenization cleanup is unrelated to FP8 dequant/loading behavior. Different subsystems, different failure modes, no shared concrete fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "#43665 is test-only noise while #44815 is a real FP8/dequant fix. They do not plausibly merge into one PR for the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 29, + "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-43488-5", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:45932", + "pull_request:46100", + "pull_request:46114" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:46114", + "pull_request:43488|pull_request:43532", + "pull_request:43532|pull_request:46114", + "pull_request:45932|pull_request:46100", + "pull_request:46100|pull_request:46114" + ], + "prepared_review_unit_hash": "79d86350bcea9c8200a743a7c0dee81b871fe482fdd73260a00b54dc7aa204e2", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 5, + "serialized_chars": 11342, + "estimated_input_tokens": 2836, + "estimated_eval_tokens": 5928 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 5, + "serialized_chars": 11342, + "estimated_input_tokens": 2836, + "estimated_eval_tokens": 5928 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Most PRs are unrelated repo-test / bot-playground changes that happen to touch the same test file or subsystem. Only the two identical ViT test-migration PRs look like true duplicates of the same change.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46100 is the cleaner canonical representative: it is earlier, has the generic test-migration title, and its diff matches #46114 exactly.", + "best_issue_reason": null, + "best_pr_reason": "PR #46100 is the best overall representative of the duplicate pair because it is the earliest and most neutral description of the identical change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:46114", + "accept": false, + "reason": "Both touch tests/models/vit/test_modeling_vit.py, but the surrounding changes differ completely (#43488 also changes setup.py/dependency plumbing). Same file is not enough; not the same underlying PR." + }, + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Different edits and different intent: #43488 is a repo-bot/version/dependency tweak, while #43532 rewrites CI workflows and multiple model files. Not a duplicate change." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46114", + "accept": false, + "reason": "They only overlap on the ViT test file, but #43532 is a broad CI/workflow/model change and #46114 is a tiny test-migration stub. Not the same concrete change." + }, + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "#45932 is a large multi-file grafana/OTEL and model-code change; #46100 is a tiny ViT test-migration edit. Shared test-file noise does not make them duplicates." + }, + { + "left": "pull_request:46100", + "right": "pull_request:46114", + "accept": true, + "reason": "These PRs have the same diff in the same file and appear to be forked/duplicate test-migration PRs. They are the same underlying change and could be merged as one." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet. The only accepted pair, 46100 and 46114, has the same single-file diff and is plausibly a forked duplicate test-migration PR, so it is mergeable as one bugfix. The rejected pairs are conservative: they share at most a filename or subsystem, but the surrounding changes are clearly unrelated in scope and intent." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 29, + "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": "a1f3f6d87e081d3bfebf3c418cb5637905430851d0df38d52789f18809979709", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are closely related feature additions, but they target different segmentation pipelines and different model/code paths, so they should not be treated as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR as a duplicate target: PR 43612 adds Promptable Concept Segmentation, while PR 43613 adds Promptable Visual Segmentation with a different set of supported models and pipeline implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the broader, more infrastructure-heavy pipeline addition, but it still represents a distinct feature rather than the same change as PR 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "They share documentation/pipeline scaffolding, but the titles, filenames, and supported model sets show different underlying tasks and code paths, so they are not the same concrete change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 29, + "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": "c36e7d070e991b31485a82fce4e83a47dd8d89d45d8dc4f49b6044c15820f24d", + "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": "The two PRs both touch `trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for overrideability. They are not the same underlying change and should not be merged as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete bugfix-style change: it alters training loss scaling behavior for gradient accumulation. PR #43663 is a separate refactor around input column filtering and model signature inspection.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the best representative of the cluster because it targets a specific behavioral issue in the training loop, whereas #43663 is an unrelated extensibility refactor in the same file.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and different goals: #43651 changes gradient-accumulation loss normalization in `training_step`, while #43663 adds an overrideable helper for signature-based column filtering. Shared file alone is not enough to treat them as the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 29, + "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": "eb3175c3196b0d1fd677d3231e1e58a7f84be221c9ef1390ef3edd74f54d0cca", + "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": "These PRs overlap on the same CLI file and the same TypeAdapter NameError theme, but they appear to fix different runtime paths with different techniques, so they should not be treated as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more precise fix for the explicit TypeAdapter NameError when pydantic is unavailable, with a minimal targeted change in the affected code path.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 best captures the concrete bug report and applies the narrowest fix directly to the failing TypeAdapter annotation path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Both mention TypeAdapter NameError in src/transformers/cli/serve.py, but 43656 is a broader annotations/typing cleanup via __future__.annotations, while 43842 is a targeted fix for the pydantic-not-installed case. Different root causes and different fixes, so not a duplicate PR pair." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 29, + "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:44018|pull_request:44066", + "pull_request:44066|pull_request:44068", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44068", + "pull_request:44073|pull_request:44161", + "pull_request:44056|pull_request:44101", + "pull_request:44101|pull_request:44161", + "pull_request:44017|pull_request:44018", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:43997|pull_request:44018", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43997|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "dfc36ac762f1cefe3d92f14dda99c02166e3ed523b56b3588bc8f8a8c2d91376", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84901, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75096, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of model-specific output-tracing refactors under the same tracking issue. The only clear duplicates are the GPT-Neo pair and the GPT-J/CodeGen variants; the rest target different models and should stay separate.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the strongest canonical PR: it is explicitly tied to the tracking issue and covers the core GPT-J/CodeGen output-tracing refactor in the shared code path.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 best represents the underlying change because it spans the shared GPT-J/CodeGen implementation and is directly linked to the issue driving this refactor wave.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths; same tracking issue alone is not enough to merge them." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "These refactor different models (CVT/FNet vs DeBERTa v2), so they are not the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both touch the GPT-J output-tracing refactor in the same file/code path, so they look like duplicate variants of one change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "RegNet/ResNet/RT-DETR and EfficientNet are separate model implementations, not one duplicate bug fix." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths; only the generic refactor theme overlaps." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen and EfficientNet are different model code paths, so this is not a duplicate." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are unrelated implementations; same meta-pattern only." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are different model changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are separate model code paths, so this is not the same underlying change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo output-tracing refactor in the same model/file; 44068 is a later revision of the same change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different implementations, so they should not be merged." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model code paths, despite both being output-tracing refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen touch different code paths and do not look like the same bug." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are separate model-specific refactors, not duplicates." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are different model code paths; only the refactor style overlaps." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are distinct code paths, so this is not a duplicate." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model refactors." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are different code paths; same broad output-tracing theme is not enough." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated model changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR target different models and different forward paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are separate model-specific refactors." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are unrelated model code paths." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different models; they are only thematically similar." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J/CodeGen are different model code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are unrelated model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are separate model refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are different model code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are unrelated model implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated models." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are different model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are separate model-specific changes." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert/XLM do not share the same concrete code path." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are different models, so this is not a duplicate." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are separate model-specific refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44059", + "accept": false, + "reason": "GPT-Neo and GPT-2 are different models and forward paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-2 and GPT-Neo are unrelated model code paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are distinct model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "SegFormer and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and SegFormer are separate model-specific changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are different code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are separate model refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are separate model-specific changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa v2 are unrelated model code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are different model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "They modify the same GPT-J/CodeGen output-tracing code paths and look like duplicate iterations of the same refactor." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both touch the GPT-J output-tracing implementation in modeling_gptj.py, so they look like 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/ResNet/RT-DETR are different model-specific refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and ViTS are different model code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays close to the packet: the canonical PR choice is supported by the shared GPT-J/CodeGen file scope and explicit issue link, and the accept/reject calls are conservative for the cross-model pairs. No pair is overstated as mergeable beyond what the diffs and titles support." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 29, + "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": "ddea8061b1c372284d2f4bc2ffc7c51f71c0076a5eb03e35272260ef41217949", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84901, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are model-specific refactors to standardized output tracing/capture_outputs across different transformer architectures. They share the same broad maintenance theme, but each PR targets a different model code path, so they should not be deduplicated together.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44129 is the most representative standalone candidate because it has the clearest end-to-end refactor shape and an explicit issue target, but it is still model-specific rather than a cluster-wide duplicate.", + "best_issue_reason": null, + "best_pr_reason": "PR 44129 is the best single PR to anchor the cluster because it includes an explicit linked issue and a concrete output-tracing refactor, making it the strongest exemplar of this batch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Different model implementations (FocalNet vs ImageGPT). They follow the same refactor pattern, but they do not fix the same concrete code-path problem and would not plausibly merge into one PR." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate model code paths; this is the same kind of standardized-output migration, not the same underlying bug or change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert are unrelated model internals. Shared output-capture refactoring alone is too broad to treat as a duplicate." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These are different model-specific migrations, not one shared fix to a single code path." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are separate implementations; the PRs are analogous refactors, but not the same bug/change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 target different forward paths and output signatures, so they should remain separate PRs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 29, + "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-43998-15", + "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", + "pull_request:46106", + "pull_request:46255" + ], + "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: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", + "pull_request:46106|pull_request:46255", + "pull_request:44027|pull_request:46106", + "pull_request:44026|pull_request:46106", + "pull_request:44001|pull_request:46106", + "pull_request:44002|pull_request:46106", + "pull_request:44027|pull_request:46255", + "pull_request:44026|pull_request:46255", + "pull_request:44001|pull_request:46255", + "pull_request:44002|pull_request:46255" + ], + "prepared_review_unit_hash": "2c32db8e887026383eaa55812bb6031337e2119d81a09ee2a9b56b61fe772805", + "component_budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 48, + "serialized_chars": 46843, + "estimated_input_tokens": 11711, + "estimated_eval_tokens": 23678 + }, + "budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 48, + "serialized_chars": 46843, + "estimated_input_tokens": 11711, + "estimated_eval_tokens": 23678 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The set is not a duplicate cluster. Most items are separate model-specific output-tracing refactors, and the last pair is an unrelated docs/kernel rename. None of the soft edges should be merged as same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR fits: the artifacts split into distinct model-family refactors plus an unrelated docs/kernel change.", + "best_issue_reason": null, + "best_pr_reason": "No PR is a good global duplicate target; 44025 is the broadest of the output-tracing refactors, but it still covers only one model family and does not unify the rest.", + "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 architectures; same refactor theme but not the same bug or change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model implementation, so this is not a duplicate change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Falcon/Mamba vs Timm backbone are unrelated code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families; not mergeable as one concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "MobileNetV1 and UnivNet are unrelated model code paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models; shared tracing refactor wording is too broad." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and different forward paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "No shared concrete bug; only similar output-tracing cleanup." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Distinct encoder-decoder model wrappers, not the same underlying issue." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Vision-text dual encoder and UnivNet are unrelated code paths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model implementations; not the same change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Unrelated model families; cannot plausibly be one PR." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models and different output signatures." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "UnivNet and UperNet are unrelated architectures." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Separate model families; only the refactor style matches." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models and different tracing/capture changes." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet and RWKV are unrelated code paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model stacks; no single concrete fix." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Falcon/Mamba and CodeGen are different model implementations." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Timm backbone and CodeGen are unrelated; only both touch model outputs." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen do not share the same underlying bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "UperNet and CodeGen are different code paths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families; too broad to treat as duplicates." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Depth Anything and Vision-Encoder-Decoder are different concrete model wrappers." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Depth Anything and Speech-Encoder-Decoder are unrelated." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models; same refactor theme is insufficient." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Depth Anything and RWKV are unrelated code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Depth Anything and DPR are different model families." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Vision-Encoder-Decoder and Speech-Encoder-Decoder are separate wrappers, not one fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models; no shared concrete defect." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures and code paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Vision-Encoder-Decoder and DPR are unrelated changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech-Encoder-Decoder and SuperPoint are unrelated models." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different code paths; only the tracing refactor is common." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Speech-Encoder-Decoder and DPR do not fix the same bug." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint and RWKV are unrelated implementations." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model families and output behavior." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are different concrete code paths." + }, + { + "left": "pull_request:46106", + "right": "pull_request:46255", + "accept": false, + "reason": "Docs toctree reshaping and kernel-package renaming are unrelated changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:46106", + "accept": false, + "reason": "Model code vs docs navigation; not the same underlying change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:46106", + "accept": false, + "reason": "Different domains: model implementation versus documentation." + }, + { + "left": "pull_request:44001", + "right": "pull_request:46106", + "accept": false, + "reason": "UnivNet output tracing and docs toctree edits are unrelated." + }, + { + "left": "pull_request:44002", + "right": "pull_request:46106", + "accept": false, + "reason": "Model refactor versus docs restructuring; no duplicate basis." + }, + { + "left": "pull_request:44027", + "right": "pull_request:46255", + "accept": false, + "reason": "Speech-encoder-decoder refactor is unrelated to kernel package naming." + }, + { + "left": "pull_request:44026", + "right": "pull_request:46255", + "accept": false, + "reason": "Vision-encoder-decoder code and kernel docs/imports are different changes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:46255", + "accept": false, + "reason": "UnivNet refactor and liger-kernels rename do not target the same bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:46255", + "accept": false, + "reason": "UperNet tracing cleanup is unrelated to the kernel package rename." + } + ] + }, + "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": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-3", + "nodes": [ + "pull_request:44662", + "pull_request:45075", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401", + "pull_request:44662|pull_request:45075", + "pull_request:45075|pull_request:45401" + ], + "prepared_review_unit_hash": "3de8a058c425e710849e633676e7c95395a5ba7e0ebe70bd81275bfa47d80e8b", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10911, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10911, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests are separate model additions, not duplicates: PenguinVL, DeepSeek-OCR-2, and Voxtral TTS. They only overlap on standard Transformers boilerplate such as auto registries, docs TOC, and test scaffolding.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the three PRs implement different models and do not appear to fix or add the same concrete code path.", + "best_issue_reason": null, + "best_pr_reason": "No single PR is a good duplicate representative for the cluster because the changes are model-specific and not mergeable into one shared fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model implementations (PenguinVL vs Voxtral TTS); shared auto-registry/docs files are generic and do not indicate the same bug or change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "PenguinVL and DeepSeek-OCR-2 are distinct model adds with different files, conversion code, and tests; overlap is only framework boilerplate." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model additions; they do not target the same concrete code-path problem and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 29, + "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": "fd00323e0e43efe37e212de4e6ba94e6ee7db76c9e5122ea658c21b2012b3d9e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4408, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4408, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs about fusion are related at a high level but cover different layers: one adds a generic Module Fusion API, the other wires n-to-1 kernel fusion through KernelConfig and model loading. They are not the same concrete change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45363 is the more user-facing, end-to-end change: it threads KernelConfig through model loading, adds integration hooks, and includes an example. It better represents the practical fusion feature in this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #45363 is the strongest standalone PR in the cluster because it delivers the broader functional behavior and has the clearest application path. PR #44979 is a foundational API piece, but not the best single representative of the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Both are about fusion, but they target different code paths and abstractions: #44979 introduces a generic Module Fusion API in src/transformers/module_fusion.py, while #45363 implements n-to-1 kernel fusion via KernelConfig and related integration changes. They could coexist in separate PRs and are not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 29, + "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": "41422635a6e8ec1b2ea1d3f382e1aca7693f43ca2b17a936a771c9b506b7c076", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4413, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4413, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two pull requests address the same Mistral regex patch bug in the same code path; the later PR adds tests and is the more complete candidate.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR: it fixes the same AttributeError in _patch_mistral_regex and also adds test coverage, making it the more complete version of the change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the strongest overall representative because it contains the core fix plus tests, while PR 45086 is a smaller earlier version of the same patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Accept: both PRs target the same AttributeError in _patch_mistral_regex, touch the same file, and the later PR is effectively an expanded version of the earlier fix with tests." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "335d4b7688821f8d4d1305ff75bbd378d5e1402e66d989b43a168d5f51ee1e95", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related NemotronH pull requests, but they address different changes: one is a conversion-mapping/refactor for inheriting GraniteMoeHybrid-style weights, while the other splits NemotronH into dedicated dense/sparse model implementations and updates docs/auto-mapping.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better canonical representative because it is the broader, more complete architecture-level change and includes the new dense/sparse model split plus the associated docs and auto-mapping updates.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest standalone candidate for the cluster because it covers the main user-facing model reorganization, while PR #45115 is a narrower conversion-mapping refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They touch overlapping NemotronH code, but they do not fix the same concrete bug or implement the same change. #45115 is about checkpoint conversion key renames and inheritance mapping; #45569 introduces separate NemotronHDense/NemotronHSparse implementations and documentation changes. These are distinct PRs and not mergeable as one duplicate change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 29, + "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": "10b259a884d9b144ed5fc7fea945fba8f636e079edacef12cdd5ff11cd8210e2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related feature PRs, but not duplicates: one adds heterogeneous per-layer config support, the other adds heterogeneous modeling/runtime support on top of that foundation.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45332 is the broader, more representative change: it covers model/runtime integration for heterogeneous layers and incorporates the config machinery it needs. PR 45333 is the narrower config-only foundation.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the best single representative for the cluster because it reflects the end-user-facing feature and the larger code-path change, while 45333 is an enabling subset.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Do not merge as duplicates: 45333 is config-layer support only, while 45332 adds heterogeneous cache/masking/modeling behavior. They share infrastructure files, but they address different layers of the feature and 45332 appears to build on 45333 rather than fix the same concrete problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 29, + "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": "cebfe99123b860b8e3e20ee7b00c33fcae340e95db3a1048a17e5dc32694c1f5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are typing-related and overlap on a few files, but they are not the same change: one is a broad repository-wide type-checking sweep, while the other adds specific typing protocols/aliases and related model utilities support. The overlap looks incidental rather than a single duplicated fix.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the more focused and coherent typing change; PR 45415 is a much broader type-checking sweep across many files and config.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best representative of the cluster because it is narrower and centers on concrete typing infrastructure additions rather than a large cross-file type-checking pass.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Related typing work, but not the same underlying change: 45415 is a broad type-checking rollout, while 45425 adds specific typing aliases/protocols and model utility support. 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": 22, + "component_count": 29, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45487-8", + "nodes": [ + "pull_request:45487", + "pull_request:46069", + "pull_request:46073", + "pull_request:46109", + "pull_request:46151", + "pull_request:46154", + "pull_request:46196", + "pull_request:46230" + ], + "soft_pairs": [ + "pull_request:45487|pull_request:46230", + "pull_request:45487|pull_request:46151", + "pull_request:46069|pull_request:46073", + "pull_request:46069|pull_request:46154", + "pull_request:46069|pull_request:46196", + "pull_request:46073|pull_request:46154", + "pull_request:46073|pull_request:46196", + "pull_request:46109|pull_request:46151", + "pull_request:46151|pull_request:46230", + "pull_request:46154|pull_request:46196", + "pull_request:46154|pull_request:46230" + ], + "prepared_review_unit_hash": "18c21345f4c0271e28bef60a51b1ee5ad32cb2592d17896aceb6b74884e44790", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 11, + "serialized_chars": 21558, + "estimated_input_tokens": 5390, + "estimated_eval_tokens": 11036 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 11, + "serialized_chars": 21558, + "estimated_input_tokens": 5390, + "estimated_eval_tokens": 11036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated, model-specific PRs that share broad themes like SDPA/dispatch support, model-parallel fixes, or test expectation updates. They are not duplicates of the same concrete bug/change, so all soft edges should be rejected.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46154 is the clearest representative of the common SDPA-dispatch theme: it makes concrete code changes and updates tests, while the other PRs are narrower model-specific variants or unrelated trainer/test adjustments.", + "best_issue_reason": null, + "best_pr_reason": "PR #46154 is the best single representative of the cluster\u2019s recurring attention-dispatch work, with a substantial implementation plus test coverage; it is more central than the purely model-specific or test-only PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45487", + "right": "pull_request:46230", + "accept": false, + "reason": "Different bugs: #45487 fixes model-parallel token type/device handling in AltCLIP/ChineseCLIP, while #46230 fixes pad_token_id device placement during generation for multimodal model-parallel setups." + }, + { + "left": "pull_request:45487", + "right": "pull_request:46151", + "accept": false, + "reason": "One is a model-parallel runtime fix; the other is a test expectation update for distributed trainer device strings." + }, + { + "left": "pull_request:46069", + "right": "pull_request:46073", + "accept": false, + "reason": "Same broad feature area, but different models and different code paths: ShieldGemma2 vs CTRL SDPA dispatch." + }, + { + "left": "pull_request:46069", + "right": "pull_request:46154", + "accept": false, + "reason": "Both add attn dispatch support, but for different architectures (ShieldGemma2 vs T5/MusicGen); not the same concrete fix." + }, + { + "left": "pull_request:46069", + "right": "pull_request:46196", + "accept": false, + "reason": "Different model families and implementation details; this is not one shared bug or mergeable fix." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Both are SDPA dispatch PRs, but they target different model implementations and are not the same code-path problem." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "Different model architectures and separate dispatch implementations; too broad to treat as duplicates." + }, + { + "left": "pull_request:46109", + "right": "pull_request:46151", + "accept": false, + "reason": "Both are test-related, but they cover different test suites and different assertions; not the same underlying change." + }, + { + "left": "pull_request:46151", + "right": "pull_request:46230", + "accept": false, + "reason": "#46151 is a pipeline test expectation update; #46230 is a generation/model-parallel runtime fix." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "Different model-specific SDPA/attention dispatch implementations; same theme, but not the same bug or code path." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46230", + "accept": false, + "reason": "T5 SDPA dispatch and multimodal generation device realignment are unrelated fixes and would not plausibly collapse 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": 23, + "component_count": 29, + "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": "ed06a07864666b1fdf51998a87b0249dc43168248f7d35806792a30cb89aec0c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4859, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4859, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs both address cache-aware forward passes with seq_len > 1, but they target different model families and code paths, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45513 is the broader representative change in the cluster: it spans multiple Qwen/Olmo hybrid model files and fixes the multi-token cached-forward path rather than a single model-specific kernel.", + "best_issue_reason": null, + "best_pr_reason": "PR 46084 is the strongest standalone PR: it has an explicit linked issue, a focused Mamba2-specific fix, and a concrete seq_len > 1 cache bug in one code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Both involve cache handling for seq_len > 1, but 45513 fixes linear-attention/GDN cached forward across Qwen/Olmo variants while 46084 fixes Mamba2 CUDA-kernel cache logic. Different model families and different underlying code paths, so they should not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 29, + "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": "ada951a51f0fcb95f66c530f02f157678afb44b63e116f0ed72590ed261ec538", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs concern expert-parallel loading, but they fix different bugs: one is about serializing/loading `distributed_config` and adding `has_ep`, while the other prevents expert weights from being overwritten during FSDP2/rank-0 broadcast. They are related by theme, not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the stronger canonical PR: it targets a specific, concrete loading corruption path (expert parameters overwritten during distributed loading) and the diff directly implements the fix across the affected code paths.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the best fit for the cluster because it addresses the clearest underlying runtime bug in expert-parallel loading, whereas PR #45548 is about a different accelerate/ZeRO-3 serialization/loading compatibility issue.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Different failure modes and code paths: #45548 handles config serialization / `has_ep` for accelerate+ZeRO-3 loading, while #45662 fixes expert parameters being overwritten during FSDP2 rank-0 broadcast." + } + ] + }, + "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": 29, + "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": "9d16a74d3b4e377fe3b4367fbee6d9d95c538a2488919fa75d36f3e453252413", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4588, + "estimated_input_tokens": 1147, + "estimated_eval_tokens": 2550 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist/captured_info problem in testing_utils, with 45645 extending the same fix to CI log collection. They are closely related, but 45645 is the more complete change set.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the broader, more complete version of the xdist captured_info fix: it includes the core testing_utils change plus the workflow/log-consumption updates needed to preserve the debug artifacts.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the best representative of the cluster because it covers the underlying bug fix and the downstream CI handling changes in one place.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "They share the same core xdist-safe captured_info fix in testing_utils and same issue target; 45645 appears to be a superset that adds CI artifact handling, so they look like 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": 26, + "component_count": 29, + "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:46229" + ], + "soft_pairs": [ + "pull_request:45670|pull_request:46229" + ], + "prepared_review_unit_hash": "1ab5092bc1a21f62285d327f6bf6bea7dbca77134db8af4cc8b438922ddd6653", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4355, + "estimated_input_tokens": 1089, + "estimated_eval_tokens": 2434 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4355, + "estimated_input_tokens": 1089, + "estimated_eval_tokens": 2434 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are unrelated: one is a small AutoModel registration tweak for GLM ASR, while the other adds a substantial MTP generation feature across multiple files. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No duplicate/canonical relationship is evident between the two PRs; if one must be treated as the representative change, #46229 is the more substantial and self-contained feature PR.", + "best_issue_reason": null, + "best_pr_reason": "#46229 is the stronger standalone PR because it implements a concrete feature across generation code paths, whereas #45670 is a narrow model-registration nit.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46229", + "accept": false, + "reason": "Different code paths and goals: AutoModel registry placement vs. MTP candidate generation. They do not fix the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 29, + "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": 28, + "component_count": 29, + "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": "4876ba5941fb0f6e6a627d9383c2f7941eb6ea298962746101710455246313a5", + "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 two PRs are related only at a thematic level: both replace file-open source reading with safer source inspection, but they touch different code paths and fix different mechanisms in different modules. They should not be treated as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46207 is the better representative of the cluster because it is a concrete, narrowly scoped fix in `modeling_utils.py` with tests, addressing a specific source-reading heuristic in `_can_set_attn_implementation`.", + "best_issue_reason": null, + "best_pr_reason": "PR #46207 is the strongest standalone candidate: it has the clearest problem statement, the most explicit behavioral fix, and includes validation via tests. PR #46152 is a separate import-utils change and not a better cluster anchor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Both PRs improve source-file access, but they operate on different modules and different call sites (`import_utils.py` vs `modeling_utils.py`). The fixes are not the same underlying bug and would not plausibly be merged into one PR without broadening into unrelated code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 29, + "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-46259-2", + "nodes": [ + "pull_request:46259", + "pull_request:46260" + ], + "soft_pairs": [ + "pull_request:46259|pull_request:46260" + ], + "prepared_review_unit_hash": "461374eb8d8bb06e43b25ff24b2739e5167316a1edcc198e726d32afafb1f680", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6291, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate modeling-rule PRs affecting different model families and different code paths: one targets image/video processor do_* flag consumption, the other targets audio/speech model rule coverage. They look superficially similar because they both add docs/rules infrastructure, but they are not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the two pull requests implement different modeling rules for different subsystems, so they should remain separate.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for the cluster; each PR is the right representative for its own distinct rule addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46259", + "right": "pull_request:46260", + "accept": false, + "reason": "Different underlying rule additions and model families; they share documentation/rules plumbing but not the same concrete bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/manifest.json b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..9cd5a8dfeb882e6b24b49e58981e967dcbddbf95 --- /dev/null +++ b/snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260528t120029z", + "artifacts": { + "hybrid": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-05-28T12:12:32Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260528T120029Z", + "variant": "hybrid" +} diff --git a/snapshots/20260528T120029Z/manifest.json b/snapshots/20260528T120029Z/manifest.json index 8a138b9b65cdafa9a6ceb66164b39041aba09a37..81d956314cac8fcfa346e10a361b670092ca9cbf 100644 --- a/snapshots/20260528T120029Z/manifest.json +++ b/snapshots/20260528T120029Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 1147 }, "extracted_at": "2026-05-28T12:00:29Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260528t120029z", + "runs": { + "hybrid-model-20260528t120029z": { + "analysis_id": "hybrid-model-20260528t120029z", + "artifacts": { + "hybrid": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260528T120029Z/analysis-runs/hybrid-model-20260528t120029z/manifest.json", + "model": null, + "published_at": "2026-05-28T12:12:32Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260528T120029Z", "watermark": {